Add support for IQ1_S, IQ3_S, IQ2_S, IQ4_XS. IQ4_NL

Co-authored-by: ManniX-ITA <20623405+mann1x@users.noreply.github.com>

.github/workflows/release.yaml

@@ -103,6 +103,7 @@ jobs:
           path: |
             llm/build/**/bin/*
             llm/build/**/*.a
+            dist/windows-amd64/**
 
   # ROCm generation step
   generate-windows-rocm:
@@ -173,7 +174,9 @@ jobs:
       - uses: actions/upload-artifact@v4
         with:
           name: generate-windows-rocm
-          path: llm/build/**/bin/*
+          path: |
+            llm/build/**/bin/*
+            dist/windows-amd64/**
       - uses: actions/upload-artifact@v4
         with:
           name: windows-rocm-deps
@@ -253,7 +256,9 @@ jobs:
       - uses: actions/upload-artifact@v4
         with:
           name: generate-windows-cuda
-          path: llm/build/**/bin/*
+          path: |
+            llm/build/**/bin/*
+            dist/windows-amd64/**
       - uses: actions/upload-artifact@v4
         with:
           name: windows-cuda-deps
@@ -306,23 +311,18 @@ jobs:
       - uses: actions/download-artifact@v4
         with:
           name: generate-windows-cpu
-          path: llm/build
       - uses: actions/download-artifact@v4
         with:
           name: generate-windows-cuda
-          path: llm/build
       - uses: actions/download-artifact@v4
         with:
           name: windows-cuda-deps
-          path: dist/deps
       - uses: actions/download-artifact@v4
         with:
           name: windows-rocm-deps
-          path: dist/deps
       - uses: actions/download-artifact@v4
         with:
           name: generate-windows-rocm
-          path: llm/build
       - run: dir llm/build
       - run: |
           $gopath=(get-command go).source | split-path -parent
@@ -331,13 +331,13 @@ jobs:
           $env:CMAKE_SYSTEM_VERSION="10.0.22621.0"
           $env:PATH="$gopath;$env:PATH"
           $env:OLLAMA_SKIP_GENERATE="1"
-          $env:NVIDIA_DIR=$(resolve-path ".\dist\deps")
-          $env:HIP_PATH=$(resolve-path ".\dist\deps")
           & .\scripts\build_windows.ps1
       - uses: actions/upload-artifact@v4
         with:
           name: dist-windows
-          path: dist/*.exe
+          path: |
+            dist/OllamaSetup.exe
+            dist/ollama-windows-*.zip
 
   # Linux x86 assets built using the container based build
   build-linux-amd64:

.github/workflows/test.yaml

@@ -1,5 +1,15 @@
 name: test
 
+concurrency:
+  # For PRs, later CI runs preempt previous ones. e.g. a force push on a PR
+  # cancels running CI jobs and starts all new ones.
+  #
+  # For non-PR pushes, concurrency.group needs to be unique for every distinct
+  # CI run we want to have happen. Use run_id, which in practice means all
+  # non-PR CI runs will be allowed to run without preempting each other.
+  group: ${{ github.workflow }}-$${{ github.pull_request.number || github.run_id }}
+  cancel-in-progress: true
+
 on:
   pull_request:
     paths:
@@ -21,7 +31,9 @@ jobs:
       - id: changes
         run: |
           changed() {
-            git diff-tree -r --no-commit-id --name-only ${{ github.event.pull_request.base.sha }} ${{ github.event.pull_request.head.sha }} \
+            git diff-tree -r --no-commit-id --name-only \
+              $(git merge-base ${{ github.event.pull_request.base.sha }} ${{ github.event.pull_request.head.sha }}) \
+              ${{ github.event.pull_request.head.sha }} \
               | xargs python3 -c "import sys; print(any([x.startswith('$1') for x in sys.argv[1:]]))"
           }
 
@@ -103,7 +115,9 @@ jobs:
       - uses: actions/upload-artifact@v4
         with:
           name: cuda-${{ matrix.cuda-version }}-libraries
-          path: llm/build/**/bin/*
+          path: |
+            llm/build/**/bin/*
+            dist/windows-amd64/**
   generate-rocm:
     needs: [changes]
     if: ${{ needs.changes.outputs.GENERATE_ROCM == 'True' }}
@@ -134,7 +148,9 @@ jobs:
       - uses: actions/upload-artifact@v4
         with:
           name: rocm-${{ matrix.rocm-version }}-libraries
-          path: llm/build/**/bin/*
+          path: |
+            llm/build/**/bin/*
+            dist/windows-amd64/**
 
   # ROCm generation step
   generate-windows-rocm:
@@ -253,14 +269,9 @@ jobs:
           mkdir -p llm/build/darwin/$ARCH/stub/bin
           touch llm/build/darwin/$ARCH/stub/bin/ollama_llama_server
         if: ${{ startsWith(matrix.os, 'macos-') }}
-      - run: |
-          mkdir -p llm/build/windows/$ARCH/stub/bin
-          touch llm/build/windows/$ARCH/stub/bin/ollama_llama_server
-        if: ${{ startsWith(matrix.os, 'windows-') }}
-        shell: bash
       - uses: golangci/golangci-lint-action@v4
         with:
-          args: --timeout 8m0s
+          args: --timeout 8m0s -v
   test:
     strategy:
       matrix:
@@ -284,7 +295,6 @@ jobs:
         with:
           go-version-file: go.mod
           cache: true
-      - run: go get
       - run: |
           case ${{ matrix.arch }} in
             amd64) echo ARCH=x86_64 ;;
@@ -299,10 +309,6 @@ jobs:
           mkdir -p llm/build/darwin/$ARCH/stub/bin
           touch llm/build/darwin/$ARCH/stub/bin/ollama_llama_server
         if: ${{ startsWith(matrix.os, 'macos-') }}
-      - run: |
-          mkdir -p llm/build/windows/$ARCH/stub/bin
-          touch llm/build/windows/$ARCH/stub/bin/ollama_llama_server
-        if: ${{ startsWith(matrix.os, 'windows-') }}
         shell: bash
       - run: go generate ./...
       - run: go build

.gitignore

@@ -11,4 +11,5 @@ ggml-metal.metal
 .idea
 test_data
 *.crt
-llm/build
+llm/build
+__debug_bin*

README.md

@@ -1,5 +1,5 @@
 <div align="center">
-  <img alt="ollama" height="200px" src="https://github.com/ollama/ollama/assets/3325447/0d0b44e2-8f4a-4e99-9b52-a5c1c741c8f7">
+ <img alt="ollama" height="200px" src="https://github.com/ollama/ollama/assets/3325447/0d0b44e2-8f4a-4e99-9b52-a5c1c741c8f7">
 </div>
 
 # Ollama
@@ -51,16 +51,12 @@ Here are some example models that can be downloaded:
 | ------------------ | ---------- | ----- | ------------------------------ |
 | Llama 3            | 8B         | 4.7GB | `ollama run llama3`            |
 | Llama 3            | 70B        | 40GB  | `ollama run llama3:70b`        |
+| Phi-3              | 3.8B       | 2.3GB | `ollama run phi3`              |
 | Mistral            | 7B         | 4.1GB | `ollama run mistral`           |
-| Dolphin Phi        | 2.7B       | 1.6GB | `ollama run dolphin-phi`       |
-| Phi-2              | 2.7B       | 1.7GB | `ollama run phi`               |
 | Neural Chat        | 7B         | 4.1GB | `ollama run neural-chat`       |
 | Starling           | 7B         | 4.1GB | `ollama run starling-lm`       |
 | Code Llama         | 7B         | 3.8GB | `ollama run codellama`         |
 | Llama 2 Uncensored | 7B         | 3.8GB | `ollama run llama2-uncensored` |
-| Llama 2 13B        | 13B        | 7.3GB | `ollama run llama2:13b`        |
-| Llama 2 70B        | 70B        | 39GB  | `ollama run llama2:70b`        |
-| Orca Mini          | 3B         | 1.9GB | `ollama run orca-mini`         |
 | LLaVA              | 7B         | 4.5GB | `ollama run llava`             |
 | Gemma              | 2B         | 1.4GB | `ollama run gemma:2b`          |
 | Gemma              | 7B         | 4.8GB | `ollama run gemma:7b`          |
@@ -177,7 +173,7 @@ I'm a basic program that prints the famous "Hello, world!" message to the consol
 The image features a yellow smiley face, which is likely the central focus of the picture.
 ```
 
-### Pass in prompt as arguments
+### Pass the prompt as an argument
 
 ```
 $ ollama run llama3 "Summarize this file: $(cat README.md)"
@@ -260,16 +256,17 @@ See the [API documentation](./docs/api.md) for all endpoints.
 
 ### Web & Desktop
 
+- [Open WebUI](https://github.com/open-webui/open-webui)
+- [Enchanted (macOS native)](https://github.com/AugustDev/enchanted)
 - [Lollms-Webui](https://github.com/ParisNeo/lollms-webui)
 - [LibreChat](https://github.com/danny-avila/LibreChat)
 - [Bionic GPT](https://github.com/bionic-gpt/bionic-gpt)
-- [Enchanted (macOS native)](https://github.com/AugustDev/enchanted)
 - [HTML UI](https://github.com/rtcfirefly/ollama-ui)
 - [Saddle](https://github.com/jikkuatwork/saddle)
 - [Chatbot UI](https://github.com/ivanfioravanti/chatbot-ollama)
+- [Chatbot UI v2](https://github.com/mckaywrigley/chatbot-ui)
 - [Typescript UI](https://github.com/ollama-interface/Ollama-Gui?tab=readme-ov-file)
 - [Minimalistic React UI for Ollama Models](https://github.com/richawo/minimal-llm-ui)
-- [Open WebUI](https://github.com/open-webui/open-webui)
 - [Ollamac](https://github.com/kevinhermawan/Ollamac)
 - [big-AGI](https://github.com/enricoros/big-AGI/blob/main/docs/config-local-ollama.md)
 - [Cheshire Cat assistant framework](https://github.com/cheshire-cat-ai/core)
@@ -291,9 +288,13 @@ See the [API documentation](./docs/api.md) for all endpoints.
 - [AnythingLLM (Docker + MacOs/Windows/Linux native app)](https://github.com/Mintplex-Labs/anything-llm)
 - [Ollama Basic Chat: Uses HyperDiv Reactive UI](https://github.com/rapidarchitect/ollama_basic_chat)
 - [Ollama-chats RPG](https://github.com/drazdra/ollama-chats)
+- [QA-Pilot: Chat with Code Repository](https://github.com/reid41/QA-Pilot)
 - [ChatOllama: Open Source Chatbot based on Ollama with Knowledge Bases](https://github.com/sugarforever/chat-ollama)
 - [CRAG Ollama Chat: Simple Web Search with Corrective RAG](https://github.com/Nagi-ovo/CRAG-Ollama-Chat)
 - [RAGFlow: Open-source Retrieval-Augmented Generation engine based on deep document understanding](https://github.com/infiniflow/ragflow)
+- [chat: chat web app for teams](https://github.com/swuecho/chat)
+- [Lobe Chat](https://github.com/lobehub/lobe-chat) with [Integrating Doc](https://lobehub.com/docs/self-hosting/examples/ollama)
+- [Ollama RAG Chatbot: Local Chat with multiple PDFs using Ollama and RAG.](https://github.com/datvodinh/rag-chatbot.git)
 
 ### Terminal
 
@@ -309,11 +310,13 @@ See the [API documentation](./docs/api.md) for all endpoints.
 - [Oatmeal](https://github.com/dustinblackman/oatmeal)
 - [cmdh](https://github.com/pgibler/cmdh)
 - [ooo](https://github.com/npahlfer/ooo)
+- [shell-pilot](https://github.com/reid41/shell-pilot)
 - [tenere](https://github.com/pythops/tenere)
 - [llm-ollama](https://github.com/taketwo/llm-ollama) for [Datasette's LLM CLI](https://llm.datasette.io/en/stable/).
 - [typechat-cli](https://github.com/anaisbetts/typechat-cli)
 - [ShellOracle](https://github.com/djcopley/ShellOracle)
 - [tlm](https://github.com/yusufcanb/tlm)
+- [podman-ollama](https://github.com/ericcurtin/podman-ollama)
 
 ### Database
 
@@ -378,6 +381,7 @@ See the [API documentation](./docs/api.md) for all endpoints.
 - [Page Assist](https://github.com/n4ze3m/page-assist) (Chrome Extension)
 - [AI Telegram Bot](https://github.com/tusharhero/aitelegrambot) (Telegram bot using Ollama in backend)
 - [AI ST Completion](https://github.com/yaroslavyaroslav/OpenAI-sublime-text) (Sublime Text 4 AI assistant plugin with Ollama support)
+- [Discord-Ollama Chat Bot](https://github.com/kevinthedang/discord-ollama) (Generalized TypeScript Discord Bot w/ Tuning Documentation)
 
 ### Supported backends 
-- [llama.cpp](https://github.com/ggerganov/llama.cpp) project founded by Georgi Gerganov. 
+- [llama.cpp](https://github.com/ggerganov/llama.cpp) project founded by Georgi Gerganov. 

api/client.go

@@ -18,6 +18,7 @@ import (
 	"net/url"
 	"os"
 	"runtime"
+	"strconv"
 	"strings"
 
 	"github.com/ollama/ollama/format"
@@ -57,12 +58,36 @@ func checkError(resp *http.Response, body []byte) error {
 // If the variable is not specified, a default ollama host and port will be
 // used.
 func ClientFromEnvironment() (*Client, error) {
+	ollamaHost, err := GetOllamaHost()
+	if err != nil {
+		return nil, err
+	}
+
+	return &Client{
+		base: &url.URL{
+			Scheme: ollamaHost.Scheme,
+			Host:   net.JoinHostPort(ollamaHost.Host, ollamaHost.Port),
+		},
+		http: http.DefaultClient,
+	}, nil
+}
+
+type OllamaHost struct {
+	Scheme string
+	Host   string
+	Port   string
+}
+
+func GetOllamaHost() (OllamaHost, error) {
 	defaultPort := "11434"
 
-	scheme, hostport, ok := strings.Cut(os.Getenv("OLLAMA_HOST"), "://")
+	hostVar := os.Getenv("OLLAMA_HOST")
+	hostVar = strings.TrimSpace(strings.Trim(strings.TrimSpace(hostVar), "\"'"))
+
+	scheme, hostport, ok := strings.Cut(hostVar, "://")
 	switch {
 	case !ok:
-		scheme, hostport = "http", os.Getenv("OLLAMA_HOST")
+		scheme, hostport = "http", hostVar
 	case scheme == "http":
 		defaultPort = "80"
 	case scheme == "https":
@@ -82,15 +107,24 @@ func ClientFromEnvironment() (*Client, error) {
 		}
 	}
 
-	return &Client{
-		base: &url.URL{
-			Scheme: scheme,
-			Host:   net.JoinHostPort(host, port),
-		},
-		http: http.DefaultClient,
+	if portNum, err := strconv.ParseInt(port, 10, 32); err != nil || portNum > 65535 || portNum < 0 {
+		return OllamaHost{}, ErrInvalidHostPort
+	}
+
+	return OllamaHost{
+		Scheme: scheme,
+		Host:   host,
+		Port:   port,
 	}, nil
 }
 
+func NewClient(base *url.URL, http *http.Client) *Client {
+	return &Client{
+		base: base,
+		http: http,
+	}
+}
+
 func (c *Client) do(ctx context.Context, method, path string, reqData, respData any) error {
 	var reqBody io.Reader
 	var data []byte

api/client_test.go

@@ -1,6 +1,12 @@
 package api
 
-import "testing"
+import (
+	"fmt"
+	"net"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
 
 func TestClientFromEnvironment(t *testing.T) {
 	type testCase struct {
@@ -40,4 +46,40 @@ func TestClientFromEnvironment(t *testing.T) {
 			}
 		})
 	}
+
+	hostTestCases := map[string]*testCase{
+		"empty":               {value: "", expect: "127.0.0.1:11434"},
+		"only address":        {value: "1.2.3.4", expect: "1.2.3.4:11434"},
+		"only port":           {value: ":1234", expect: ":1234"},
+		"address and port":    {value: "1.2.3.4:1234", expect: "1.2.3.4:1234"},
+		"hostname":            {value: "example.com", expect: "example.com:11434"},
+		"hostname and port":   {value: "example.com:1234", expect: "example.com:1234"},
+		"zero port":           {value: ":0", expect: ":0"},
+		"too large port":      {value: ":66000", err: ErrInvalidHostPort},
+		"too small port":      {value: ":-1", err: ErrInvalidHostPort},
+		"ipv6 localhost":      {value: "[::1]", expect: "[::1]:11434"},
+		"ipv6 world open":     {value: "[::]", expect: "[::]:11434"},
+		"ipv6 no brackets":    {value: "::1", expect: "[::1]:11434"},
+		"ipv6 + port":         {value: "[::1]:1337", expect: "[::1]:1337"},
+		"extra space":         {value: " 1.2.3.4 ", expect: "1.2.3.4:11434"},
+		"extra quotes":        {value: "\"1.2.3.4\"", expect: "1.2.3.4:11434"},
+		"extra space+quotes":  {value: " \" 1.2.3.4 \" ", expect: "1.2.3.4:11434"},
+		"extra single quotes": {value: "'1.2.3.4'", expect: "1.2.3.4:11434"},
+	}
+
+	for k, v := range hostTestCases {
+		t.Run(k, func(t *testing.T) {
+			t.Setenv("OLLAMA_HOST", v.value)
+
+			oh, err := GetOllamaHost()
+			if err != v.err {
+				t.Fatalf("expected %s, got %s", v.err, err)
+			}
+
+			if err == nil {
+				host := net.JoinHostPort(oh.Host, oh.Port)
+				assert.Equal(t, v.expect, host, fmt.Sprintf("%s: expected %s, got %s", k, v.expect, host))
+			}
+		})
+	}
 }

api/types.go

@@ -98,8 +98,7 @@ type ChatResponse struct {
 	CreatedAt time.Time `json:"created_at"`
 	Message   Message   `json:"message"`
 
-	Done       bool   `json:"done"`
-	DoneReason string `json:"done_reason,omitempty"`
+	Done bool `json:"done"`
 
 	Metrics
 }
@@ -266,9 +265,8 @@ type GenerateResponse struct {
 	CreatedAt time.Time `json:"created_at"`
 	Response  string    `json:"response"`
 
-	Done       bool   `json:"done"`
-	DoneReason string `json:"done_reason,omitempty"`
-	Context    []int  `json:"context,omitempty"`
+	Done    bool  `json:"done"`
+	Context []int `json:"context,omitempty"`
 
 	Metrics
 }
@@ -311,6 +309,7 @@ func (m *Metrics) Summary() {
 }
 
 var ErrInvalidOpts = errors.New("invalid options")
+var ErrInvalidHostPort = errors.New("invalid port specified in OLLAMA_HOST")
 
 func (opts *Options) FromMap(m map[string]interface{}) error {
 	valueOpts := reflect.ValueOf(opts).Elem() // names of the fields in the options struct
@@ -398,8 +397,10 @@ func (opts *Options) FromMap(m map[string]interface{}) error {
 func DefaultOptions() Options {
 	return Options{
 		// options set on request to runner
-		NumPredict:       -1,
-		NumKeep:          0,
+		NumPredict: -1,
+
+		// set a minimal num_keep to avoid issues on context shifts
+		NumKeep:          4,
 		Temperature:      0.8,
 		TopK:             40,
 		TopP:             0.9,

app/lifecycle/server.go

@@ -43,37 +43,36 @@ func getCLIFullPath(command string) string {
 	return command
 }
 
-func SpawnServer(ctx context.Context, command string) (chan int, error) {
-	done := make(chan int)
-
-	logDir := filepath.Dir(ServerLogFile)
-	_, err := os.Stat(logDir)
-	if errors.Is(err, os.ErrNotExist) {
-		if err := os.MkdirAll(logDir, 0o755); err != nil {
-			return done, fmt.Errorf("create ollama server log dir %s: %v", logDir, err)
-		}
-	}
-
+func start(ctx context.Context, command string) (*exec.Cmd, error) {
 	cmd := getCmd(ctx, getCLIFullPath(command))
-	// send stdout and stderr to a file
 	stdout, err := cmd.StdoutPipe()
 	if err != nil {
-		return done, fmt.Errorf("failed to spawn server stdout pipe %s", err)
+		return nil, fmt.Errorf("failed to spawn server stdout pipe: %w", err)
 	}
 	stderr, err := cmd.StderrPipe()
 	if err != nil {
-		return done, fmt.Errorf("failed to spawn server stderr pipe %s", err)
-	}
-	stdin, err := cmd.StdinPipe()
-	if err != nil {
-		return done, fmt.Errorf("failed to spawn server stdin pipe %s", err)
+		return nil, fmt.Errorf("failed to spawn server stderr pipe: %w", err)
 	}
 
 	// TODO - rotation
 	logFile, err := os.OpenFile(ServerLogFile, os.O_APPEND|os.O_WRONLY|os.O_CREATE, 0755)
 	if err != nil {
-		return done, fmt.Errorf("failed to create server log %w", err)
+		return nil, fmt.Errorf("failed to create server log: %w", err)
 	}
+
+	logDir := filepath.Dir(ServerLogFile)
+	_, err = os.Stat(logDir)
+	if err != nil {
+		if !errors.Is(err, os.ErrNotExist) {
+			return nil, fmt.Errorf("stat ollama server log dir %s: %v", logDir, err)
+
+		}
+
+		if err := os.MkdirAll(logDir, 0o755); err != nil {
+			return nil, fmt.Errorf("create ollama server log dir %s: %v", logDir, err)
+		}
+	}
+
 	go func() {
 		defer logFile.Close()
 		io.Copy(logFile, stdout) //nolint:errcheck
@@ -117,19 +116,33 @@ func SpawnServer(ctx context.Context, command string) (chan int, error) {
 
 	// run the command and wait for it to finish
 	if err := cmd.Start(); err != nil {
-		return done, fmt.Errorf("failed to start server %w", err)
+		return nil, fmt.Errorf("failed to start server %w", err)
 	}
 	if cmd.Process != nil {
 		slog.Info(fmt.Sprintf("started ollama server with pid %d", cmd.Process.Pid))
 	}
 	slog.Info(fmt.Sprintf("ollama server logs %s", ServerLogFile))
 
+	return cmd, nil
+}
+
+func SpawnServer(ctx context.Context, command string) (chan int, error) {
+	done := make(chan int)
+
 	go func() {
 		// Keep the server running unless we're shuttind down the app
 		crashCount := 0
 		for {
+			slog.Info("starting server...")
+			cmd, err := start(ctx, command)
+			if err != nil {
+				crashCount++
+				slog.Error(fmt.Sprintf("failed to start server %s", err))
+				time.Sleep(500 * time.Millisecond * time.Duration(crashCount))
+				continue
+			}
+
 			cmd.Wait() //nolint:errcheck
-			stdin.Close()
 			var code int
 			if cmd.ProcessState != nil {
 				code = cmd.ProcessState.ExitCode()
@@ -143,15 +156,12 @@ func SpawnServer(ctx context.Context, command string) (chan int, error) {
 			default:
 				crashCount++
 				slog.Warn(fmt.Sprintf("server crash %d - exit code %d - respawning", crashCount, code))
-				time.Sleep(500 * time.Millisecond)
-				if err := cmd.Start(); err != nil {
-					slog.Error(fmt.Sprintf("failed to restart server %s", err))
-					// Keep trying, but back off if we keep failing
-					time.Sleep(time.Duration(crashCount) * time.Second)
-				}
+				time.Sleep(500 * time.Millisecond * time.Duration(crashCount))
+				break
 			}
 		}
 	}()
+
 	return done, nil
 }
 

app/ollama.iss

@@ -88,15 +88,12 @@ DialogFontSize=12
 [Files]
 Source: ".\app.exe"; DestDir: "{app}"; DestName: "{#MyAppExeName}" ; Flags: ignoreversion 64bit
 Source: "..\ollama.exe"; DestDir: "{app}"; Flags: ignoreversion 64bit
-Source: "..\dist\windeps\*.dll"; DestDir: "{app}"; Flags: ignoreversion 64bit
+Source: "..\dist\windows-{#ARCH}\*.dll"; DestDir: "{app}"; Flags: ignoreversion 64bit
+Source: "..\dist\windows-{#ARCH}\ollama_runners\*"; DestDir: "{app}\ollama_runners"; Flags: ignoreversion 64bit recursesubdirs
 Source: "..\dist\ollama_welcome.ps1"; DestDir: "{app}"; Flags: ignoreversion
 Source: ".\assets\app.ico"; DestDir: "{app}"; Flags: ignoreversion
-; Assumes v5.7, may need adjustments for v6
-#if GetEnv("HIP_PATH") != ""
-  Source: "{#GetEnv('HIP_PATH')}\bin\hipblas.dll"; DestDir: "{app}\rocm\"; Flags: ignoreversion
-  Source: "{#GetEnv('HIP_PATH')}\bin\rocblas.dll"; DestDir: "{app}\rocm\"; Flags: ignoreversion
-  ; amdhip64.dll dependency comes from the driver and must be installed already
-  Source: "{#GetEnv('HIP_PATH')}\bin\rocblas\library\*"; DestDir: "{app}\rocm\rocblas\library\"; Flags: ignoreversion
+#if DirExists("..\dist\windows-amd64\rocm")
+  Source: "..\dist\windows-amd64\rocm\*"; DestDir: "{app}\rocm\"; Flags: ignoreversion recursesubdirs
 #endif
 
 
@@ -132,7 +129,7 @@ SetupAppRunningError=Another Ollama installer is running.%n%nPlease cancel or fi
 
 
 ;FinishedHeadingLabel=Run your first model
-;FinishedLabel=%nRun this command in a PowerShell or cmd terminal.%n%n%n    ollama run llama2
+;FinishedLabel=%nRun this command in a PowerShell or cmd terminal.%n%n%n    ollama run llama3
 ;ClickFinish=%n
 
 [Registry]

auth/auth.go

@@ -10,12 +10,44 @@ import (
 	"log/slog"
 	"os"
 	"path/filepath"
+	"strings"
 
 	"golang.org/x/crypto/ssh"
 )
 
 const defaultPrivateKey = "id_ed25519"
 
+func keyPath() (string, error) {
+	home, err := os.UserHomeDir()
+	if err != nil {
+		return "", err
+	}
+
+	return filepath.Join(home, ".ollama", defaultPrivateKey), nil
+}
+
+func GetPublicKey() (string, error) {
+	keyPath, err := keyPath()
+	if err != nil {
+		return "", err
+	}
+
+	privateKeyFile, err := os.ReadFile(keyPath)
+	if err != nil {
+		slog.Info(fmt.Sprintf("Failed to load private key: %v", err))
+		return "", err
+	}
+
+	privateKey, err := ssh.ParsePrivateKey(privateKeyFile)
+	if err != nil {
+		return "", err
+	}
+
+	publicKey := ssh.MarshalAuthorizedKey(privateKey.PublicKey())
+
+	return strings.TrimSpace(string(publicKey)), nil
+}
+
 func NewNonce(r io.Reader, length int) (string, error) {
 	nonce := make([]byte, length)
 	if _, err := io.ReadFull(r, nonce); err != nil {
@@ -26,13 +58,11 @@ func NewNonce(r io.Reader, length int) (string, error) {
 }
 
 func Sign(ctx context.Context, bts []byte) (string, error) {
-	home, err := os.UserHomeDir()
+	keyPath, err := keyPath()
 	if err != nil {
 		return "", err
 	}
 
-	keyPath := filepath.Join(home, ".ollama", defaultPrivateKey)
-
 	privateKeyFile, err := os.ReadFile(keyPath)
 	if err != nil {
 		slog.Info(fmt.Sprintf("Failed to load private key: %v", err))

cmd/cmd.go

@@ -17,6 +17,7 @@ import (
 	"os"
 	"os/signal"
 	"path/filepath"
+	"regexp"
 	"runtime"
 	"strings"
 	"syscall"
@@ -31,10 +32,12 @@ import (
 	"golang.org/x/term"
 
 	"github.com/ollama/ollama/api"
+	"github.com/ollama/ollama/auth"
 	"github.com/ollama/ollama/format"
-	"github.com/ollama/ollama/parser"
 	"github.com/ollama/ollama/progress"
 	"github.com/ollama/ollama/server"
+	"github.com/ollama/ollama/types/errtypes"
+	"github.com/ollama/ollama/types/model"
 	"github.com/ollama/ollama/version"
 )
 
@@ -53,14 +56,13 @@ func CreateHandler(cmd *cobra.Command, args []string) error {
 	p := progress.NewProgress(os.Stderr)
 	defer p.Stop()
 
-	bars := make(map[string]*progress.Bar)
-
-	modelfile, err := os.ReadFile(filename)
+	f, err := os.Open(filename)
 	if err != nil {
 		return err
 	}
+	defer f.Close()
 
-	commands, err := parser.Parse(bytes.NewReader(modelfile))
+	modelfile, err := model.ParseFile(f)
 	if err != nil {
 		return err
 	}
@@ -74,10 +76,10 @@ func CreateHandler(cmd *cobra.Command, args []string) error {
 	spinner := progress.NewSpinner(status)
 	p.Add(status, spinner)
 
-	for _, c := range commands {
-		switch c.Name {
+	for i := range modelfile.Commands {
+		switch modelfile.Commands[i].Name {
 		case "model", "adapter":
-			path := c.Args
+			path := modelfile.Commands[i].Args
 			if path == "~" {
 				path = home
 			} else if strings.HasPrefix(path, "~/") {
@@ -89,101 +91,22 @@ func CreateHandler(cmd *cobra.Command, args []string) error {
 			}
 
 			fi, err := os.Stat(path)
-			if errors.Is(err, os.ErrNotExist) && c.Name == "model" {
+			if errors.Is(err, os.ErrNotExist) && modelfile.Commands[i].Name == "model" {
 				continue
 			} else if err != nil {
 				return err
 			}
 
-			// TODO make this work w/ adapters
 			if fi.IsDir() {
-				tf, err := os.CreateTemp("", "ollama-tf")
+				// this is likely a safetensors or pytorch directory
+				// TODO make this work w/ adapters
+				tempfile, err := tempZipFiles(path)
 				if err != nil {
 					return err
 				}
-				defer os.RemoveAll(tf.Name())
+				defer os.RemoveAll(tempfile)
 
-				zf := zip.NewWriter(tf)
-
-				files := []string{}
-
-				tfiles, err := filepath.Glob(filepath.Join(path, "pytorch_model-*.bin"))
-				if err != nil {
-					return err
-				} else if len(tfiles) == 0 {
-					tfiles, err = filepath.Glob(filepath.Join(path, "model-*.safetensors"))
-					if err != nil {
-						return err
-					}
-				}
-
-				files = append(files, tfiles...)
-
-				if len(files) == 0 {
-					return fmt.Errorf("no models were found in '%s'", path)
-				}
-
-				// add the safetensor/torch config file + tokenizer
-				files = append(files, filepath.Join(path, "config.json"))
-				files = append(files, filepath.Join(path, "params.json"))
-				files = append(files, filepath.Join(path, "added_tokens.json"))
-				files = append(files, filepath.Join(path, "tokenizer.model"))
-
-				for _, fn := range files {
-					f, err := os.Open(fn)
-
-					// just skip whatever files aren't there
-					if os.IsNotExist(err) {
-						if strings.HasSuffix(fn, "tokenizer.model") {
-							// try the parent dir before giving up
-							parentDir := filepath.Dir(path)
-							newFn := filepath.Join(parentDir, "tokenizer.model")
-							f, err = os.Open(newFn)
-							if os.IsNotExist(err) {
-								continue
-							} else if err != nil {
-								return err
-							}
-						} else {
-							continue
-						}
-					} else if err != nil {
-						return err
-					}
-
-					fi, err := f.Stat()
-					if err != nil {
-						return err
-					}
-
-					h, err := zip.FileInfoHeader(fi)
-					if err != nil {
-						return err
-					}
-
-					h.Name = filepath.Base(fn)
-					h.Method = zip.Store
-
-					w, err := zf.CreateHeader(h)
-					if err != nil {
-						return err
-					}
-
-					_, err = io.Copy(w, f)
-					if err != nil {
-						return err
-					}
-
-				}
-
-				if err := zf.Close(); err != nil {
-					return err
-				}
-
-				if err := tf.Close(); err != nil {
-					return err
-				}
-				path = tf.Name()
+				path = tempfile
 			}
 
 			digest, err := createBlob(cmd, client, path)
@@ -191,10 +114,11 @@ func CreateHandler(cmd *cobra.Command, args []string) error {
 				return err
 			}
 
-			modelfile = bytes.ReplaceAll(modelfile, []byte(c.Args), []byte("@"+digest))
+			modelfile.Commands[i].Args = "@" + digest
 		}
 	}
 
+	bars := make(map[string]*progress.Bar)
 	fn := func(resp api.ProgressResponse) error {
 		if resp.Digest != "" {
 			spinner.Stop()
@@ -220,7 +144,7 @@ func CreateHandler(cmd *cobra.Command, args []string) error {
 
 	quantization, _ := cmd.Flags().GetString("quantization")
 
-	request := api.CreateRequest{Name: args[0], Modelfile: string(modelfile), Quantization: quantization}
+	request := api.CreateRequest{Name: args[0], Modelfile: modelfile.String(), Quantization: quantization}
 	if err := client.Create(cmd.Context(), &request, fn); err != nil {
 		return err
 	}
@@ -228,6 +152,114 @@ func CreateHandler(cmd *cobra.Command, args []string) error {
 	return nil
 }
 
+func tempZipFiles(path string) (string, error) {
+	tempfile, err := os.CreateTemp("", "ollama-tf")
+	if err != nil {
+		return "", err
+	}
+	defer tempfile.Close()
+
+	zipfile := zip.NewWriter(tempfile)
+	defer zipfile.Close()
+
+	detectContentType := func(path string) (string, error) {
+		f, err := os.Open(path)
+		if err != nil {
+			return "", err
+		}
+		defer f.Close()
+
+		var b bytes.Buffer
+		b.Grow(512)
+
+		if _, err := io.CopyN(&b, f, 512); err != nil && !errors.Is(err, io.EOF) {
+			return "", err
+		}
+
+		contentType, _, _ := strings.Cut(http.DetectContentType(b.Bytes()), ";")
+		return contentType, nil
+	}
+
+	glob := func(pattern, contentType string) ([]string, error) {
+		matches, err := filepath.Glob(pattern)
+		if err != nil {
+			return nil, err
+		}
+
+		for _, safetensor := range matches {
+			if ct, err := detectContentType(safetensor); err != nil {
+				return nil, err
+			} else if ct != contentType {
+				return nil, fmt.Errorf("invalid content type: expected %s for %s", ct, safetensor)
+			}
+		}
+
+		return matches, nil
+	}
+
+	var files []string
+	if st, _ := glob(filepath.Join(path, "model*.safetensors"), "application/octet-stream"); len(st) > 0 {
+		// safetensors files might be unresolved git lfs references; skip if they are
+		// covers model-x-of-y.safetensors, model.fp32-x-of-y.safetensors, model.safetensors
+		files = append(files, st...)
+	} else if pt, _ := glob(filepath.Join(path, "pytorch_model*.bin"), "application/zip"); len(pt) > 0 {
+		// pytorch files might also be unresolved git lfs references; skip if they are
+		// covers pytorch_model-x-of-y.bin, pytorch_model.fp32-x-of-y.bin, pytorch_model.bin
+		files = append(files, pt...)
+	} else if pt, _ := glob(filepath.Join(path, "consolidated*.pth"), "application/octet-stream"); len(pt) > 0 {
+		// pytorch files might also be unresolved git lfs references; skip if they are
+		// covers consolidated.x.pth, consolidated.pth
+		files = append(files, pt...)
+	} else {
+		return "", errors.New("no safetensors or torch files found")
+	}
+
+	// add configuration files, json files are detected as text/plain
+	js, err := glob(filepath.Join(path, "*.json"), "text/plain")
+	if err != nil {
+		return "", err
+	}
+	files = append(files, js...)
+
+	if tks, _ := glob(filepath.Join(path, "tokenizer.model"), "application/octet-stream"); len(tks) > 0 {
+		// add tokenizer.model if it exists, tokenizer.json is automatically picked up by the previous glob
+		// tokenizer.model might be a unresolved git lfs reference; error if it is
+		files = append(files, tks...)
+	} else if tks, _ := glob(filepath.Join(path, "**/tokenizer.model"), "text/plain"); len(tks) > 0 {
+		// some times tokenizer.model is in a subdirectory (e.g. meta-llama/Meta-Llama-3-8B)
+		files = append(files, tks...)
+	}
+
+	for _, file := range files {
+		f, err := os.Open(file)
+		if err != nil {
+			return "", err
+		}
+		defer f.Close()
+
+		fi, err := f.Stat()
+		if err != nil {
+			return "", err
+		}
+
+		zfi, err := zip.FileInfoHeader(fi)
+		if err != nil {
+			return "", err
+		}
+
+		zf, err := zipfile.CreateHeader(zfi)
+		if err != nil {
+			return "", err
+		}
+
+		if _, err := io.Copy(zf, f); err != nil {
+			return "", err
+		}
+	}
+
+	return tempfile.Name(), nil
+}
+
 func createBlob(cmd *cobra.Command, client *api.Client, path string) (string, error) {
 	bin, err := os.Open(path)
 	if err != nil {
@@ -322,6 +354,47 @@ func RunHandler(cmd *cobra.Command, args []string) error {
 	return generateInteractive(cmd, opts)
 }
 
+func errFromUnknownKey(unknownKeyErr error) error {
+	// find SSH public key in the error message
+	sshKeyPattern := `ssh-\w+ [^\s"]+`
+	re := regexp.MustCompile(sshKeyPattern)
+	matches := re.FindStringSubmatch(unknownKeyErr.Error())
+
+	if len(matches) > 0 {
+		serverPubKey := matches[0]
+
+		localPubKey, err := auth.GetPublicKey()
+		if err != nil {
+			return unknownKeyErr
+		}
+
+		if runtime.GOOS == "linux" && serverPubKey != localPubKey {
+			// try the ollama service public key
+			svcPubKey, err := os.ReadFile("/usr/share/ollama/.ollama/id_ed25519.pub")
+			if err != nil {
+				return unknownKeyErr
+			}
+			localPubKey = strings.TrimSpace(string(svcPubKey))
+		}
+
+		// check if the returned public key matches the local public key, this prevents adding a remote key to the user's account
+		if serverPubKey != localPubKey {
+			return unknownKeyErr
+		}
+
+		var msg strings.Builder
+		msg.WriteString(unknownKeyErr.Error())
+		msg.WriteString("\n\nYour ollama key is:\n")
+		msg.WriteString(localPubKey)
+		msg.WriteString("\nAdd your key at:\n")
+		msg.WriteString("https://ollama.com/settings/keys")
+
+		return errors.New(msg.String())
+	}
+
+	return unknownKeyErr
+}
+
 func PushHandler(cmd *cobra.Command, args []string) error {
 	client, err := api.ClientFromEnvironment()
 	if err != nil {
@@ -369,6 +442,20 @@ func PushHandler(cmd *cobra.Command, args []string) error {
 
 	request := api.PushRequest{Name: args[0], Insecure: insecure}
 	if err := client.Push(cmd.Context(), &request, fn); err != nil {
+		if spinner != nil {
+			spinner.Stop()
+		}
+		if strings.Contains(err.Error(), "access denied") {
+			return errors.New("you are not authorized to push to this namespace, create the model under a namespace you own")
+		}
+		host := model.ParseName(args[0]).Host
+		isOllamaHost := strings.HasSuffix(host, ".ollama.ai") || strings.HasSuffix(host, ".ollama.com")
+		if strings.Contains(err.Error(), errtypes.UnknownOllamaKeyErrMsg) && isOllamaHost {
+			// the user has not added their ollama key to ollama.com
+			// re-throw an error with a more user-friendly message
+			return errFromUnknownKey(err)
+		}
+
 		return err
 	}
 
@@ -796,19 +883,17 @@ func generate(cmd *cobra.Command, opts runOptions) error {
 }
 
 func RunServer(cmd *cobra.Command, _ []string) error {
-	host, port, err := net.SplitHostPort(strings.Trim(os.Getenv("OLLAMA_HOST"), "\"'"))
+	// retrieve the OLLAMA_HOST environment variable
+	ollamaHost, err := api.GetOllamaHost()
 	if err != nil {
-		host, port = "127.0.0.1", "11434"
-		if ip := net.ParseIP(strings.Trim(os.Getenv("OLLAMA_HOST"), "[]")); ip != nil {
-			host = ip.String()
-		}
+		return err
 	}
 
 	if err := initializeKeypair(); err != nil {
 		return err
 	}
 
-	ln, err := net.Listen("tcp", net.JoinHostPort(host, port))
+	ln, err := net.Listen("tcp", net.JoinHostPort(ollamaHost.Host, ollamaHost.Port))
 	if err != nil {
 		return err
 	}
@@ -1034,7 +1119,7 @@ Environment Variables:
 		RunE:    ListHandler,
 	}
 	copyCmd := &cobra.Command{
-		Use:     "cp SOURCE TARGET",
+		Use:     "cp SOURCE DESTINATION",
 		Short:   "Copy a model",
 		Args:    cobra.ExactArgs(2),
 		PreRunE: checkServerHeartbeat,

cmd/interactive.go

@@ -94,6 +94,7 @@ func generateInteractive(cmd *cobra.Command, opts runOptions) error {
 		fmt.Fprintln(os.Stderr, "  /show           Show model information")
 		fmt.Fprintln(os.Stderr, "  /load <model>   Load a session or model")
 		fmt.Fprintln(os.Stderr, "  /save <model>   Save your current session")
+		fmt.Fprintln(os.Stderr, "  /clear          Clear session context")
 		fmt.Fprintln(os.Stderr, "  /bye            Exit")
 		fmt.Fprintln(os.Stderr, "  /?, /help       Help for a command")
 		fmt.Fprintln(os.Stderr, "  /? shortcuts    Help for keyboard shortcuts")
@@ -280,6 +281,10 @@ func generateInteractive(cmd *cobra.Command, opts runOptions) error {
 			}
 			fmt.Printf("Created new model '%s'\n", args[1])
 			continue
+		case strings.HasPrefix(line, "/clear"):
+			opts.Messages = []api.Message{}
+			fmt.Println("Cleared session context")
+			continue
 		case strings.HasPrefix(line, "/set"):
 			args := strings.Fields(line)
 			if len(args) > 1 {

convert/convert.go

@@ -5,6 +5,7 @@ import (
 	"encoding/binary"
 	"encoding/json"
 	"fmt"
+	"io"
 	"log/slog"
 	"os"
 	"path/filepath"
@@ -18,19 +19,23 @@ import (
 )
 
 type Params struct {
-	Architectures    []string `json:"architectures"`
-	VocabSize        int      `json:"vocab_size"`
-	HiddenSize       int      `json:"hidden_size"`       // n_embd
-	HiddenLayers     int      `json:"num_hidden_layers"` // n_layer
-	ContextSize      int      `json:"max_position_embeddings"`
-	IntermediateSize int      `json:"intermediate_size"`
-	AttentionHeads   int      `json:"num_attention_heads"` // n_head
-	KeyValHeads      int      `json:"num_key_value_heads"`
-	NormEPS          float64  `json:"rms_norm_eps"`
-	BoSTokenID       int      `json:"bos_token_id"`
-	EoSTokenID       int      `json:"eos_token_id"`
-	HeadDimension    int      `json:"head_dim"`
-	PaddingTokenID   int      `json:"pad_token_id"`
+	Architectures     []string `json:"architectures"`
+	VocabSize         int      `json:"vocab_size"`
+	HiddenSize        int      `json:"hidden_size"`       // n_embd
+	HiddenLayers      int      `json:"num_hidden_layers"` // n_layer
+	ContextSize       int      `json:"max_position_embeddings"`
+	IntermediateSize  int      `json:"intermediate_size"`
+	AttentionHeads    int      `json:"num_attention_heads"` // n_head
+	KeyValHeads       int      `json:"num_key_value_heads"`
+	NormEPS           float64  `json:"rms_norm_eps"`
+	BoSTokenID        int      `json:"bos_token_id"`
+	EoSTokenID        int      `json:"eos_token_id"`
+	HeadDimension     int      `json:"head_dim"`
+	PaddingTokenID    int      `json:"pad_token_id"`
+	RopeFrequencyBase float64  `json:"rope_theta"`
+
+	Experts     int `json:"num_local_experts"`
+	ExpertsUsed int `json:"num_experts_per_tok"`
 
 	ByteOrder
 }
@@ -43,7 +48,7 @@ type ByteOrder interface {
 type ModelArch interface {
 	GetTensors() error
 	LoadVocab() error
-	WriteGGUF() (string, error)
+	WriteGGUF(io.WriteSeeker) error
 }
 
 type ModelFormat interface {

convert/gemma.go

@@ -94,7 +94,7 @@ func (m *GemmaModel) LoadVocab() error {
 	return nil
 }
 
-func (m *GemmaModel) WriteGGUF() (string, error) {
+func (m *GemmaModel) WriteGGUF(ws io.WriteSeeker) error {
 	kv := llm.KV{
 		"general.architecture":                   "gemma",
 		"general.name":                           m.Name,
@@ -122,16 +122,5 @@ func (m *GemmaModel) WriteGGUF() (string, error) {
 		"tokenizer.ggml.add_eos_token":    false,
 	}
 
-	f, err := os.CreateTemp("", "ollama-gguf")
-	if err != nil {
-		return "", err
-	}
-	defer f.Close()
-
-	mod := llm.NewGGUFV3(m.Params.ByteOrder)
-	if err := mod.Encode(f, kv, m.Tensors); err != nil {
-		return "", err
-	}
-
-	return f.Name(), nil
+	return llm.NewGGUFV3(m.Params.ByteOrder).Encode(ws, kv, m.Tensors)
 }

convert/llama.go

@@ -132,7 +132,7 @@ func (m *LlamaModel) LoadVocab() error {
 	return nil
 }
 
-func (m *LlamaModel) WriteGGUF() (string, error) {
+func (m *LlamaModel) WriteGGUF(ws io.WriteSeeker) error {
 	kv := llm.KV{
 		"general.architecture":                   "llama",
 		"general.name":                           m.Name,
@@ -161,16 +161,9 @@ func (m *LlamaModel) WriteGGUF() (string, error) {
 
 	f, err := os.CreateTemp("", "ollama-gguf")
 	if err != nil {
-		return "", err
+		return err
 	}
 	defer f.Close()
 
-	mod := llm.NewGGUFV3(m.Params.ByteOrder)
-	if err := mod.Encode(f, kv, m.Tensors); err != nil {
-		return "", err
-	}
-
-	slog.Debug(fmt.Sprintf("gguf file = %s", f.Name()))
-
-	return f.Name(), nil
+	return llm.NewGGUFV3(m.Params.ByteOrder).Encode(f, kv, m.Tensors)
 }

convert/mistral.go

@@ -132,7 +132,7 @@ func (m *MistralModel) LoadVocab() error {
 	return nil
 }
 
-func (m *MistralModel) WriteGGUF() (string, error) {
+func (m *MistralModel) WriteGGUF(ws io.WriteSeeker) error {
 	kv := llm.KV{
 		"general.architecture":                   "llama",
 		"general.name":                           m.Name,
@@ -158,16 +158,5 @@ func (m *MistralModel) WriteGGUF() (string, error) {
 		"tokenizer.ggml.unknown_token_id": uint32(0),
 	}
 
-	f, err := os.CreateTemp("", "ollama-gguf")
-	if err != nil {
-		return "", err
-	}
-	defer f.Close()
-
-	mod := llm.NewGGUFV3(m.Params.ByteOrder)
-	if err := mod.Encode(f, kv, m.Tensors); err != nil {
-		return "", err
-	}
-
-	return f.Name(), nil
+	return llm.NewGGUFV3(m.Params.ByteOrder).Encode(ws, kv, m.Tensors)
 }

convert/mixtral.go

@@ -0,0 +1,85 @@
+package convert
+
+import (
+	"io"
+	"regexp"
+
+	"github.com/ollama/ollama/llm"
+)
+
+type MixtralModel struct {
+	ModelData
+}
+
+func (m *MixtralModel) GetTensors() error {
+	t, err := m.Format.GetTensors(m.Path, m.Params)
+	if err != nil {
+		return err
+	}
+
+	m.Tensors = []llm.Tensor{}
+
+	pattern := `^blk\.[0-9]+\.attn_(?P<layer>q|k)\.weight$`
+	re, err := regexp.Compile(pattern)
+	if err != nil {
+		return err
+	}
+
+	for _, l := range t {
+		matches := re.FindAllStringSubmatch(l.Name, -1)
+		if len(matches) > 0 {
+			wt := l.WriterTo.(safetensorWriterTo)
+			wt.handler = mistralLayerHandler
+			l.WriterTo = wt
+		}
+		m.Tensors = append(m.Tensors, l)
+	}
+
+	return nil
+}
+
+func (m *MixtralModel) LoadVocab() error {
+	v, err := LoadSentencePieceTokens(m.Path, m.Params)
+	if err != nil {
+		return err
+	}
+	m.Vocab = v
+	return nil
+}
+
+func (m *MixtralModel) WriteGGUF(ws io.WriteSeeker) error {
+	kv := llm.KV{
+		"general.architecture":          "llama",
+		"general.name":                  m.Name,
+		"llama.block_count":             uint32(m.Params.HiddenLayers),
+		"llama.context_length":          uint32(m.Params.ContextSize),
+		"llama.embedding_length":        uint32(m.Params.HiddenSize),
+		"llama.feed_forward_length":     uint32(m.Params.IntermediateSize),
+		"llama.attention.head_count":    uint32(m.Params.AttentionHeads),
+		"llama.attention.head_count_kv": uint32(m.Params.KeyValHeads),
+
+		"llama.rope.freq_base":                   float32(m.Params.RopeFrequencyBase),
+		"llama.attention.layer_norm_rms_epsilon": float32(m.Params.NormEPS),
+
+		"llama.expert_count":      uint32(m.Params.Experts),
+		"llama.expert_used_count": uint32(m.Params.ExpertsUsed),
+
+		"llama.vocab_size":           uint32(len(m.Vocab.Tokens)),
+		"llama.rope.dimension_count": uint32(m.Params.HiddenSize / m.Params.AttentionHeads),
+
+		"general.file_type":    uint32(1),
+		"tokenizer.ggml.model": "llama",
+
+		"tokenizer.ggml.tokens":     m.Vocab.Tokens,
+		"tokenizer.ggml.scores":     m.Vocab.Scores,
+		"tokenizer.ggml.token_type": m.Vocab.Types,
+
+		"tokenizer.ggml.bos_token_id":     uint32(m.Params.BoSTokenID),
+		"tokenizer.ggml.eos_token_id":     uint32(m.Params.EoSTokenID),
+		"tokenizer.ggml.unknown_token_id": uint32(0),
+		"tokenizer.ggml.add_bos_token":    true,
+		"tokenizer.ggml.add_eos_token":    false,
+	}
+
+	return llm.NewGGUFV3(m.Params.ByteOrder).Encode(ws, kv, m.Tensors)
+}

convert/safetensors.go

@@ -93,7 +93,6 @@ func (m *SafetensorFormat) readTensors(fn string, offset uint64, params *Params)
 	}
 
 	slices.Sort(keys)
-
 	slog.Info("converting layers")
 
 	var tensors []llm.Tensor
@@ -105,7 +104,6 @@ func (m *SafetensorFormat) readTensors(fn string, offset uint64, params *Params)
 			return nil, 0, err
 		}
 
-		slog.Debug(fmt.Sprintf("metadata = %#v", data))
 		var size uint64
 		var kind uint32
 		switch len(data.Shape) {
@@ -150,11 +148,13 @@ func (m *SafetensorFormat) readTensors(fn string, offset uint64, params *Params)
 			padding:  8 + jsonSize,
 		}
 
-		tensors = append(tensors, t)
 		offset += size
+		tensors = append(tensors, t)
 	}
+
 	slog.Debug(fmt.Sprintf("total tensors for file = %d", len(tensors)))
 	slog.Debug(fmt.Sprintf("offset = %d", offset))
+
 	return tensors, offset, nil
 }
 
@@ -185,15 +185,19 @@ func (m *SafetensorFormat) GetLayerName(n string) (string, error) {
 	}
 
 	tMap := map[string]string{
-		"model.layers.(\\d+).input_layernorm.weight":          "blk.$1.attn_norm.weight",
-		"model.layers.(\\d+).mlp.down_proj.weight":            "blk.$1.ffn_down.weight",
-		"model.layers.(\\d+).mlp.gate_proj.weight":            "blk.$1.ffn_gate.weight",
-		"model.layers.(\\d+).mlp.up_proj.weight":              "blk.$1.ffn_up.weight",
-		"model.layers.(\\d+).post_attention_layernorm.weight": "blk.$1.ffn_norm.weight",
-		"model.layers.(\\d+).self_attn.k_proj.weight":         "blk.$1.attn_k.weight",
-		"model.layers.(\\d+).self_attn.o_proj.weight":         "blk.$1.attn_output.weight",
-		"model.layers.(\\d+).self_attn.q_proj.weight":         "blk.$1.attn_q.weight",
-		"model.layers.(\\d+).self_attn.v_proj.weight":         "blk.$1.attn_v.weight",
+		"model.layers.(\\d+).input_layernorm.weight":                    "blk.$1.attn_norm.weight",
+		"model.layers.(\\d+).mlp.down_proj.weight":                      "blk.$1.ffn_down.weight",
+		"model.layers.(\\d+).mlp.gate_proj.weight":                      "blk.$1.ffn_gate.weight",
+		"model.layers.(\\d+).mlp.up_proj.weight":                        "blk.$1.ffn_up.weight",
+		"model.layers.(\\d+).post_attention_layernorm.weight":           "blk.$1.ffn_norm.weight",
+		"model.layers.(\\d+).self_attn.k_proj.weight":                   "blk.$1.attn_k.weight",
+		"model.layers.(\\d+).self_attn.o_proj.weight":                   "blk.$1.attn_output.weight",
+		"model.layers.(\\d+).self_attn.q_proj.weight":                   "blk.$1.attn_q.weight",
+		"model.layers.(\\d+).self_attn.v_proj.weight":                   "blk.$1.attn_v.weight",
+		"model.layers.(\\d+).block_sparse_moe.gate.weight":              "blk.$1.ffn_gate_inp.weight",
+		"model.layers.(\\d+).block_sparse_moe.experts.(\\d+).w1.weight": "blk.$1.ffn_gate.$2.weight",
+		"model.layers.(\\d+).block_sparse_moe.experts.(\\d+).w2.weight": "blk.$1.ffn_down.$2.weight",
+		"model.layers.(\\d+).block_sparse_moe.experts.(\\d+).w3.weight": "blk.$1.ffn_up.$2.weight",
 	}
 
 	v, ok := directMap[n]
@@ -286,6 +290,15 @@ func (m *SafetensorFormat) GetModelArch(name, dirPath string, params *Params) (M
 					Format: m,
 				},
 			}, nil
+		case "MixtralForCausalLM":
+			return &MixtralModel{
+				ModelData{
+					Name:   name,
+					Path:   dirPath,
+					Params: params,
+					Format: m,
+				},
+			}, nil
 		case "GemmaForCausalLM":
 			return &GemmaModel{
 				ModelData{

docs/api.md

@@ -17,7 +17,7 @@
 
 ### Model names
 
-Model names follow a `model:tag` format, where `model` can have an optional namespace such as `example/model`. Some examples are `orca-mini:3b-q4_1` and `llama2:70b`. The tag is optional and, if not provided, will default to `latest`. The tag is used to identify a specific version.
+Model names follow a `model:tag` format, where `model` can have an optional namespace such as `example/model`. Some examples are `orca-mini:3b-q4_1` and `llama3:70b`. The tag is optional and, if not provided, will default to `latest`. The tag is used to identify a specific version.
 
 ### Durations
 
@@ -66,7 +66,7 @@ Enable JSON mode by setting the `format` parameter to `json`. This will structur
 
 ```shell
 curl http://localhost:11434/api/generate -d '{
-  "model": "llama2",
+  "model": "llama3",
   "prompt": "Why is the sky blue?"
 }'
 ```
@@ -77,7 +77,7 @@ A stream of JSON objects is returned:
 
 ```json
 {
-  "model": "llama2",
+  "model": "llama3",
   "created_at": "2023-08-04T08:52:19.385406455-07:00",
   "response": "The",
   "done": false
@@ -99,7 +99,7 @@ To calculate how fast the response is generated in tokens per second (token/s),
 
 ```json
 {
-  "model": "llama2",
+  "model": "llama3",
   "created_at": "2023-08-04T19:22:45.499127Z",
   "response": "",
   "done": true,
@@ -121,7 +121,7 @@ A response can be received in one reply when streaming is off.
 
 ```shell
 curl http://localhost:11434/api/generate -d '{
-  "model": "llama2",
+  "model": "llama3",
   "prompt": "Why is the sky blue?",
   "stream": false
 }'
@@ -133,7 +133,7 @@ If `stream` is set to `false`, the response will be a single JSON object:
 
 ```json
 {
-  "model": "llama2",
+  "model": "llama3",
   "created_at": "2023-08-04T19:22:45.499127Z",
   "response": "The sky is blue because it is the color of the sky.",
   "done": true,
@@ -155,7 +155,7 @@ If `stream` is set to `false`, the response will be a single JSON object:
 
 ```shell
 curl http://localhost:11434/api/generate -d '{
-  "model": "llama2",
+  "model": "llama3",
   "prompt": "What color is the sky at different times of the day? Respond using JSON",
   "format": "json",
   "stream": false
@@ -166,7 +166,7 @@ curl http://localhost:11434/api/generate -d '{
 
 ```json
 {
-  "model": "llama2",
+  "model": "llama3",
   "created_at": "2023-11-09T21:07:55.186497Z",
   "response": "{\n\"morning\": {\n\"color\": \"blue\"\n},\n\"noon\": {\n\"color\": \"blue-gray\"\n},\n\"afternoon\": {\n\"color\": \"warm gray\"\n},\n\"evening\": {\n\"color\": \"orange\"\n}\n}\n",
   "done": true,
@@ -289,7 +289,7 @@ If you want to set custom options for the model at runtime rather than in the Mo
 
 ```shell
 curl http://localhost:11434/api/generate -d '{
-  "model": "llama2",
+  "model": "llama3",
   "prompt": "Why is the sky blue?",
   "stream": false,
   "options": {
@@ -332,7 +332,7 @@ curl http://localhost:11434/api/generate -d '{
 
 ```json
 {
-  "model": "llama2",
+  "model": "llama3",
   "created_at": "2023-08-04T19:22:45.499127Z",
   "response": "The sky is blue because it is the color of the sky.",
   "done": true,
@@ -354,7 +354,7 @@ If an empty prompt is provided, the model will be loaded into memory.
 
 ```shell
 curl http://localhost:11434/api/generate -d '{
-  "model": "llama2"
+  "model": "llama3"
 }'
 ```
 
@@ -364,7 +364,7 @@ A single JSON object is returned:
 
 ```json
 {
-  "model": "llama2",
+  "model": "llama3",
   "created_at": "2023-12-18T19:52:07.071755Z",
   "response": "",
   "done": true
@@ -407,7 +407,7 @@ Send a chat message with a streaming response.
 
 ```shell
 curl http://localhost:11434/api/chat -d '{
-  "model": "llama2",
+  "model": "llama3",
   "messages": [
     {
       "role": "user",
@@ -423,7 +423,7 @@ A stream of JSON objects is returned:
 
 ```json
 {
-  "model": "llama2",
+  "model": "llama3",
   "created_at": "2023-08-04T08:52:19.385406455-07:00",
   "message": {
     "role": "assistant",
@@ -438,7 +438,7 @@ Final response:
 
 ```json
 {
-  "model": "llama2",
+  "model": "llama3",
   "created_at": "2023-08-04T19:22:45.499127Z",
   "done": true,
   "total_duration": 4883583458,
@@ -456,7 +456,7 @@ Final response:
 
 ```shell
 curl http://localhost:11434/api/chat -d '{
-  "model": "llama2",
+  "model": "llama3",
   "messages": [
     {
       "role": "user",
@@ -471,7 +471,7 @@ curl http://localhost:11434/api/chat -d '{
 
 ```json
 {
-  "model": "registry.ollama.ai/library/llama2:latest",
+  "model": "registry.ollama.ai/library/llama3:latest",
   "created_at": "2023-12-12T14:13:43.416799Z",
   "message": {
     "role": "assistant",
@@ -495,7 +495,7 @@ Send a chat message with a conversation history. You can use this same approach
 
 ```shell
 curl http://localhost:11434/api/chat -d '{
-  "model": "llama2",
+  "model": "llama3",
   "messages": [
     {
       "role": "user",
@@ -519,7 +519,7 @@ A stream of JSON objects is returned:
 
 ```json
 {
-  "model": "llama2",
+  "model": "llama3",
   "created_at": "2023-08-04T08:52:19.385406455-07:00",
   "message": {
     "role": "assistant",
@@ -533,7 +533,7 @@ Final response:
 
 ```json
 {
-  "model": "llama2",
+  "model": "llama3",
   "created_at": "2023-08-04T19:22:45.499127Z",
   "done": true,
   "total_duration": 8113331500,
@@ -591,7 +591,7 @@ curl http://localhost:11434/api/chat -d '{
 
 ```shell
 curl http://localhost:11434/api/chat -d '{
-  "model": "llama2",
+  "model": "llama3",
   "messages": [
     {
       "role": "user",
@@ -609,7 +609,7 @@ curl http://localhost:11434/api/chat -d '{
 
 ```json
 {
-  "model": "registry.ollama.ai/library/llama2:latest",
+  "model": "registry.ollama.ai/library/llama3:latest",
   "created_at": "2023-12-12T14:13:43.416799Z",
   "message": {
     "role": "assistant",
@@ -651,7 +651,7 @@ Create a new model from a `Modelfile`.
 ```shell
 curl http://localhost:11434/api/create -d '{
   "name": "mario",
-  "modelfile": "FROM llama2\nSYSTEM You are mario from Super Mario Bros."
+  "modelfile": "FROM llama3\nSYSTEM You are mario from Super Mario Bros."
 }'
 ```
 
@@ -758,7 +758,7 @@ A single JSON object will be returned.
       }
     },
     {
-      "name": "llama2:latest",
+      "name": "llama3:latest",
       "modified_at": "2023-12-07T09:32:18.757212583-08:00",
       "size": 3825819519,
       "digest": "fe938a131f40e6f6d40083c9f0f430a515233eb2edaa6d72eb85c50d64f2300e",
@@ -792,7 +792,7 @@ Show information about a model including details, modelfile, template, parameter
 
 ```shell
 curl http://localhost:11434/api/show -d '{
-  "name": "llama2"
+  "name": "llama3"
 }'
 ```
 
@@ -827,8 +827,8 @@ Copy a model. Creates a model with another name from an existing model.
 
 ```shell
 curl http://localhost:11434/api/copy -d '{
-  "source": "llama2",
-  "destination": "llama2-backup"
+  "source": "llama3",
+  "destination": "llama3-backup"
 }'
 ```
 
@@ -854,7 +854,7 @@ Delete a model and its data.
 
 ```shell
 curl -X DELETE http://localhost:11434/api/delete -d '{
-  "name": "llama2:13b"
+  "name": "llama3:13b"
 }'
 ```
 
@@ -882,7 +882,7 @@ Download a model from the ollama library. Cancelled pulls are resumed from where
 
 ```shell
 curl http://localhost:11434/api/pull -d '{
-  "name": "llama2"
+  "name": "llama3"
 }'
 ```
 

docs/development.md

@@ -51,7 +51,7 @@ Typically the build scripts will auto-detect CUDA, however, if your Linux distro
 or installation approach uses unusual paths, you can specify the location by
 specifying an environment variable `CUDA_LIB_DIR` to the location of the shared
 libraries, and `CUDACXX` to the location of the nvcc compiler. You can customize
-set set of target CUDA architectues by setting `CMAKE_CUDA_ARCHITECTURES` (e.g. "50;60;70")
+a set of target CUDA architectures by setting `CMAKE_CUDA_ARCHITECTURES` (e.g. "50;60;70")
 
 Then generate dependencies:
 
@@ -142,4 +142,4 @@ In addition to the common Windows development tools described above, install AMD
 - [AMD HIP](https://www.amd.com/en/developer/resources/rocm-hub/hip-sdk.html)
 - [Strawberry Perl](https://strawberryperl.com/)
 
-Lastly, add `ninja.exe` included with MSVC to the system path (e.g. `C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\Common7\IDE\CommonExtensions\Microsoft\CMake\Ninja`).
+Lastly, add `ninja.exe` included with MSVC to the system path (e.g. `C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\Common7\IDE\CommonExtensions\Microsoft\CMake\Ninja`).

docs/faq.md

@@ -32,7 +32,7 @@ When using the API, specify the `num_ctx` parameter:
 
 ```
 curl http://localhost:11434/api/generate -d '{
-  "model": "llama2",
+  "model": "llama3",
   "prompt": "Why is the sky blue?",
   "options": {
     "num_ctx": 4096
@@ -88,9 +88,9 @@ On windows, Ollama inherits your user and system environment variables.
 
 3. Edit or create New variable(s) for your user account for `OLLAMA_HOST`, `OLLAMA_MODELS`, etc.
 
-4. Click OK/Apply to save 
+4. Click OK/Apply to save
 
-5. Run `ollama` from a new terminal window 
+5. Run `ollama` from a new terminal window
 
 
 ## How can I expose Ollama on my network?
@@ -221,10 +221,14 @@ The `keep_alive` parameter can be set to:
 
 For example, to preload a model and leave it in memory use:
 ```shell
-curl http://localhost:11434/api/generate -d '{"model": "llama2", "keep_alive": -1}'
+curl http://localhost:11434/api/generate -d '{"model": "llama3", "keep_alive": -1}'
 ```
 
 To unload the model and free up memory use:
 ```shell
-curl http://localhost:11434/api/generate -d '{"model": "llama2", "keep_alive": 0}'
+curl http://localhost:11434/api/generate -d '{"model": "llama3", "keep_alive": 0}'
 ```
+
+Alternatively, you can change the amount of time all models are loaded into memory by setting the `OLLAMA_KEEP_ALIVE` environment variable when starting the Ollama server. The `OLLAMA_KEEP_ALIVE` variable uses the same parameter types as the `keep_alive` parameter types mentioned above. Refer to section explaining [how to configure the Ollama server](#how-do-i-configure-ollama-server) to correctly set the environment variable.
+
+If you wish to override the `OLLAMA_KEEP_ALIVE` setting, use the `keep_alive` API parameter with the `/api/generate` or `/api/chat` API endpoints.

docs/modelfile.md

@@ -10,7 +10,7 @@ A model file is the blueprint to create and share models with Ollama.
 - [Examples](#examples)
 - [Instructions](#instructions)
   - [FROM (Required)](#from-required)
-    - [Build from llama2](#build-from-llama2)
+    - [Build from llama3](#build-from-llama3)
     - [Build from a bin file](#build-from-a-bin-file)
   - [PARAMETER](#parameter)
     - [Valid Parameters and Values](#valid-parameters-and-values)
@@ -48,7 +48,7 @@ INSTRUCTION arguments
 An example of a `Modelfile` creating a mario blueprint:
 
 ```modelfile
-FROM llama2
+FROM llama3
 # sets the temperature to 1 [higher is more creative, lower is more coherent]
 PARAMETER temperature 1
 # sets the context window size to 4096, this controls how many tokens the LLM can use as context to generate the next token
@@ -67,33 +67,25 @@ To use this:
 
 More examples are available in the [examples directory](../examples).
 
-### `Modelfile`s in [ollama.com/library][1]
-
-There are two ways to view `Modelfile`s underlying the models in [ollama.com/library][1]:
-
-- Option 1: view a details page from a model's tags page:
-  1.  Go to a particular model's tags (e.g. https://ollama.com/library/llama2/tags)
-  2.  Click on a tag (e.g. https://ollama.com/library/llama2:13b)
-  3.  Scroll down to "Layers"
-      - Note: if the [`FROM` instruction](#from-required) is not present,
-        it means the model was created from a local file
-- Option 2: use `ollama show` to print the `Modelfile` for any local models like so:
+To view the Modelfile of a given model, use the `ollama show --modelfile` command.
 
   ```bash
-  > ollama show --modelfile llama2:13b
+  > ollama show --modelfile llama3
   # Modelfile generated by "ollama show"
   # To build a new Modelfile based on this one, replace the FROM line with:
-  # FROM llama2:13b
+  # FROM llama3:latest
+  FROM /Users/pdevine/.ollama/models/blobs/sha256-00e1317cbf74d901080d7100f57580ba8dd8de57203072dc6f668324ba545f29
+  TEMPLATE """{{ if .System }}<|start_header_id|>system<|end_header_id|>
 
-  FROM /root/.ollama/models/blobs/sha256:123abc
-  TEMPLATE """[INST] {{ if .System }}<<SYS>>{{ .System }}<</SYS>>
+  {{ .System }}<|eot_id|>{{ end }}{{ if .Prompt }}<|start_header_id|>user<|end_header_id|>
 
-  {{ end }}{{ .Prompt }} [/INST] """
-  SYSTEM """"""
-  PARAMETER stop [INST]
-  PARAMETER stop [/INST]
-  PARAMETER stop <<SYS>>
-  PARAMETER stop <</SYS>>
+  {{ .Prompt }}<|eot_id|>{{ end }}<|start_header_id|>assistant<|end_header_id|>
+
+  {{ .Response }}<|eot_id|>"""
+  PARAMETER stop "<|start_header_id|>"
+  PARAMETER stop "<|end_header_id|>"
+  PARAMETER stop "<|eot_id|>"
+  PARAMETER stop "<|reserved_special_token"
   ```
 
 ## Instructions
@@ -106,10 +98,10 @@ The `FROM` instruction defines the base model to use when creating a model.
 FROM <model name>:<tag>
 ```
 
-#### Build from llama2
+#### Build from llama3
 
 ```modelfile
-FROM llama2
+FROM llama3
 ```
 
 A list of available base models:

docs/openai.md

@@ -25,7 +25,7 @@ chat_completion = client.chat.completions.create(
             'content': 'Say this is a test',
         }
     ],
-    model='llama2',
+    model='llama3',
 )
 ```
 
@@ -43,7 +43,7 @@ const openai = new OpenAI({
 
 const chatCompletion = await openai.chat.completions.create({
   messages: [{ role: 'user', content: 'Say this is a test' }],
-  model: 'llama2',
+  model: 'llama3',
 })
 ```
 
@@ -53,7 +53,7 @@ const chatCompletion = await openai.chat.completions.create({
 curl http://localhost:11434/v1/chat/completions \
     -H "Content-Type: application/json" \
     -d '{
-        "model": "llama2",
+        "model": "llama3",
         "messages": [
             {
                 "role": "system",
@@ -113,7 +113,7 @@ curl http://localhost:11434/v1/chat/completions \
 Before using a model, pull it locally `ollama pull`:
 
 ```shell
-ollama pull llama2
+ollama pull llama3
 ```
 
 ### Default model names
@@ -121,7 +121,7 @@ ollama pull llama2
 For tooling that relies on default OpenAI model names such as `gpt-3.5-turbo`, use `ollama cp` to copy an existing model name to a temporary name:
 
 ```
-ollama cp llama2 gpt-3.5-turbo
+ollama cp llama3 gpt-3.5-turbo
 ```
 
 Afterwards, this new model name can be specified the `model` field:

docs/tutorials/langchainjs.md

@@ -15,7 +15,7 @@ import { Ollama } from "langchain/llms/ollama";
 
 const ollama = new Ollama({
   baseUrl: "http://localhost:11434",
-  model: "llama2",
+  model: "llama3",
 });
 
 const answer = await ollama.invoke(`why is the sky blue?`);
@@ -23,10 +23,10 @@ const answer = await ollama.invoke(`why is the sky blue?`);
 console.log(answer);
 ```
 
-That will get us the same thing as if we ran `ollama run llama2 "why is the sky blue"` in the terminal. But we want to load a document from the web to ask a question against. **Cheerio** is a great library for ingesting a webpage, and **LangChain** uses it in their **CheerioWebBaseLoader**. So let's install **Cheerio** and build that part of the app.
+That will get us the same thing as if we ran `ollama run llama3 "why is the sky blue"` in the terminal. But we want to load a document from the web to ask a question against. **Cheerio** is a great library for ingesting a webpage, and **LangChain** uses it in their **CheerioWebBaseLoader**. So let's install **Cheerio** and build that part of the app.
 
 ```bash
-npm install cheerio 
+npm install cheerio
 ```
 
 ```javascript

docs/tutorials/langchainpy.md

@@ -17,10 +17,12 @@ Let's start by asking a simple question that we can get an answer to from the **
 Then we can create a model and ask the question:
 
 ```python
-from langchain.llms import Ollama
-ollama = Ollama(base_url='http://localhost:11434',
-model="llama2")
-print(ollama("why is the sky blue"))
+from langchain_community.llms import Ollama
+ollama = Ollama(
+    base_url='http://localhost:11434',
+    model="llama3"
+)
+print(ollama.invoke("why is the sky blue"))
 ```
 
 Notice that we are defining the model and the base URL for Ollama.

docs/windows.md

@@ -1,47 +1,48 @@
-# Ollama Windows Preview
-
-Welcome to the Ollama Windows preview.
-
-No more WSL required!
-
-Ollama now runs as a native Windows application, including NVIDIA and AMD Radeon GPU support.
-After installing Ollama Windows Preview, Ollama will run in the background and
-the `ollama` command line is available in `cmd`, `powershell` or your favorite
-terminal application. As usual the Ollama [api](./api.md) will be served on
-`http://localhost:11434`.
-
-As this is a preview release, you should expect a few bugs here and there.  If
-you run into a problem you can reach out on
-[Discord](https://discord.gg/ollama), or file an 
-[issue](https://github.com/ollama/ollama/issues).
-Logs will often be helpful in dianosing the problem (see
-[Troubleshooting](#troubleshooting) below)
-
-## System Requirements
-
-* Windows 10 or newer, Home or Pro
-* NVIDIA 452.39 or newer Drivers if you have an NVIDIA card
-* AMD Radeon Driver https://www.amd.com/en/support if you have a Radeon card
-
-## API Access
-
-Here's a quick example showing API access from `powershell`
-```powershell
-(Invoke-WebRequest -method POST -Body '{"model":"llama2", "prompt":"Why is the sky blue?", "stream": false}' -uri http://localhost:11434/api/generate ).Content | ConvertFrom-json
-```
-
-## Troubleshooting
-
-While we're in preview, `OLLAMA_DEBUG` is always enabled, which adds
-a "view logs" menu item to the app, and increses logging for the GUI app and
-server.
-
-Ollama on Windows stores files in a few different locations.  You can view them in
-the explorer window by hitting `<cmd>+R` and type in:
-- `explorer %LOCALAPPDATA%\Ollama` contains logs, and downloaded updates
-    - *app.log* contains logs from the GUI application
-    - *server.log* contains the server logs
-    - *upgrade.log* contains log output for upgrades
-- `explorer %LOCALAPPDATA%\Programs\Ollama` contains the binaries (The installer adds this to your user PATH)
-- `explorer %HOMEPATH%\.ollama` contains models and configuration
-- `explorer %TEMP%` contains temporary executable files in one or more `ollama*` directories
+<<<<<<< HEAD
+# Ollama Windows Preview
+
+Welcome to the Ollama Windows preview.
+
+No more WSL required!
+
+Ollama now runs as a native Windows application, including NVIDIA and AMD Radeon GPU support.
+After installing Ollama Windows Preview, Ollama will run in the background and
+the `ollama` command line is available in `cmd`, `powershell` or your favorite
+terminal application. As usual the Ollama [api](./api.md) will be served on
+`http://localhost:11434`.
+
+As this is a preview release, you should expect a few bugs here and there.  If
+you run into a problem you can reach out on
+[Discord](https://discord.gg/ollama), or file an
+[issue](https://github.com/ollama/ollama/issues).
+Logs will often be helpful in diagnosing the problem (see
+[Troubleshooting](#troubleshooting) below)
+
+## System Requirements
+
+* Windows 10 or newer, Home or Pro
+* NVIDIA 452.39 or newer Drivers if you have an NVIDIA card
+* AMD Radeon Driver https://www.amd.com/en/support if you have a Radeon card
+
+## API Access
+
+Here's a quick example showing API access from `powershell`
+```powershell
+(Invoke-WebRequest -method POST -Body '{"model":"llama3", "prompt":"Why is the sky blue?", "stream": false}' -uri http://localhost:11434/api/generate ).Content | ConvertFrom-json
+```
+
+## Troubleshooting
+
+While we're in preview, `OLLAMA_DEBUG` is always enabled, which adds
+a "view logs" menu item to the app, and increses logging for the GUI app and
+server.
+
+Ollama on Windows stores files in a few different locations.  You can view them in
+the explorer window by hitting `<cmd>+R` and type in:
+- `explorer %LOCALAPPDATA%\Ollama` contains logs, and downloaded updates
+    - *app.log* contains logs from the GUI application
+    - *server.log* contains the server logs
+    - *upgrade.log* contains log output for upgrades
+- `explorer %LOCALAPPDATA%\Programs\Ollama` contains the binaries (The installer adds this to your user PATH)
+- `explorer %HOMEPATH%\.ollama` contains models and configuration
+- `explorer %TEMP%` contains temporary executable files in one or more `ollama*` directories

examples/bash-comparemodels/README.md

@@ -2,7 +2,7 @@
 
 When calling `ollama`, you can pass it a file to run all the prompts in the file, one after the other:
 
-`ollama run llama2 < sourcequestions.txt`
+`ollama run llama3 < sourcequestions.txt`
 
 This concept is used in the following example.
 

examples/go-chat/main.go

@@ -35,7 +35,7 @@ func main() {
 
 	ctx := context.Background()
 	req := &api.ChatRequest{
-		Model:    "llama2",
+		Model:    "llama3",
 		Messages: messages,
 	}
 

examples/langchain-python-rag-document/main.py

@@ -40,9 +40,9 @@ while True:
         continue
 
     # Prompt
-    template = """Use the following pieces of context to answer the question at the end. 
-    If you don't know the answer, just say that you don't know, don't try to make up an answer. 
-    Use three sentences maximum and keep the answer as concise as possible. 
+    template = """Use the following pieces of context to answer the question at the end.
+    If you don't know the answer, just say that you don't know, don't try to make up an answer.
+    Use three sentences maximum and keep the answer as concise as possible.
     {context}
     Question: {question}
     Helpful Answer:"""
@@ -51,11 +51,11 @@ while True:
         template=template,
     )
 
-    llm = Ollama(model="llama2:13b", callback_manager=CallbackManager([StreamingStdOutCallbackHandler()]))
+    llm = Ollama(model="llama3:8b", callback_manager=CallbackManager([StreamingStdOutCallbackHandler()]))
     qa_chain = RetrievalQA.from_chain_type(
         llm,
         retriever=vectorstore.as_retriever(),
         chain_type_kwargs={"prompt": QA_CHAIN_PROMPT},
     )
 
-    result = qa_chain({"query": query})
+    result = qa_chain({"query": query})

examples/langchain-python-simple/README.md

@@ -4,10 +4,10 @@ This example is a basic "hello world" of using LangChain with Ollama.
 
 ## Running the Example
 
-1. Ensure you have the `llama2` model installed:
+1. Ensure you have the `llama3` model installed:
 
    ```bash
-   ollama pull llama2
+   ollama pull llama3
    ```
 
 2. Install the Python Requirements.
@@ -21,4 +21,3 @@ This example is a basic "hello world" of using LangChain with Ollama.
    ```bash
    python main.py
    ```
-  

examples/langchain-python-simple/main.py

@@ -1,6 +1,6 @@
 from langchain.llms import Ollama
 
 input = input("What is your question?")
-llm = Ollama(model="llama2")
+llm = Ollama(model="llama3")
 res = llm.predict(input)
 print (res)

examples/modelfile-mario/Modelfile

@@ -1,4 +1,4 @@
-FROM llama2
+FROM llama3
 PARAMETER temperature 1
 SYSTEM """
 You are Mario from super mario bros, acting as an assistant.

examples/modelfile-mario/readme.md

@@ -2,12 +2,12 @@
 
 # Example character: Mario
 
-This example shows how to create a basic character using Llama2 as the base model.
+This example shows how to create a basic character using Llama3 as the base model.
 
 To run this example:
 
 1. Download the Modelfile
-2. `ollama pull llama2` to get the base model used in the model file.
+2. `ollama pull llama3` to get the base model used in the model file.
 3. `ollama create NAME -f ./Modelfile`
 4. `ollama run NAME`
 
@@ -18,7 +18,7 @@ Ask it some questions like "Who are you?" or "Is Peach in trouble again?"
 What the model file looks like:
 
 ```
-FROM llama2
+FROM llama3
 PARAMETER temperature 1
 SYSTEM """
 You are Mario from Super Mario Bros, acting as an assistant.

examples/python-json-datagenerator/predefinedschema.py

@@ -2,16 +2,16 @@ import requests
 import json
 import random
 
-model = "llama2"
+model = "llama3"
 template = {
-  "firstName": "", 
-  "lastName": "", 
+  "firstName": "",
+  "lastName": "",
   "address": {
-    "street": "", 
-    "city": "", 
-    "state": "", 
+    "street": "",
+    "city": "",
+    "state": "",
     "zipCode": ""
-  }, 
+  },
   "phoneNumber": ""
 }
 

examples/python-json-datagenerator/randomaddresses.py

@@ -12,7 +12,7 @@ countries = [
     "France",
 ]
 country = random.choice(countries)
-model = "llama2"
+model = "llama3"
 
 prompt = f"generate one realistically believable sample data set of a persons first name, last name, address in {country}, and phone number. Do not use common names. Respond using JSON. Key names should have no backslashes, values should use plain ascii with no special characters."
 

examples/python-json-datagenerator/readme.md

@@ -6,10 +6,10 @@ There are two python scripts in this example. `randomaddresses.py` generates ran
 
 ## Running the Example
 
-1. Ensure you have the `llama2` model installed:
+1. Ensure you have the `llama3` model installed:
 
    ```bash
-   ollama pull llama2
+   ollama pull llama3
    ```
 
 2. Install the Python Requirements.

examples/python-simplechat/client.py

@@ -2,7 +2,7 @@ import json
 import requests
 
 # NOTE: ollama must be running for this to work, start the ollama app or run `ollama serve`
-model = "llama2"  # TODO: update this for whatever model you wish to use
+model = "llama3"  # TODO: update this for whatever model you wish to use
 
 
 def chat(messages):

examples/python-simplechat/readme.md

@@ -4,10 +4,10 @@ The **chat** endpoint is one of two ways to generate text from an LLM with Ollam
 
 ## Running the Example
 
-1. Ensure you have the `llama2` model installed:
+1. Ensure you have the `llama3` model installed:
 
    ```bash
-   ollama pull llama2
+   ollama pull llama3
    ```
 
 2. Install the Python Requirements.

examples/typescript-mentors/README.md

@@ -4,10 +4,10 @@ This example demonstrates how one would create a set of 'mentors' you can have a
 
 ## Usage
 
-1. Add llama2 to have the mentors ask your questions:
+1. Add llama3 to have the mentors ask your questions:
 
    ```bash
-   ollama pull llama2
+   ollama pull llama3
    ```
 
 2. Install prerequisites:

examples/typescript-mentors/character-generator.ts

@@ -15,7 +15,7 @@ async function characterGenerator() {
   ollama.setModel("stablebeluga2:70b-q4_K_M");
   const bio = await ollama.generate(`create a bio of ${character} in a single long paragraph. Instead of saying '${character} is...' or '${character} was...' use language like 'You are...' or 'You were...'. Then create a paragraph describing the speaking mannerisms and style of ${character}. Don't include anything about how ${character} looked or what they sounded like, just focus on the words they said. Instead of saying '${character} would say...' use language like 'You should say...'. If you use quotes, always use single quotes instead of double quotes. If there are any specific words or phrases you used a lot, show how you used them. `);
 
-  const thecontents = `FROM llama2\nSYSTEM """\n${bio.response.replace(/(\r\n|\n|\r)/gm, " ").replace('would', 'should')} All answers to questions should be related back to what you are most known for.\n"""`;
+  const thecontents = `FROM llama3\nSYSTEM """\n${bio.response.replace(/(\r\n|\n|\r)/gm, " ").replace('would', 'should')} All answers to questions should be related back to what you are most known for.\n"""`;
 
   fs.writeFile(path.join(directory, 'Modelfile'), thecontents, (err: any) => {
     if (err) throw err;
@@ -23,4 +23,4 @@ async function characterGenerator() {
   });
 }
 
-characterGenerator();
+characterGenerator();

examples/typescript-simplechat/client.ts

@@ -1,6 +1,6 @@
 import * as readline from "readline";
 
-const model = "llama2";
+const model = "llama3";
 type Message = {
   role: "assistant" | "user" | "system";
   content: string;
@@ -74,4 +74,4 @@ async function main() {
 
 }
 
-main();
+main();

format/bytes.go

@@ -15,6 +15,7 @@ const (
 
 	KibiByte = Byte * 1024
 	MebiByte = KibiByte * 1024
+	GibiByte = MebiByte * 1024
 )
 
 func HumanBytes(b int64) string {

gpu/amd_common.go

@@ -7,7 +7,7 @@ import (
 	"log/slog"
 	"os"
 	"path/filepath"
-	"strconv"
+	"runtime"
 	"strings"
 )
 
@@ -35,22 +35,64 @@ func GetSupportedGFX(libDir string) ([]string, error) {
 	return ret, nil
 }
 
-func amdSetVisibleDevices(ids []int, skip map[int]interface{}) {
-	// Set the visible devices if not already set
-	// TODO - does sort order matter?
-	devices := []string{}
-	for i := range ids {
-		if _, skipped := skip[i]; skipped {
+func rocmGetVisibleDevicesEnv(gpuInfo []GpuInfo) (string, string) {
+	ids := []string{}
+	for _, info := range gpuInfo {
+		if info.Library != "rocm" {
+			// TODO shouldn't happen if things are wired correctly...
+			slog.Debug("rocmGetVisibleDevicesEnv skipping over non-rocm device", "library", info.Library)
 			continue
 		}
-		devices = append(devices, strconv.Itoa(i))
+		ids = append(ids, info.ID)
+	}
+	return "HIP_VISIBLE_DEVICES", strings.Join(ids, ",")
+}
+
+func commonAMDValidateLibDir() (string, error) {
+	// We try to favor system paths first, so that we can wire up the subprocess to use
+	// the system version.  Only use our bundled version if the system version doesn't work
+	// This gives users a more recovery options if versions have subtle problems at runtime
+
+	// Prefer explicit HIP env var
+	hipPath := os.Getenv("HIP_PATH")
+	if hipPath != "" {
+		hipLibDir := filepath.Join(hipPath, "bin")
+		if rocmLibUsable(hipLibDir) {
+			slog.Debug("detected ROCM via HIP_PATH=" + hipPath)
+			return hipLibDir, nil
+		}
 	}
 
-	val := strings.Join(devices, ",")
-	err := os.Setenv("HIP_VISIBLE_DEVICES", val)
-	if err != nil {
-		slog.Warn(fmt.Sprintf("failed to set env: %s", err))
-	} else {
-		slog.Info("Setting HIP_VISIBLE_DEVICES=" + val)
+	// Scan the LD_LIBRARY_PATH or PATH
+	pathEnv := "LD_LIBRARY_PATH"
+	if runtime.GOOS == "windows" {
+		pathEnv = "PATH"
 	}
+
+	paths := os.Getenv(pathEnv)
+	for _, path := range filepath.SplitList(paths) {
+		d, err := filepath.Abs(path)
+		if err != nil {
+			continue
+		}
+		if rocmLibUsable(d) {
+			return d, nil
+		}
+	}
+
+	// Well known location(s)
+	if rocmLibUsable(RocmStandardLocation) {
+		return RocmStandardLocation, nil
+	}
+
+	// Installer payload location if we're running the installed binary
+	exe, err := os.Executable()
+	if err == nil {
+		rocmTargetDir := filepath.Join(filepath.Dir(exe), "rocm")
+		if rocmLibUsable(rocmTargetDir) {
+			slog.Debug("detected ROCM next to ollama executable " + rocmTargetDir)
+			return rocmTargetDir, nil
+		}
+	}
+	return "", fmt.Errorf("no suitable rocm found, falling back to CPU")
 }

gpu/amd_hip_windows.go

@@ -69,7 +69,7 @@ func NewHipLib() (*HipLib, error) {
 func (hl *HipLib) Release() {
 	err := windows.FreeLibrary(hl.dll)
 	if err != nil {
-		slog.Warn(fmt.Sprintf("failed to unload amdhip64.dll: %s", err))
+		slog.Warn("failed to unload amdhip64.dll", "error", err)
 	}
 	hl.dll = 0
 }
@@ -98,7 +98,7 @@ func (hl *HipLib) HipGetDeviceCount() int {
 		return 0
 	}
 	if status != hipSuccess {
-		slog.Warn(fmt.Sprintf("failed call to hipGetDeviceCount: %d %s", status, err))
+		slog.Warn("failed call to hipGetDeviceCount", "status", status, "error", err)
 	}
 	return count
 }

gpu/amd_linux.go

@@ -11,6 +11,8 @@ import (
 	"slices"
 	"strconv"
 	"strings"
+
+	"github.com/ollama/ollama/format"
 )
 
 // Discovery logic for AMD/ROCm GPUs
@@ -24,9 +26,6 @@ const (
 	GPUTotalMemoryFileGlob = "mem_banks/*/properties" // size_in_bytes line
 	GPUUsedMemoryFileGlob  = "mem_banks/*/used_memory"
 	RocmStandardLocation   = "/opt/rocm/lib"
-
-	// TODO find a better way to detect iGPU instead of minimum memory
-	IGPUMemLimit = 1024 * 1024 * 1024 // 512G is what they typically report, so anything less than 1G must be iGPU
 )
 
 var (
@@ -35,14 +34,11 @@ var (
 )
 
 // Gather GPU information from the amdgpu driver if any supported GPUs are detected
-// HIP_VISIBLE_DEVICES will be set if we detect a mix of unsupported and supported devices
-// and the user hasn't already set this variable
-func AMDGetGPUInfo(resp *GpuInfo) {
-	// TODO - DRY this out with windows
+func AMDGetGPUInfo() []GpuInfo {
+	resp := []GpuInfo{}
 	if !AMDDetected() {
-		return
+		return resp
 	}
-	skip := map[int]interface{}{}
 
 	// Opportunistic logging of driver version to aid in troubleshooting
 	ver, err := AMDDriverVersion()
@@ -50,160 +46,117 @@ func AMDGetGPUInfo(resp *GpuInfo) {
 		slog.Info("AMD Driver: " + ver)
 	} else {
 		// TODO - if we see users crash and burn with the upstreamed kernel this can be adjusted to hard-fail rocm support and fallback to CPU
-		slog.Warn(fmt.Sprintf("ollama recommends running the https://www.amd.com/en/support/linux-drivers: %s", err))
+		slog.Warn("ollama recommends running the https://www.amd.com/en/support/linux-drivers", "error", err)
 	}
 
-	// If the user has specified exactly which GPUs to use, look up their memory
-	visibleDevices := os.Getenv("HIP_VISIBLE_DEVICES")
-	if visibleDevices != "" {
-		ids := []int{}
-		for _, idStr := range strings.Split(visibleDevices, ",") {
-			id, err := strconv.Atoi(idStr)
-			if err != nil {
-				slog.Warn(fmt.Sprintf("malformed HIP_VISIBLE_DEVICES=%s %s", visibleDevices, err))
-			} else {
-				ids = append(ids, id)
-			}
-		}
-		amdProcMemLookup(resp, nil, ids)
-		return
+	// Determine if the user has already pre-selected which GPUs to look at, then ignore the others
+	var visibleDevices []string
+	hipVD := os.Getenv("HIP_VISIBLE_DEVICES")   // zero based index only
+	rocrVD := os.Getenv("ROCR_VISIBLE_DEVICES") // zero based index or UUID, but consumer cards seem to not support UUID
+	gpuDO := os.Getenv("GPU_DEVICE_ORDINAL")    // zero based index
+	switch {
+	// TODO is this priorty order right?
+	case hipVD != "":
+		visibleDevices = strings.Split(hipVD, ",")
+	case rocrVD != "":
+		visibleDevices = strings.Split(rocrVD, ",")
+		// TODO - since we don't yet support UUIDs, consider detecting and reporting here
+		// all our test systems show GPU-XX indicating UUID is not supported
+	case gpuDO != "":
+		visibleDevices = strings.Split(gpuDO, ",")
 	}
 
-	// Gather GFX version information from all detected cards
-	gfx := AMDGFXVersions()
-	verStrings := []string{}
-	for i, v := range gfx {
-		verStrings = append(verStrings, v.ToGFXString())
-		if v.Major == 0 {
-			// Silently skip CPUs
-			skip[i] = struct{}{}
-			continue
-		}
-		if v.Major < 9 {
-			// TODO consider this a build-time setting if we can support 8xx family GPUs
-			slog.Warn(fmt.Sprintf("amdgpu [%d] too old %s", i, v.ToGFXString()))
-			skip[i] = struct{}{}
-		}
-	}
-	slog.Info(fmt.Sprintf("detected amdgpu versions %v", verStrings))
-
-	// Abort if all GPUs are skipped
-	if len(skip) >= len(gfx) {
-		slog.Info("all detected amdgpus are skipped, falling back to CPU")
-		return
-	}
-
-	// If we got this far, then we have at least 1 GPU that's a ROCm candidate, so make sure we have a lib
-	libDir, err := AMDValidateLibDir()
-	if err != nil {
-		slog.Warn(fmt.Sprintf("unable to verify rocm library, will use cpu: %s", err))
-		return
-	}
-
-	updateLibPath(libDir)
-
 	gfxOverride := os.Getenv("HSA_OVERRIDE_GFX_VERSION")
-	if gfxOverride == "" {
-		supported, err := GetSupportedGFX(libDir)
+	var supported []string
+	libDir := ""
+
+	// The amdgpu driver always exposes the host CPU(s) first, but we have to skip them and subtract
+	// from the other IDs to get alignment with the HIP libraries expectations (zero is the first GPU, not the CPU)
+	matches, _ := filepath.Glob(GPUPropertiesFileGlob)
+	cpuCount := 0
+	for _, match := range matches {
+		slog.Debug("evaluating amdgpu node " + match)
+		fp, err := os.Open(match)
 		if err != nil {
-			slog.Warn(fmt.Sprintf("failed to lookup supported GFX types, falling back to CPU mode: %s", err))
-			return
-		}
-		slog.Debug(fmt.Sprintf("rocm supported GPU types %v", supported))
-
-		for i, v := range gfx {
-			if !slices.Contains[[]string, string](supported, v.ToGFXString()) {
-				slog.Warn(fmt.Sprintf("amdgpu [%d] %s is not supported by %s %v", i, v.ToGFXString(), libDir, supported))
-				// TODO - consider discrete markdown just for ROCM troubleshooting?
-				slog.Warn("See https://github.com/ollama/ollama/blob/main/docs/gpu.md#overrides for HSA_OVERRIDE_GFX_VERSION usage")
-				skip[i] = struct{}{}
-			} else {
-				slog.Info(fmt.Sprintf("amdgpu [%d] %s is supported", i, v.ToGFXString()))
-			}
-		}
-	} else {
-		slog.Debug("skipping rocm gfx compatibility check with HSA_OVERRIDE_GFX_VERSION=" + gfxOverride)
-	}
-
-	if len(skip) >= len(gfx) {
-		slog.Info("all detected amdgpus are skipped, falling back to CPU")
-		return
-	}
-
-	ids := make([]int, len(gfx))
-	i := 0
-	for k := range gfx {
-		ids[i] = k
-		i++
-	}
-	amdProcMemLookup(resp, skip, ids)
-	if resp.memInfo.DeviceCount == 0 {
-		return
-	}
-	if len(skip) > 0 {
-		amdSetVisibleDevices(ids, skip)
-	}
-}
-
-func updateLibPath(libDir string) {
-	ldPaths := []string{}
-	if val, ok := os.LookupEnv("LD_LIBRARY_PATH"); ok {
-		ldPaths = strings.Split(val, ":")
-	}
-	for _, d := range ldPaths {
-		if d == libDir {
-			return
-		}
-	}
-	val := strings.Join(append(ldPaths, libDir), ":")
-	slog.Debug("updated lib path", "LD_LIBRARY_PATH", val)
-	os.Setenv("LD_LIBRARY_PATH", val)
-}
-
-// Walk the sysfs nodes for the available GPUs and gather information from them
-// skipping over any devices in the skip map
-func amdProcMemLookup(resp *GpuInfo, skip map[int]interface{}, ids []int) {
-	resp.memInfo.DeviceCount = 0
-	resp.memInfo.TotalMemory = 0
-	resp.memInfo.FreeMemory = 0
-	slog.Debug("discovering VRAM for amdgpu devices")
-	if len(ids) == 0 {
-		entries, err := os.ReadDir(AMDNodesSysfsDir)
-		if err != nil {
-			slog.Warn(fmt.Sprintf("failed to read amdgpu sysfs %s - %s", AMDNodesSysfsDir, err))
-			return
-		}
-		for _, node := range entries {
-			if !node.IsDir() {
-				continue
-			}
-			id, err := strconv.Atoi(node.Name())
-			if err != nil {
-				slog.Warn("malformed amdgpu sysfs node id " + node.Name())
-				continue
-			}
-			ids = append(ids, id)
-		}
-	}
-	slog.Debug(fmt.Sprintf("amdgpu devices %v", ids))
-
-	for _, id := range ids {
-		if _, skipped := skip[id]; skipped {
+			slog.Debug("failed to open sysfs node", "file", match, "error", err)
 			continue
 		}
+		defer fp.Close()
+		nodeID, err := strconv.Atoi(filepath.Base(filepath.Dir(match)))
+		if err != nil {
+			slog.Debug("failed to parse node ID", "error", err)
+			continue
+		}
+
+		scanner := bufio.NewScanner(fp)
+		isCPU := false
+		var major, minor, patch uint64
+		for scanner.Scan() {
+			line := strings.TrimSpace(scanner.Text())
+			// Note: we could also use "cpu_cores_count X" where X is greater than zero to detect CPUs
+			if strings.HasPrefix(line, "gfx_target_version") {
+				ver := strings.Fields(line)
+
+				// Detect CPUs
+				if len(ver) == 2 && ver[1] == "0" {
+					slog.Debug("detected CPU " + match)
+					isCPU = true
+					break
+				}
+
+				if len(ver) != 2 || len(ver[1]) < 5 {
+					slog.Warn("malformed "+match, "gfx_target_version", line)
+					// If this winds up being a CPU, our offsets may be wrong
+					continue
+				}
+				l := len(ver[1])
+				var err1, err2, err3 error
+				patch, err1 = strconv.ParseUint(ver[1][l-2:l], 10, 32)
+				minor, err2 = strconv.ParseUint(ver[1][l-4:l-2], 10, 32)
+				major, err3 = strconv.ParseUint(ver[1][:l-4], 10, 32)
+				if err1 != nil || err2 != nil || err3 != nil {
+					slog.Debug("malformed int " + line)
+					continue
+				}
+			}
+
+			// TODO - any other properties we want to extract and record?
+			// vendor_id + device_id -> pci lookup for "Name"
+			// Other metrics that may help us understand relative performance between multiple GPUs
+		}
+
+		if isCPU {
+			cpuCount++
+			continue
+		}
+
+		// CPUs are always first in the list
+		gpuID := nodeID - cpuCount
+
+		// Shouldn't happen, but just in case...
+		if gpuID < 0 {
+			slog.Error("unexpected amdgpu sysfs data resulted in negative GPU ID, please set OLLAMA_DEBUG=1 and report an issue")
+			return []GpuInfo{}
+		}
+
+		if int(major) < RocmComputeMin {
+			slog.Warn(fmt.Sprintf("amdgpu too old gfx%d%d%x", major, minor, patch), "gpu", gpuID)
+			continue
+		}
+
+		// Look up the memory for the current node
 		totalMemory := uint64(0)
 		usedMemory := uint64(0)
-		// Adjust for sysfs vs HIP ids
-		propGlob := filepath.Join(AMDNodesSysfsDir, strconv.Itoa(id+1), GPUTotalMemoryFileGlob)
+		propGlob := filepath.Join(AMDNodesSysfsDir, strconv.Itoa(nodeID), GPUTotalMemoryFileGlob)
 		propFiles, err := filepath.Glob(propGlob)
 		if err != nil {
-			slog.Warn(fmt.Sprintf("error looking up total GPU memory: %s %s", propGlob, err))
+			slog.Warn("error looking up total GPU memory", "glob", propGlob, "error", err)
 		}
 		// 1 or more memory banks - sum the values of all of them
 		for _, propFile := range propFiles {
 			fp, err := os.Open(propFile)
 			if err != nil {
-				slog.Warn(fmt.Sprintf("failed to open sysfs node file %s: %s", propFile, err))
+				slog.Warn("failed to open sysfs node", "file", propFile, "erroir", err)
 				continue
 			}
 			defer fp.Close()
@@ -226,49 +179,113 @@ func amdProcMemLookup(resp *GpuInfo, skip map[int]interface{}, ids []int) {
 			}
 		}
 		if totalMemory == 0 {
-			slog.Warn(fmt.Sprintf("amdgpu [%d] reports zero total memory, skipping", id))
-			skip[id] = struct{}{}
+			slog.Warn("amdgpu reports zero total memory", "gpu", gpuID)
 			continue
 		}
-		if totalMemory < IGPUMemLimit {
-			slog.Info(fmt.Sprintf("amdgpu [%d] appears to be an iGPU with %dM reported total memory, skipping", id, totalMemory/1024/1024))
-			skip[id] = struct{}{}
-			continue
-		}
-		usedGlob := filepath.Join(AMDNodesSysfsDir, strconv.Itoa(id), GPUUsedMemoryFileGlob)
+		usedGlob := filepath.Join(AMDNodesSysfsDir, strconv.Itoa(nodeID), GPUUsedMemoryFileGlob)
 		usedFiles, err := filepath.Glob(usedGlob)
 		if err != nil {
-			slog.Warn(fmt.Sprintf("error looking up used GPU memory: %s %s", usedGlob, err))
+			slog.Warn("error looking up used GPU memory", "glob", usedGlob, "error", err)
 			continue
 		}
 		for _, usedFile := range usedFiles {
 			fp, err := os.Open(usedFile)
 			if err != nil {
-				slog.Warn(fmt.Sprintf("failed to open sysfs node file %s: %s", usedFile, err))
+				slog.Warn("failed to open sysfs node", "file", usedFile, "error", err)
 				continue
 			}
 			defer fp.Close()
 			data, err := io.ReadAll(fp)
 			if err != nil {
-				slog.Warn(fmt.Sprintf("failed to read sysfs node file %s: %s", usedFile, err))
+				slog.Warn("failed to read sysfs node", "file", usedFile, "error", err)
 				continue
 			}
 			used, err := strconv.ParseUint(strings.TrimSpace(string(data)), 10, 64)
 			if err != nil {
-				slog.Warn(fmt.Sprintf("malformed used memory %s: %s", string(data), err))
+				slog.Warn("malformed used memory", "data", string(data), "error", err)
 				continue
 			}
 			usedMemory += used
 		}
-		slog.Info(fmt.Sprintf("[%d] amdgpu totalMemory %dM", id, totalMemory/1024/1024))
-		slog.Info(fmt.Sprintf("[%d] amdgpu freeMemory  %dM", id, (totalMemory-usedMemory)/1024/1024))
-		resp.memInfo.DeviceCount++
-		resp.memInfo.TotalMemory += totalMemory
-		resp.memInfo.FreeMemory += (totalMemory - usedMemory)
+
+		// iGPU detection, remove this check once we can support an iGPU variant of the rocm library
+		if totalMemory < IGPUMemLimit {
+			slog.Info("amdgpu appears to be an iGPU, skipping", "gpu", gpuID, "total", format.HumanBytes2(totalMemory))
+			continue
+		}
+
+		slog.Info("amdgpu memory", "gpu", gpuID, "total", format.HumanBytes2(totalMemory))
+		slog.Info("amdgpu memory", "gpu", gpuID, "available", format.HumanBytes2(totalMemory-usedMemory))
+		gpuInfo := GpuInfo{
+			Library: "rocm",
+			memInfo: memInfo{
+				TotalMemory: totalMemory,
+				FreeMemory:  (totalMemory - usedMemory),
+			},
+			ID: fmt.Sprintf("%d", gpuID),
+			// Name: not exposed in sysfs directly, would require pci device id lookup
+			Major:         int(major),
+			Minor:         int(minor),
+			Patch:         int(patch),
+			MinimumMemory: rocmMinimumMemory,
+		}
+
+		// If the user wants to filter to a subset of devices, filter out if we aren't a match
+		if len(visibleDevices) > 0 {
+			include := false
+			for _, visible := range visibleDevices {
+				if visible == gpuInfo.ID {
+					include = true
+					break
+				}
+			}
+			if !include {
+				slog.Info("filtering out device per user request", "id", gpuInfo.ID, "visible_devices", visibleDevices)
+				continue
+			}
+		}
+
+		// Final validation is gfx compatibility - load the library if we haven't already loaded it
+		// even if the user overrides, we still need to validate the library
+		if libDir == "" {
+			libDir, err = AMDValidateLibDir()
+			if err != nil {
+				slog.Warn("unable to verify rocm library, will use cpu", "error", err)
+				return []GpuInfo{}
+			}
+		}
+		gpuInfo.DependencyPath = libDir
+
+		if gfxOverride == "" {
+			// Only load supported list once
+			if len(supported) == 0 {
+				supported, err = GetSupportedGFX(libDir)
+				if err != nil {
+					slog.Warn("failed to lookup supported GFX types, falling back to CPU mode", "error", err)
+					return []GpuInfo{}
+				}
+				slog.Debug("rocm supported GPUs", "types", supported)
+			}
+			gfx := fmt.Sprintf("gfx%d%d%x", gpuInfo.Major, gpuInfo.Minor, gpuInfo.Patch)
+			if !slices.Contains[[]string, string](supported, gfx) {
+				slog.Warn("amdgpu is not supported", "gpu", gpuInfo.ID, "gpu_type", gfx, "library", libDir, "supported_types", supported)
+				// TODO - consider discrete markdown just for ROCM troubleshooting?
+				slog.Warn("See https://github.com/ollama/ollama/blob/main/docs/gpu.md#overrides for HSA_OVERRIDE_GFX_VERSION usage")
+				continue
+			} else {
+				slog.Info("amdgpu is supported", "gpu", gpuInfo.ID, "gpu_type", gfx)
+			}
+		} else {
+			slog.Debug("skipping rocm gfx compatibility check with HSA_OVERRIDE_GFX_VERSION=" + gfxOverride)
+		}
+
+		// The GPU has passed all the verification steps and is supported
+		resp = append(resp, gpuInfo)
 	}
-	if resp.memInfo.DeviceCount > 0 {
-		resp.Library = "rocm"
+	if len(resp) == 0 {
+		slog.Info("no compatible amdgpu devices detected")
 	}
+	return resp
 }
 
 // Quick check for AMD driver so we can skip amdgpu discovery if not present
@@ -280,87 +297,24 @@ func AMDDetected() bool {
 		slog.Debug("amdgpu driver not detected " + sysfsDir)
 		return false
 	} else if err != nil {
-		slog.Debug(fmt.Sprintf("error looking up amd driver %s %s", sysfsDir, err))
+		slog.Debug("error looking up amd driver", "path", sysfsDir, "error", err)
 		return false
 	}
 	return true
 }
 
-func setupLink(source, target string) error {
-	if err := os.RemoveAll(target); err != nil {
-		return fmt.Errorf("failed to remove old rocm directory %s %w", target, err)
-	}
-	if err := os.Symlink(source, target); err != nil {
-		return fmt.Errorf("failed to create link %s => %s %w", source, target, err)
-	}
-	slog.Debug(fmt.Sprintf("host rocm linked %s => %s", source, target))
-	return nil
-}
-
-// Ensure the AMD rocm lib dir is wired up
 // Prefer to use host installed ROCm, as long as it meets our minimum requirements
 // failing that, tell the user how to download it on their own
 func AMDValidateLibDir() (string, error) {
-	// We rely on the rpath compiled into our library to find rocm
-	// so we establish a symlink to wherever we find it on the system
-	// to <payloads>/rocm
-	payloadsDir, err := PayloadsDir()
-	if err != nil {
-		return "", err
-	}
-
-	// If we already have a rocm dependency wired, nothing more to do
-	rocmTargetDir := filepath.Clean(filepath.Join(payloadsDir, "..", "rocm"))
-	if rocmLibUsable(rocmTargetDir) {
-		return rocmTargetDir, nil
-	}
-
-	// next to the running binary
-	exe, err := os.Executable()
+	libDir, err := commonAMDValidateLibDir()
 	if err == nil {
-		peerDir := filepath.Dir(exe)
-		if rocmLibUsable(peerDir) {
-			slog.Debug("detected ROCM next to ollama executable " + peerDir)
-			return rocmTargetDir, setupLink(peerDir, rocmTargetDir)
-		}
-		peerDir = filepath.Join(filepath.Dir(exe), "rocm")
-		if rocmLibUsable(peerDir) {
-			slog.Debug("detected ROCM next to ollama executable " + peerDir)
-			return rocmTargetDir, setupLink(peerDir, rocmTargetDir)
-		}
+		return libDir, nil
 	}
 
 	// Well known ollama installer path
 	installedRocmDir := "/usr/share/ollama/lib/rocm"
 	if rocmLibUsable(installedRocmDir) {
-		return rocmTargetDir, setupLink(installedRocmDir, rocmTargetDir)
-	}
-
-	// Prefer explicit HIP env var
-	hipPath := os.Getenv("HIP_PATH")
-	if hipPath != "" {
-		hipLibDir := filepath.Join(hipPath, "lib")
-		if rocmLibUsable(hipLibDir) {
-			slog.Debug("detected ROCM via HIP_PATH=" + hipPath)
-			return rocmTargetDir, setupLink(hipLibDir, rocmTargetDir)
-		}
-	}
-
-	// Scan the library path for potential matches
-	ldPaths := strings.Split(os.Getenv("LD_LIBRARY_PATH"), ":")
-	for _, ldPath := range ldPaths {
-		d, err := filepath.Abs(ldPath)
-		if err != nil {
-			continue
-		}
-		if rocmLibUsable(d) {
-			return rocmTargetDir, setupLink(d, rocmTargetDir)
-		}
-	}
-
-	// Well known location(s)
-	if rocmLibUsable("/opt/rocm/lib") {
-		return rocmTargetDir, setupLink("/opt/rocm/lib", rocmTargetDir)
+		return installedRocmDir, nil
 	}
 
 	// If we still haven't found a usable rocm, the user will have to install it on their own
@@ -384,68 +338,3 @@ func AMDDriverVersion() (string, error) {
 	}
 	return strings.TrimSpace(string(verString)), nil
 }
-
-func AMDGFXVersions() map[int]Version {
-	// The amdgpu driver always exposes the host CPU as node 0, but we have to skip that and subtract one
-	// from the other IDs to get alignment with the HIP libraries expectations (zero is the first GPU, not the CPU)
-	res := map[int]Version{}
-	matches, _ := filepath.Glob(GPUPropertiesFileGlob)
-	for _, match := range matches {
-		fp, err := os.Open(match)
-		if err != nil {
-			slog.Debug(fmt.Sprintf("failed to open sysfs node file %s: %s", match, err))
-			continue
-		}
-		defer fp.Close()
-		i, err := strconv.Atoi(filepath.Base(filepath.Dir(match)))
-		if err != nil {
-			slog.Debug(fmt.Sprintf("failed to parse node ID %s", err))
-			continue
-		}
-
-		if i == 0 {
-			// Skipping the CPU
-			continue
-		}
-		// Align with HIP IDs (zero is first GPU, not CPU)
-		i -= 1
-
-		scanner := bufio.NewScanner(fp)
-		for scanner.Scan() {
-			line := strings.TrimSpace(scanner.Text())
-			if strings.HasPrefix(line, "gfx_target_version") {
-				ver := strings.Fields(line)
-				if len(ver) != 2 || len(ver[1]) < 5 {
-					if ver[1] != "0" {
-						slog.Debug("malformed " + line)
-					}
-					res[i] = Version{
-						Major: 0,
-						Minor: 0,
-						Patch: 0,
-					}
-					continue
-				}
-				l := len(ver[1])
-				patch, err1 := strconv.ParseUint(ver[1][l-2:l], 10, 32)
-				minor, err2 := strconv.ParseUint(ver[1][l-4:l-2], 10, 32)
-				major, err3 := strconv.ParseUint(ver[1][:l-4], 10, 32)
-				if err1 != nil || err2 != nil || err3 != nil {
-					slog.Debug("malformed int " + line)
-					continue
-				}
-
-				res[i] = Version{
-					Major: uint(major),
-					Minor: uint(minor),
-					Patch: uint(patch),
-				}
-			}
-		}
-	}
-	return res
-}
-
-func (v Version) ToGFXString() string {
-	return fmt.Sprintf("gfx%d%d%d", v.Major, v.Minor, v.Patch)
-}

gpu/amd_windows.go

@@ -7,7 +7,10 @@ import (
 	"os"
 	"path/filepath"
 	"slices"
+	"strconv"
 	"strings"
+
+	"github.com/ollama/ollama/format"
 )
 
 const (
@@ -22,36 +25,32 @@ var (
 	ROCmLibGlobs = []string{"hipblas.dll", "rocblas"} // TODO - probably include more coverage of files here...
 )
 
-func AMDGetGPUInfo(resp *GpuInfo) {
+func AMDGetGPUInfo() []GpuInfo {
+	resp := []GpuInfo{}
 	hl, err := NewHipLib()
 	if err != nil {
 		slog.Debug(err.Error())
-		return
+		return nil
 	}
 	defer hl.Release()
-	skip := map[int]interface{}{}
-	ids := []int{}
-	resp.memInfo.DeviceCount = 0
-	resp.memInfo.TotalMemory = 0
-	resp.memInfo.FreeMemory = 0
 
 	ver, err := hl.AMDDriverVersion()
 	if err == nil {
 		slog.Info("AMD Driver: " + ver)
 	} else {
 		// For now this is benign, but we may eventually need to fail compatibility checks
-		slog.Debug(fmt.Sprintf("error looking up amd driver version: %s", err))
+		slog.Debug("error looking up amd driver version", "error", err)
 	}
 
-	// Note: the HIP library automatically handles HIP_VISIBLE_DEVICES
+	// Note: the HIP library automatically handles subsetting to any HIP_VISIBLE_DEVICES the user specified
 	count := hl.HipGetDeviceCount()
 	if count == 0 {
-		return
+		return nil
 	}
 	libDir, err := AMDValidateLibDir()
 	if err != nil {
-		slog.Warn(fmt.Sprintf("unable to verify rocm library, will use cpu: %s", err))
-		return
+		slog.Warn("unable to verify rocm library, will use cpu", "error", err)
+		return nil
 	}
 
 	var supported []string
@@ -59,95 +58,120 @@ func AMDGetGPUInfo(resp *GpuInfo) {
 	if gfxOverride == "" {
 		supported, err = GetSupportedGFX(libDir)
 		if err != nil {
-			slog.Warn(fmt.Sprintf("failed to lookup supported GFX types, falling back to CPU mode: %s", err))
-			return
+			slog.Warn("failed to lookup supported GFX types, falling back to CPU mode", "error", err)
+			return nil
 		}
 	} else {
 		slog.Debug("skipping rocm gfx compatibility check with HSA_OVERRIDE_GFX_VERSION=" + gfxOverride)
 	}
 
-	slog.Info(fmt.Sprintf("detected %d hip devices", count))
+	slog.Info("detected hip devices", "count", count)
+	// TODO how to determine the underlying device ID when visible devices is causing this to subset?
 	for i := 0; i < count; i++ {
-		ids = append(ids, i)
 		err = hl.HipSetDevice(i)
 		if err != nil {
-			slog.Warn(fmt.Sprintf("[%d] %s", i, err))
-			skip[i] = struct{}{}
+			slog.Warn("set device", "id", i, "error", err)
 			continue
 		}
 
 		props, err := hl.HipGetDeviceProperties(i)
 		if err != nil {
-			slog.Warn(fmt.Sprintf("[%d] %s", i, err))
-			skip[i] = struct{}{}
+			slog.Warn("get properties", "id", i, "error", err)
 			continue
 		}
 		n := bytes.IndexByte(props.Name[:], 0)
 		name := string(props.Name[:n])
-		slog.Info(fmt.Sprintf("[%d] Name: %s", i, name))
+		// TODO is UUID actually populated on windows?
+		// Can luid be used on windows for setting visible devices (and is it actually set?)
 		n = bytes.IndexByte(props.GcnArchName[:], 0)
 		gfx := string(props.GcnArchName[:n])
-		slog.Info(fmt.Sprintf("[%d] GcnArchName: %s", i, gfx))
+		slog.Info("hip device", "id", i, "name", name, "gfx", gfx)
+		var major, minor, patch string
+		switch len(gfx) {
+		case 6:
+			major, minor, patch = gfx[3:4], gfx[4:5], gfx[5:]
+		case 7:
+			major, minor, patch = gfx[3:5], gfx[5:6], gfx[6:]
+		}
 		//slog.Info(fmt.Sprintf("[%d] Integrated: %d", i, props.iGPU)) // DOESN'T REPORT CORRECTLY!  Always 0
 		// TODO  Why isn't props.iGPU accurate!?
 		if strings.EqualFold(name, iGPUName) {
-			slog.Info(fmt.Sprintf("iGPU detected [%d] skipping", i))
-			skip[i] = struct{}{}
+			slog.Info("iGPU detected skipping", "id", i)
 			continue
 		}
 		if gfxOverride == "" {
 			if !slices.Contains[[]string, string](supported, gfx) {
-				slog.Warn(fmt.Sprintf("amdgpu [%d] %s is not supported by %s %v", i, gfx, libDir, supported))
+				slog.Warn("amdgpu is not supported", "gpu", i, "gpu_type", gfx, "library", libDir, "supported_types", supported)
 				// TODO - consider discrete markdown just for ROCM troubleshooting?
 				slog.Warn("See https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md for HSA_OVERRIDE_GFX_VERSION usage")
-				skip[i] = struct{}{}
 				continue
 			} else {
-				slog.Info(fmt.Sprintf("amdgpu [%d] %s is supported", i, gfx))
+				slog.Info("amdgpu is supported", "gpu", i, "gpu_type", gfx)
 			}
 		}
 
-		totalMemory, freeMemory, err := hl.HipMemGetInfo()
+		freeMemory, totalMemory, err := hl.HipMemGetInfo()
 		if err != nil {
-			slog.Warn(fmt.Sprintf("[%d] %s", i, err))
+			slog.Warn("get mem info", "id", i, "error", err)
 			continue
 		}
 
-		// TODO according to docs, freeMem may lie on windows!
-		slog.Info(fmt.Sprintf("[%d] Total Mem: %d", i, totalMemory))
-		slog.Info(fmt.Sprintf("[%d] Free Mem:  %d", i, freeMemory))
-		resp.memInfo.DeviceCount++
-		resp.memInfo.TotalMemory += totalMemory
-		resp.memInfo.FreeMemory += freeMemory
+		// iGPU detection, remove this check once we can support an iGPU variant of the rocm library
+		if totalMemory < IGPUMemLimit {
+			slog.Info("amdgpu appears to be an iGPU, skipping", "gpu", i, "total", format.HumanBytes2(totalMemory))
+			continue
+		}
+
+		// TODO revisit this once ROCm v6 is available on windows.
+		// v5.7 only reports VRAM used by this process, so it's completely wrong and unusable
+		slog.Info("amdgpu memory", "gpu", i, "total", format.HumanBytes2(totalMemory))
+		slog.Info("amdgpu memory", "gpu", i, "available", format.HumanBytes2(freeMemory))
+		gpuInfo := GpuInfo{
+			Library: "rocm",
+			memInfo: memInfo{
+				TotalMemory: totalMemory,
+				FreeMemory:  freeMemory,
+			},
+			ID:             fmt.Sprintf("%d", i), // TODO this is probably wrong if we specify visible devices
+			DependencyPath: libDir,
+			MinimumMemory:  rocmMinimumMemory,
+		}
+		if major != "" {
+			gpuInfo.Major, err = strconv.Atoi(major)
+			if err != nil {
+				slog.Info("failed to parse version", "version", gfx, "error", err)
+			}
+		}
+		if minor != "" {
+			gpuInfo.Minor, err = strconv.Atoi(minor)
+			if err != nil {
+				slog.Info("failed to parse version", "version", gfx, "error", err)
+			}
+		}
+		if patch != "" {
+			// Patch rev is hex; e.g. gfx90a
+			p, err := strconv.ParseInt(patch, 16, 0)
+			if err != nil {
+				slog.Info("failed to parse version", "version", gfx, "error", err)
+			} else {
+				gpuInfo.Patch = int(p)
+			}
+		}
+		if gpuInfo.Major < RocmComputeMin {
+			slog.Warn(fmt.Sprintf("amdgpu [%s] too old gfx%d%d%x", gpuInfo.ID, gpuInfo.Major, gpuInfo.Minor, gpuInfo.Patch))
+			continue
+		}
+
+		resp = append(resp, gpuInfo)
 	}
-	if resp.memInfo.DeviceCount > 0 {
-		resp.Library = "rocm"
-	}
-	// Abort if all GPUs are skipped
-	if len(skip) >= count {
-		slog.Info("all detected amdgpus are skipped, falling back to CPU")
-		return
-	}
-	if len(skip) > 0 {
-		amdSetVisibleDevices(ids, skip)
-	}
-	UpdatePath(libDir)
+
+	return resp
 }
 
 func AMDValidateLibDir() (string, error) {
-	// On windows non-admins typically can't create links
-	// so instead of trying to rely on rpath and a link in
-	// $LibDir/rocm, we instead rely on setting PATH to point
-	// to the location of the ROCm library
-
-	// Installer payload location if we're running the installed binary
-	exe, err := os.Executable()
+	libDir, err := commonAMDValidateLibDir()
 	if err == nil {
-		rocmTargetDir := filepath.Join(filepath.Dir(exe), "rocm")
-		if rocmLibUsable(rocmTargetDir) {
-			slog.Debug("detected ROCM next to ollama executable " + rocmTargetDir)
-			return rocmTargetDir, nil
-		}
+		return libDir, nil
 	}
 
 	// Installer payload (if we're running from some other location)
@@ -159,21 +183,6 @@ func AMDValidateLibDir() (string, error) {
 		return rocmTargetDir, nil
 	}
 
-	// Prefer explicit HIP env var
-	hipPath := os.Getenv("HIP_PATH")
-	if hipPath != "" {
-		hipLibDir := filepath.Join(hipPath, "bin")
-		if rocmLibUsable(hipLibDir) {
-			slog.Debug("detected ROCM via HIP_PATH=" + hipPath)
-			return hipLibDir, nil
-		}
-	}
-
-	// Well known location(s)
-	if rocmLibUsable(RocmStandardLocation) {
-		return RocmStandardLocation, nil
-	}
-
 	// Should not happen on windows since we include it in the installer, but stand-alone binary might hit this
 	slog.Warn("amdgpu detected, but no compatible rocm library found.  Please install ROCm")
 	return "", fmt.Errorf("no suitable rocm found, falling back to CPU")

gpu/assets.go

@@ -24,6 +24,51 @@ func PayloadsDir() (string, error) {
 	defer lock.Unlock()
 	var err error
 	if payloadsDir == "" {
+		runnersDir := os.Getenv("OLLAMA_RUNNERS_DIR")
+		// On Windows we do not carry the payloads inside the main executable
+		if runtime.GOOS == "windows" && runnersDir == "" {
+			appExe, err := os.Executable()
+			if err != nil {
+				slog.Error("failed to lookup executable path", "error", err)
+				return "", err
+			}
+
+			cwd, err := os.Getwd()
+			if err != nil {
+				slog.Error("failed to lookup working directory", "error", err)
+				return "", err
+			}
+
+			var paths []string
+			for _, root := range []string{filepath.Dir(appExe), cwd} {
+				paths = append(paths,
+					filepath.Join(root),
+					filepath.Join(root, "windows-"+runtime.GOARCH),
+					filepath.Join(root, "dist", "windows-"+runtime.GOARCH),
+				)
+			}
+
+			// Try a few variations to improve developer experience when building from source in the local tree
+			for _, p := range paths {
+				candidate := filepath.Join(p, "ollama_runners")
+				_, err := os.Stat(candidate)
+				if err == nil {
+					runnersDir = candidate
+					break
+				}
+			}
+			if runnersDir == "" {
+				err = fmt.Errorf("unable to locate llm runner directory.  Set OLLAMA_RUNNERS_DIR to the location of 'ollama_runners'")
+				slog.Error("incomplete distribution", "error", err)
+				return "", err
+			}
+		}
+		if runnersDir != "" {
+			payloadsDir = runnersDir
+			return payloadsDir, nil
+		}
+
+		// The remainder only applies on non-windows where we still carry payloads in the main executable
 		cleanupTmpDirs()
 		tmpDir := os.Getenv("OLLAMA_TMPDIR")
 		if tmpDir == "" {
@@ -80,7 +125,7 @@ func cleanupTmpDirs() {
 		}
 		err = os.RemoveAll(d)
 		if err != nil {
-			slog.Debug(fmt.Sprintf("unable to cleanup stale tmpdir %s: %s", d, err))
+			slog.Debug("unable to cleanup stale tmpdir", "path", d, "error", err)
 		}
 	}
 }
@@ -88,7 +133,8 @@ func cleanupTmpDirs() {
 func Cleanup() {
 	lock.Lock()
 	defer lock.Unlock()
-	if payloadsDir != "" {
+	runnersDir := os.Getenv("OLLAMA_RUNNERS_DIR")
+	if payloadsDir != "" && runnersDir == "" && runtime.GOOS != "windows" {
 		// We want to fully clean up the tmpdir parent of the payloads dir
 		tmpDir := filepath.Clean(filepath.Join(payloadsDir, ".."))
 		slog.Debug("cleaning up", "dir", tmpDir)
@@ -120,7 +166,7 @@ func UpdatePath(dir string) {
 			}
 		}
 		newPath := strings.Join(append([]string{dir}, pathComponents...), ";")
-		slog.Info(fmt.Sprintf("Updating PATH to %s", newPath))
+		slog.Info("updating", "PATH", newPath)
 		os.Setenv("PATH", newPath)
 	}
 	// linux and darwin rely on rpath

gpu/cuda_common.go

@@ -0,0 +1,22 @@
+//go:build linux || windows
+
+package gpu
+
+import (
+	"log/slog"
+	"strings"
+)
+
+func cudaGetVisibleDevicesEnv(gpuInfo []GpuInfo) (string, string) {
+	ids := []string{}
+	for _, info := range gpuInfo {
+		if info.Library != "cuda" {
+			// TODO shouldn't happen if things are wired correctly...
+			slog.Debug("cudaGetVisibleDevicesEnv skipping over non-cuda device", "library", info.Library)
+			continue
+		}
+		ids = append(ids, info.ID)
+	}
+	return "CUDA_VISIBLE_DEVICES", strings.Join(ids, ",")
+
+}

gpu/gpu.go

@@ -16,7 +16,6 @@ import (
 	"os"
 	"path/filepath"
 	"runtime"
-	"strconv"
 	"strings"
 	"sync"
 	"unsafe"
@@ -25,8 +24,8 @@ import (
 )
 
 type handles struct {
-	nvml   *C.nvml_handle_t
-	cudart *C.cudart_handle_t
+	deviceCount int
+	cudart      *C.cudart_handle_t
 }
 
 const (
@@ -39,26 +38,10 @@ var gpuMutex sync.Mutex
 // With our current CUDA compile flags, older than 5.0 will not work properly
 var CudaComputeMin = [2]C.int{5, 0}
 
-// Possible locations for the nvidia-ml library
-var NvmlLinuxGlobs = []string{
-	"/usr/local/cuda/lib64/libnvidia-ml.so*",
-	"/usr/lib/x86_64-linux-gnu/nvidia/current/libnvidia-ml.so*",
-	"/usr/lib/x86_64-linux-gnu/libnvidia-ml.so*",
-	"/usr/lib/wsl/lib/libnvidia-ml.so*",
-	"/usr/lib/wsl/drivers/*/libnvidia-ml.so*",
-	"/opt/cuda/lib64/libnvidia-ml.so*",
-	"/usr/lib*/libnvidia-ml.so*",
-	"/usr/lib/aarch64-linux-gnu/nvidia/current/libnvidia-ml.so*",
-	"/usr/lib/aarch64-linux-gnu/libnvidia-ml.so*",
-	"/usr/local/lib*/libnvidia-ml.so*",
+var RocmComputeMin = 9
 
-	// TODO: are these stubs ever valid?
-	"/opt/cuda/targets/x86_64-linux/lib/stubs/libnvidia-ml.so*",
-}
-
-var NvmlWindowsGlobs = []string{
-	"c:\\Windows\\System32\\nvml.dll",
-}
+// TODO find a better way to detect iGPU instead of minimum memory
+const IGPUMemLimit = 1 * format.GibiByte // 512G is what they typically report, so anything less than 1G must be iGPU
 
 var CudartLinuxGlobs = []string{
 	"/usr/local/cuda/lib64/libcudart.so*",
@@ -88,26 +71,18 @@ func initGPUHandles() *handles {
 
 	// TODO - if the ollama build is CPU only, don't do these checks as they're irrelevant and confusing
 
-	gpuHandles := &handles{nil, nil}
-	var nvmlMgmtName string
-	var nvmlMgmtPatterns []string
+	gpuHandles := &handles{}
 	var cudartMgmtName string
 	var cudartMgmtPatterns []string
 
 	tmpDir, _ := PayloadsDir()
 	switch runtime.GOOS {
 	case "windows":
-		nvmlMgmtName = "nvml.dll"
-		nvmlMgmtPatterns = make([]string, len(NvmlWindowsGlobs))
-		copy(nvmlMgmtPatterns, NvmlWindowsGlobs)
 		cudartMgmtName = "cudart64_*.dll"
 		localAppData := os.Getenv("LOCALAPPDATA")
 		cudartMgmtPatterns = []string{filepath.Join(localAppData, "Programs", "Ollama", cudartMgmtName)}
 		cudartMgmtPatterns = append(cudartMgmtPatterns, CudartWindowsGlobs...)
 	case "linux":
-		nvmlMgmtName = "libnvidia-ml.so"
-		nvmlMgmtPatterns = make([]string, len(NvmlLinuxGlobs))
-		copy(nvmlMgmtPatterns, NvmlLinuxGlobs)
 		cudartMgmtName = "libcudart.so*"
 		if tmpDir != "" {
 			// TODO - add "payloads" for subprocess
@@ -118,31 +93,21 @@ func initGPUHandles() *handles {
 		return gpuHandles
 	}
 
-	slog.Info("Detecting GPU type")
+	slog.Info("Detecting GPUs")
 	cudartLibPaths := FindGPULibs(cudartMgmtName, cudartMgmtPatterns)
 	if len(cudartLibPaths) > 0 {
-		cudart := LoadCUDARTMgmt(cudartLibPaths)
+		deviceCount, cudart, libPath := LoadCUDARTMgmt(cudartLibPaths)
 		if cudart != nil {
-			slog.Info("Nvidia GPU detected via cudart")
+			slog.Info("detected GPUs", "library", libPath, "count", deviceCount)
 			gpuHandles.cudart = cudart
-			return gpuHandles
-		}
-	}
-
-	// TODO once we build confidence, remove this and the gpu_info_nvml.[ch] files
-	nvmlLibPaths := FindGPULibs(nvmlMgmtName, nvmlMgmtPatterns)
-	if len(nvmlLibPaths) > 0 {
-		nvml := LoadNVMLMgmt(nvmlLibPaths)
-		if nvml != nil {
-			slog.Info("Nvidia GPU detected via nvidia-ml")
-			gpuHandles.nvml = nvml
+			gpuHandles.deviceCount = deviceCount
 			return gpuHandles
 		}
 	}
 	return gpuHandles
 }
 
-func GetGPUInfo() GpuInfo {
+func GetGPUInfo() GpuInfoList {
 	// TODO - consider exploring lspci (and equivalent on windows) to check for
 	// GPUs so we can report warnings if we see Nvidia/AMD but fail to load the libraries
 	gpuMutex.Lock()
@@ -150,9 +115,6 @@ func GetGPUInfo() GpuInfo {
 
 	gpuHandles := initGPUHandles()
 	defer func() {
-		if gpuHandles.nvml != nil {
-			C.nvml_release(*gpuHandles.nvml)
-		}
 		if gpuHandles.cudart != nil {
 			C.cudart_release(*gpuHandles.cudart)
 		}
@@ -165,72 +127,63 @@ func GetGPUInfo() GpuInfo {
 	}
 
 	var memInfo C.mem_info_t
-	resp := GpuInfo{}
-	if gpuHandles.nvml != nil && (cpuVariant != "" || runtime.GOARCH != "amd64") {
-		C.nvml_check_vram(*gpuHandles.nvml, &memInfo)
+	resp := []GpuInfo{}
+
+	// NVIDIA first
+	for i := 0; i < gpuHandles.deviceCount; i++ {
+		// TODO once we support CPU compilation variants of GPU libraries refine this...
+		if cpuVariant == "" && runtime.GOARCH == "amd64" {
+			continue
+		}
+		gpuInfo := GpuInfo{
+			Library: "cuda",
+		}
+		C.cudart_check_vram(*gpuHandles.cudart, C.int(i), &memInfo)
 		if memInfo.err != nil {
-			slog.Info(fmt.Sprintf("[nvidia-ml] error looking up NVML GPU memory: %s", C.GoString(memInfo.err)))
+			slog.Info("error looking up nvidia GPU memory", "error", C.GoString(memInfo.err))
 			C.free(unsafe.Pointer(memInfo.err))
-		} else if memInfo.count > 0 {
-			// Verify minimum compute capability
-			var cc C.nvml_compute_capability_t
-			C.nvml_compute_capability(*gpuHandles.nvml, &cc)
-			if cc.err != nil {
-				slog.Info(fmt.Sprintf("[nvidia-ml] error looking up NVML GPU compute capability: %s", C.GoString(cc.err)))
-				C.free(unsafe.Pointer(cc.err))
-			} else if cc.major > CudaComputeMin[0] || (cc.major == CudaComputeMin[0] && cc.minor >= CudaComputeMin[1]) {
-				slog.Info(fmt.Sprintf("[nvidia-ml] NVML CUDA Compute Capability detected: %d.%d", cc.major, cc.minor))
-				resp.Library = "cuda"
-				resp.MinimumMemory = cudaMinimumMemory
-			} else {
-				slog.Info(fmt.Sprintf("[nvidia-ml] CUDA GPU is too old. Falling back to CPU mode. Compute Capability detected: %d.%d", cc.major, cc.minor))
-			}
+			continue
 		}
-	} else if gpuHandles.cudart != nil && (cpuVariant != "" || runtime.GOARCH != "amd64") {
-		C.cudart_check_vram(*gpuHandles.cudart, &memInfo)
-		if memInfo.err != nil {
-			slog.Info(fmt.Sprintf("[cudart] error looking up CUDART GPU memory: %s", C.GoString(memInfo.err)))
-			C.free(unsafe.Pointer(memInfo.err))
-		} else if memInfo.count > 0 {
-			// Verify minimum compute capability
-			var cc C.cudart_compute_capability_t
-			C.cudart_compute_capability(*gpuHandles.cudart, &cc)
-			if cc.err != nil {
-				slog.Info(fmt.Sprintf("[cudart] error looking up CUDA compute capability: %s", C.GoString(cc.err)))
-				C.free(unsafe.Pointer(cc.err))
-			} else if cc.major > CudaComputeMin[0] || (cc.major == CudaComputeMin[0] && cc.minor >= CudaComputeMin[1]) {
-				slog.Info(fmt.Sprintf("[cudart] CUDART CUDA Compute Capability detected: %d.%d", cc.major, cc.minor))
-				resp.Library = "cuda"
-				resp.MinimumMemory = cudaMinimumMemory
-			} else {
-				slog.Info(fmt.Sprintf("[cudart] CUDA GPU is too old. Falling back to CPU mode. Compute Capability detected: %d.%d", cc.major, cc.minor))
-			}
+		if memInfo.major < CudaComputeMin[0] || (memInfo.major == CudaComputeMin[0] && memInfo.minor < CudaComputeMin[1]) {
+			slog.Info(fmt.Sprintf("[%d] CUDA GPU is too old. Compute Capability detected: %d.%d", i, memInfo.major, memInfo.minor))
+			continue
 		}
-	} else {
-		AMDGetGPUInfo(&resp)
-		if resp.Library != "" {
-			resp.MinimumMemory = rocmMinimumMemory
-			return resp
-		}
-	}
-	if resp.Library == "" {
-		C.cpu_check_ram(&memInfo)
-		resp.Library = "cpu"
-		resp.Variant = cpuVariant
-	}
-	if memInfo.err != nil {
-		slog.Info(fmt.Sprintf("error looking up CPU memory: %s", C.GoString(memInfo.err)))
-		C.free(unsafe.Pointer(memInfo.err))
-		return resp
+		gpuInfo.TotalMemory = uint64(memInfo.total)
+		gpuInfo.FreeMemory = uint64(memInfo.free)
+		gpuInfo.ID = C.GoString(&memInfo.gpu_id[0])
+		gpuInfo.Major = int(memInfo.major)
+		gpuInfo.Minor = int(memInfo.minor)
+		gpuInfo.MinimumMemory = cudaMinimumMemory
+
+		// TODO potentially sort on our own algorithm instead of what the underlying GPU library does...
+		resp = append(resp, gpuInfo)
+	}
+
+	// Then AMD
+	resp = append(resp, AMDGetGPUInfo()...)
+
+	if len(resp) == 0 {
+		C.cpu_check_ram(&memInfo)
+		if memInfo.err != nil {
+			slog.Info("error looking up CPU memory", "error", C.GoString(memInfo.err))
+			C.free(unsafe.Pointer(memInfo.err))
+			return resp
+		}
+		gpuInfo := GpuInfo{
+			Library: "cpu",
+			Variant: cpuVariant,
+		}
+		gpuInfo.TotalMemory = uint64(memInfo.total)
+		gpuInfo.FreeMemory = uint64(memInfo.free)
+		gpuInfo.ID = C.GoString(&memInfo.gpu_id[0])
+
+		resp = append(resp, gpuInfo)
 	}
 
-	resp.DeviceCount = uint32(memInfo.count)
-	resp.FreeMemory = uint64(memInfo.free)
-	resp.TotalMemory = uint64(memInfo.total)
 	return resp
 }
 
-func getCPUMem() (memInfo, error) {
+func GetCPUMem() (memInfo, error) {
 	var ret memInfo
 	var info C.mem_info_t
 	C.cpu_check_ram(&info)
@@ -243,29 +196,11 @@ func getCPUMem() (memInfo, error) {
 	return ret, nil
 }
 
-func CheckVRAM() (uint64, error) {
-	userLimit := os.Getenv("OLLAMA_MAX_VRAM")
-	if userLimit != "" {
-		avail, err := strconv.ParseInt(userLimit, 10, 64)
-		if err != nil {
-			return 0, fmt.Errorf("Invalid OLLAMA_MAX_VRAM setting %s: %s", userLimit, err)
-		}
-		slog.Info(fmt.Sprintf("user override OLLAMA_MAX_VRAM=%d", avail))
-		return uint64(avail), nil
-	}
-	gpuInfo := GetGPUInfo()
-	if gpuInfo.FreeMemory > 0 && (gpuInfo.Library == "cuda" || gpuInfo.Library == "rocm") {
-		return gpuInfo.FreeMemory, nil
-	}
-
-	return 0, fmt.Errorf("no GPU detected") // TODO - better handling of CPU based memory determiniation
-}
-
 func FindGPULibs(baseLibName string, patterns []string) []string {
 	// Multiple GPU libraries may exist, and some may not work, so keep trying until we exhaust them
 	var ldPaths []string
 	gpuLibPaths := []string{}
-	slog.Info(fmt.Sprintf("Searching for GPU management library %s", baseLibName))
+	slog.Debug("Searching for GPU library", "name", baseLibName)
 
 	switch runtime.GOOS {
 	case "windows":
@@ -283,7 +218,7 @@ func FindGPULibs(baseLibName string, patterns []string) []string {
 		}
 		patterns = append(patterns, filepath.Join(d, baseLibName+"*"))
 	}
-	slog.Debug(fmt.Sprintf("gpu management search paths: %v", patterns))
+	slog.Debug("gpu library search", "globs", patterns)
 	for _, pattern := range patterns {
 		// Ignore glob discovery errors
 		matches, _ := filepath.Glob(pattern)
@@ -311,28 +246,11 @@ func FindGPULibs(baseLibName string, patterns []string) []string {
 			}
 		}
 	}
-	slog.Info(fmt.Sprintf("Discovered GPU libraries: %v", gpuLibPaths))
+	slog.Debug("discovered GPU libraries", "paths", gpuLibPaths)
 	return gpuLibPaths
 }
 
-func LoadNVMLMgmt(nvmlLibPaths []string) *C.nvml_handle_t {
-	var resp C.nvml_init_resp_t
-	resp.ch.verbose = getVerboseState()
-	for _, libPath := range nvmlLibPaths {
-		lib := C.CString(libPath)
-		defer C.free(unsafe.Pointer(lib))
-		C.nvml_init(lib, &resp)
-		if resp.err != nil {
-			slog.Info(fmt.Sprintf("Unable to load NVML management library %s: %s", libPath, C.GoString(resp.err)))
-			C.free(unsafe.Pointer(resp.err))
-		} else {
-			return &resp.ch
-		}
-	}
-	return nil
-}
-
-func LoadCUDARTMgmt(cudartLibPaths []string) *C.cudart_handle_t {
+func LoadCUDARTMgmt(cudartLibPaths []string) (int, *C.cudart_handle_t, string) {
 	var resp C.cudart_init_resp_t
 	resp.ch.verbose = getVerboseState()
 	for _, libPath := range cudartLibPaths {
@@ -340,13 +258,13 @@ func LoadCUDARTMgmt(cudartLibPaths []string) *C.cudart_handle_t {
 		defer C.free(unsafe.Pointer(lib))
 		C.cudart_init(lib, &resp)
 		if resp.err != nil {
-			slog.Info(fmt.Sprintf("Unable to load cudart CUDA management library %s: %s", libPath, C.GoString(resp.err)))
+			slog.Debug("Unable to load cudart", "library", libPath, "error", C.GoString(resp.err))
 			C.free(unsafe.Pointer(resp.err))
 		} else {
-			return &resp.ch
+			return int(resp.num_devices), &resp.ch, libPath
 		}
 	}
-	return nil
+	return 0, nil, ""
 }
 
 func getVerboseState() C.uint16_t {
@@ -355,3 +273,22 @@ func getVerboseState() C.uint16_t {
 	}
 	return C.uint16_t(0)
 }
+
+// Given the list of GPUs this instantiation is targeted for,
+// figure out the visible devices environment variable
+//
+// If different libraries are detected, the first one is what we use
+func (l GpuInfoList) GetVisibleDevicesEnv() (string, string) {
+	if len(l) == 0 {
+		return "", ""
+	}
+	switch l[0].Library {
+	case "cuda":
+		return cudaGetVisibleDevicesEnv(l)
+	case "rocm":
+		return rocmGetVisibleDevicesEnv(l)
+	default:
+		slog.Debug("no filter required for library " + l[0].Library)
+		return "", ""
+	}
+}

gpu/gpu_darwin.go

@@ -9,52 +9,47 @@ package gpu
 */
 import "C"
 import (
-	"fmt"
-	"log/slog"
-	"os"
 	"runtime"
-	"strconv"
+
+	"github.com/ollama/ollama/format"
 )
 
-// CheckVRAM returns the free VRAM in bytes on Linux machines with NVIDIA GPUs
-func CheckVRAM() (uint64, error) {
-	userLimit := os.Getenv("OLLAMA_MAX_VRAM")
-	if userLimit != "" {
-		avail, err := strconv.ParseInt(userLimit, 10, 64)
-		if err != nil {
-			return 0, fmt.Errorf("Invalid OLLAMA_MAX_VRAM setting %s: %s", userLimit, err)
-		}
-		slog.Info(fmt.Sprintf("user override OLLAMA_MAX_VRAM=%d", avail))
-		return uint64(avail), nil
-	}
+const (
+	metalMinimumMemory = 512 * format.MebiByte
+)
 
+func GetGPUInfo() GpuInfoList {
+	mem, _ := GetCPUMem()
 	if runtime.GOARCH == "amd64" {
-		// gpu not supported, this may not be metal
-		return 0, nil
-	}
-
-	return uint64(C.getRecommendedMaxVRAM()), nil
-}
-
-func GetGPUInfo() GpuInfo {
-	mem, _ := getCPUMem()
-	if runtime.GOARCH == "amd64" {
-		return GpuInfo{
-			Library: "cpu",
-			Variant: GetCPUVariant(),
-			memInfo: mem,
+		return []GpuInfo{
+			{
+				Library: "cpu",
+				Variant: GetCPUVariant(),
+				memInfo: mem,
+			},
 		}
 	}
-	return GpuInfo{
+	info := GpuInfo{
 		Library: "metal",
-		memInfo: mem,
+		ID:      "0",
 	}
+	info.TotalMemory = uint64(C.getRecommendedMaxVRAM())
+
+	// TODO is there a way to gather actual allocated video memory? (currentAllocatedSize doesn't work)
+	info.FreeMemory = info.TotalMemory
+
+	info.MinimumMemory = metalMinimumMemory
+	return []GpuInfo{info}
 }
 
-func getCPUMem() (memInfo, error) {
+func GetCPUMem() (memInfo, error) {
 	return memInfo{
 		TotalMemory: uint64(C.getPhysicalMemory()),
 		FreeMemory:  0,
-		DeviceCount: 1,
 	}, nil
 }
+
+func (l GpuInfoList) GetVisibleDevicesEnv() (string, string) {
+	// No-op on darwin
+	return "", ""
+}

gpu/gpu_info.h

@@ -38,12 +38,17 @@
 extern "C" {
 #endif
 
+#define GPU_ID_LEN 64
+
 typedef struct mem_info {
+  char *err;  // If non-nill, caller responsible for freeing
+  char gpu_id[GPU_ID_LEN];
   uint64_t total;
   uint64_t free;
-  unsigned int count;
-  int igpu_index; // If >= 0, we detected an integrated GPU to ignore
-  char *err;  // If non-nill, caller responsible for freeing
+
+  // Compute Capability
+  int major; 
+  int minor;
 } mem_info_t;
 
 void cpu_check_ram(mem_info_t *resp);
@@ -52,7 +57,6 @@ void cpu_check_ram(mem_info_t *resp);
 }
 #endif
 
-#include "gpu_info_nvml.h"
 #include "gpu_info_cudart.h"
 
 #endif  // __GPU_INFO_H__

gpu/gpu_info_cpu.c

@@ -8,9 +8,11 @@ void cpu_check_ram(mem_info_t *resp) {
   MEMORYSTATUSEX info;
   info.dwLength = sizeof(info);
   if (GlobalMemoryStatusEx(&info) != 0) {
-    resp->count = 1;
     resp->total = info.ullTotalPhys;
     resp->free = info.ullAvailPhys;
+    resp->major = 0;
+    resp->minor = 0;
+    snprintf(&resp->gpu_id[0], GPU_ID_LEN, "0");
   } else {
     resp->err = LOAD_ERR();
   }
@@ -27,9 +29,11 @@ void cpu_check_ram(mem_info_t *resp) {
   if (sysinfo(&info) != 0) {
     resp->err = strdup(strerror(errno));
   } else {
-    resp->count = 1;
     resp->total = info.totalram * info.mem_unit;
     resp->free = info.freeram * info.mem_unit;
+    resp->major = 0;
+    resp->minor = 0;
+    snprintf(&resp->gpu_id[0], GPU_ID_LEN, "0");
   }
   return;
 }

gpu/gpu_info_cudart.c

@@ -6,6 +6,7 @@
 void cudart_init(char *cudart_lib_path, cudart_init_resp_t *resp) {
   cudartReturn_t ret;
   resp->err = NULL;
+  resp->num_devices = 0;
   const int buflen = 256;
   char buf[buflen + 1];
   int i;
@@ -21,6 +22,7 @@ void cudart_init(char *cudart_lib_path, cudart_init_resp_t *resp) {
       {"cudaGetDeviceCount", (void *)&resp->ch.cudaGetDeviceCount},
       {"cudaDeviceGetAttribute", (void *)&resp->ch.cudaDeviceGetAttribute},
       {"cudaDriverGetVersion", (void *)&resp->ch.cudaDriverGetVersion},
+      {"cudaGetDeviceProperties", (void *)&resp->ch.cudaGetDeviceProperties},
       {NULL, NULL},
   };
 
@@ -36,13 +38,7 @@ void cudart_init(char *cudart_lib_path, cudart_init_resp_t *resp) {
     return;
   }
 
-  // TODO once we've squashed the remaining corner cases remove this log
-  LOG(resp->ch.verbose, "wiring cudart library functions in %s\n", cudart_lib_path);
-  
   for (i = 0; l[i].s != NULL; i++) {
-    // TODO once we've squashed the remaining corner cases remove this log
-    LOG(resp->ch.verbose, "dlsym: %s\n", l[i].s);
-
     *l[i].p = LOAD_SYMBOL(resp->ch.handle, l[i].s);
     if (!l[i].p) {
       char *msg = LOAD_ERR();
@@ -63,7 +59,7 @@ void cudart_init(char *cudart_lib_path, cudart_init_resp_t *resp) {
     UNLOAD_LIBRARY(resp->ch.handle);
     resp->ch.handle = NULL;
     if (ret == CUDA_ERROR_INSUFFICIENT_DRIVER) {
-      resp->err = strdup("your nvidia driver is too old or missing, please upgrade to run ollama");
+      resp->err = strdup("your nvidia driver is too old or missing.  If you have a CUDA GPU please upgrade to run ollama");
       return;
     }
     snprintf(buf, buflen, "cudart init failure: %d", ret);
@@ -85,110 +81,95 @@ void cudart_init(char *cudart_lib_path, cudart_init_resp_t *resp) {
     driverVersion.minor = (version - (driverVersion.major * 1000)) / 10;
     LOG(resp->ch.verbose, "CUDA driver version: %d-%d\n", driverVersion.major, driverVersion.minor);
   }
+
+  ret = (*resp->ch.cudaGetDeviceCount)(&resp->num_devices);
+  if (ret != CUDART_SUCCESS) {
+    LOG(resp->ch.verbose, "cudaGetDeviceCount err: %d\n", ret);
+    UNLOAD_LIBRARY(resp->ch.handle);
+    resp->ch.handle = NULL;
+    snprintf(buf, buflen, "unable to get device count: %d", ret);
+    resp->err = strdup(buf);
+    return;
+  }
 }
 
 
-void cudart_check_vram(cudart_handle_t h, mem_info_t *resp) {
+void cudart_check_vram(cudart_handle_t h, int i, mem_info_t *resp) {
   resp->err = NULL;
   cudartMemory_t memInfo = {0,0,0};
   cudartReturn_t ret;
   const int buflen = 256;
   char buf[buflen + 1];
-  int i;
 
   if (h.handle == NULL) {
     resp->err = strdup("cudart handle isn't initialized");
     return;
   }
 
-  // cudaGetDeviceCount takes int type, resp-> count is uint
-  int deviceCount;
-  ret = (*h.cudaGetDeviceCount)(&deviceCount);
+  ret = (*h.cudaSetDevice)(i);
   if (ret != CUDART_SUCCESS) {
-    snprintf(buf, buflen, "unable to get device count: %d", ret);
+    snprintf(buf, buflen, "cudart device failed to initialize");
     resp->err = strdup(buf);
     return;
+  }
+
+  cudaDeviceProp_t props;
+  ret = (*h.cudaGetDeviceProperties)(&props, i);
+  if (ret != CUDART_SUCCESS) {
+    LOG(h.verbose, "[%d] device properties lookup failure: %d\n", i, ret);
+    snprintf(&resp->gpu_id[0], GPU_ID_LEN, "%d", i);
+    resp->major = 0;
+    resp->minor = 0;
   } else {
-    resp->count = (unsigned int)deviceCount;
-  }
-
-  resp->total = 0;
-  resp->free = 0;
-  for (i = 0; i < resp-> count; i++) {  
-    ret = (*h.cudaSetDevice)(i);
-    if (ret != CUDART_SUCCESS) {
-      snprintf(buf, buflen, "cudart device failed to initialize");
-      resp->err = strdup(buf);
-      return;
+    int allNull = 1;
+    for (int j = 0; j < 16; j++) {
+      if (props.uuid.bytes[j] != 0) {
+        allNull = 0;
+        break;
+      }
     }
-    ret = (*h.cudaMemGetInfo)(&memInfo.free, &memInfo.total);
-    if (ret != CUDART_SUCCESS) {
-      snprintf(buf, buflen, "cudart device memory info lookup failure %d", ret);
-      resp->err = strdup(buf);
-      return;
+    if (allNull != 0) {
+      snprintf(&resp->gpu_id[0], GPU_ID_LEN, "%d", i);
+    } else {
+      // GPU-d110a105-ac29-1d54-7b49-9c90440f215b
+      snprintf(&resp->gpu_id[0], GPU_ID_LEN,
+          "GPU-%02x%02x%02x%02x-%02x%02x-%02x%02x-%02x%02x-%02x%02x%02x%02x%02x%02x",
+          props.uuid.bytes[0],
+          props.uuid.bytes[1],
+          props.uuid.bytes[2],
+          props.uuid.bytes[3],
+          props.uuid.bytes[4],
+          props.uuid.bytes[5],
+          props.uuid.bytes[6],
+          props.uuid.bytes[7],
+          props.uuid.bytes[8],
+          props.uuid.bytes[9],
+          props.uuid.bytes[10],
+          props.uuid.bytes[11],
+          props.uuid.bytes[12],
+          props.uuid.bytes[13],
+          props.uuid.bytes[14],
+          props.uuid.bytes[15]
+        );
     }
+    resp->major = props.major;
+    resp->minor = props.minor;
 
-    LOG(h.verbose, "[%d] CUDA totalMem %lu\n", i, memInfo.total);
-    LOG(h.verbose, "[%d] CUDA freeMem %lu\n", i, memInfo.free);
-
-    resp->total += memInfo.total;
-    resp->free += memInfo.free;
+    // TODO add other useful properties from props
   }
-}
-
-void cudart_compute_capability(cudart_handle_t h, cudart_compute_capability_t *resp) {
-  resp->err = NULL;
-  resp->major = 0;
-  resp->minor = 0;
-  int major = 0;
-  int minor = 0;
-  cudartReturn_t ret;
-  const int buflen = 256;
-  char buf[buflen + 1];
-  int i;
-
-  if (h.handle == NULL) {
-    resp->err = strdup("cudart handle not initialized");
-    return;
-  }
-
-  int devices;
-  ret = (*h.cudaGetDeviceCount)(&devices);
+  ret = (*h.cudaMemGetInfo)(&memInfo.free, &memInfo.total);
   if (ret != CUDART_SUCCESS) {
-    snprintf(buf, buflen, "unable to get cudart device count: %d", ret);
+    snprintf(buf, buflen, "cudart device memory info lookup failure %d", ret);
     resp->err = strdup(buf);
     return;
   }
 
-  for (i = 0; i < devices; i++) {
-    ret = (*h.cudaSetDevice)(i);
-    if (ret != CUDART_SUCCESS) {
-      snprintf(buf, buflen, "cudart device failed to initialize");
-      resp->err = strdup(buf);
-      return;
-    }
+  resp->total = memInfo.total;
+  resp->free = memInfo.free;
 
-    ret = (*h.cudaDeviceGetAttribute)(&major, cudartDevAttrComputeCapabilityMajor, i);
-    if (ret != CUDART_SUCCESS) {
-      snprintf(buf, buflen, "device compute capability lookup failure %d: %d", i, ret);
-      resp->err = strdup(buf);
-      return;
-    }
-    ret = (*h.cudaDeviceGetAttribute)(&minor, cudartDevAttrComputeCapabilityMinor, i);
-    if (ret != CUDART_SUCCESS) {
-      snprintf(buf, buflen, "device compute capability lookup failure %d: %d", i, ret);
-      resp->err = strdup(buf);
-      return;
-    }
-      
-    // Report the lowest major.minor we detect as that limits our compatibility
-    if (resp->major == 0 || resp->major > major ) {
-      resp->major = major;
-      resp->minor = minor;
-    } else if ( resp->major == major && resp->minor > minor ) {
-      resp->minor = minor;
-    }
-  }
+  LOG(h.verbose, "[%s] CUDA totalMem %lu\n", resp->gpu_id, resp->total);
+  LOG(h.verbose, "[%s] CUDA freeMem %lu\n", resp->gpu_id, resp->free);
+  LOG(h.verbose, "[%s] Compute Capability %d.%d\n", resp->gpu_id, resp->major, resp->minor);
 }
 
 void cudart_release(cudart_handle_t h) {

gpu/gpu_info_cudart.h

@@ -6,7 +6,8 @@
 // Just enough typedef's to dlopen/dlsym for memory information
 typedef enum cudartReturn_enum {
   CUDART_SUCCESS = 0,
-  CUDART_UNSUPPORTED = 1,
+  CUDA_ERROR_INVALID_VALUE = 1,
+  CUDA_ERROR_MEMORY_ALLOCATION = 2,
   CUDA_ERROR_INSUFFICIENT_DRIVER = 35,
   // Other values omitted for now...
 } cudartReturn_t;
@@ -14,6 +15,11 @@ typedef enum cudartReturn_enum {
 typedef enum cudartDeviceAttr_enum {
   cudartDevAttrComputeCapabilityMajor = 75,
   cudartDevAttrComputeCapabilityMinor = 76,
+
+  // TODO - not yet wired up but may be useful for Jetson or other
+  // integrated GPU scenarios with shared memory
+  cudaDevAttrIntegrated = 18
+
 } cudartDeviceAttr_t;
 
 typedef void *cudartDevice_t;  // Opaque is sufficient
@@ -28,6 +34,92 @@ typedef struct cudartDriverVersion {
   int minor;
 } cudartDriverVersion_t;
 
+typedef struct cudaUUID {
+    unsigned char bytes[16];
+} cudaUUID_t;
+typedef struct cudaDeviceProp {
+    char         name[256];                  /**< ASCII string identifying device */
+    cudaUUID_t   uuid;                       /**< 16-byte unique identifier */
+    char         luid[8];                    /**< 8-byte locally unique identifier. Value is undefined on TCC and non-Windows platforms */
+    unsigned int luidDeviceNodeMask;         /**< LUID device node mask. Value is undefined on TCC and non-Windows platforms */
+    size_t       totalGlobalMem;             /**< Global memory available on device in bytes */
+    size_t       sharedMemPerBlock;          /**< Shared memory available per block in bytes */
+    int          regsPerBlock;               /**< 32-bit registers available per block */
+    int          warpSize;                   /**< Warp size in threads */
+    size_t       memPitch;                   /**< Maximum pitch in bytes allowed by memory copies */
+    int          maxThreadsPerBlock;         /**< Maximum number of threads per block */
+    int          maxThreadsDim[3];           /**< Maximum size of each dimension of a block */
+    int          maxGridSize[3];             /**< Maximum size of each dimension of a grid */
+    int          clockRate;                  /**< Clock frequency in kilohertz */
+    size_t       totalConstMem;              /**< Constant memory available on device in bytes */
+    int          major;                      /**< Major compute capability */
+    int          minor;                      /**< Minor compute capability */
+    size_t       textureAlignment;           /**< Alignment requirement for textures */
+    size_t       texturePitchAlignment;      /**< Pitch alignment requirement for texture references bound to pitched memory */
+    int          deviceOverlap;              /**< Device can concurrently copy memory and execute a kernel. Deprecated. Use instead asyncEngineCount. */
+    int          multiProcessorCount;        /**< Number of multiprocessors on device */
+    int          kernelExecTimeoutEnabled;   /**< Specified whether there is a run time limit on kernels */
+    int          integrated;                 /**< Device is integrated as opposed to discrete */
+    int          canMapHostMemory;           /**< Device can map host memory with cudaHostAlloc/cudaHostGetDevicePointer */
+    int          computeMode;                /**< Compute mode (See ::cudaComputeMode) */
+    int          maxTexture1D;               /**< Maximum 1D texture size */
+    int          maxTexture1DMipmap;         /**< Maximum 1D mipmapped texture size */
+    int          maxTexture1DLinear;         /**< Deprecated, do not use. Use cudaDeviceGetTexture1DLinearMaxWidth() or cuDeviceGetTexture1DLinearMaxWidth() instead. */
+    int          maxTexture2D[2];            /**< Maximum 2D texture dimensions */
+    int          maxTexture2DMipmap[2];      /**< Maximum 2D mipmapped texture dimensions */
+    int          maxTexture2DLinear[3];      /**< Maximum dimensions (width, height, pitch) for 2D textures bound to pitched memory */
+    int          maxTexture2DGather[2];      /**< Maximum 2D texture dimensions if texture gather operations have to be performed */
+    int          maxTexture3D[3];            /**< Maximum 3D texture dimensions */
+    int          maxTexture3DAlt[3];         /**< Maximum alternate 3D texture dimensions */
+    int          maxTextureCubemap;          /**< Maximum Cubemap texture dimensions */
+    int          maxTexture1DLayered[2];     /**< Maximum 1D layered texture dimensions */
+    int          maxTexture2DLayered[3];     /**< Maximum 2D layered texture dimensions */
+    int          maxTextureCubemapLayered[2];/**< Maximum Cubemap layered texture dimensions */
+    int          maxSurface1D;               /**< Maximum 1D surface size */
+    int          maxSurface2D[2];            /**< Maximum 2D surface dimensions */
+    int          maxSurface3D[3];            /**< Maximum 3D surface dimensions */
+    int          maxSurface1DLayered[2];     /**< Maximum 1D layered surface dimensions */
+    int          maxSurface2DLayered[3];     /**< Maximum 2D layered surface dimensions */
+    int          maxSurfaceCubemap;          /**< Maximum Cubemap surface dimensions */
+    int          maxSurfaceCubemapLayered[2];/**< Maximum Cubemap layered surface dimensions */
+    size_t       surfaceAlignment;           /**< Alignment requirements for surfaces */
+    int          concurrentKernels;          /**< Device can possibly execute multiple kernels concurrently */
+    int          ECCEnabled;                 /**< Device has ECC support enabled */
+    int          pciBusID;                   /**< PCI bus ID of the device */
+    int          pciDeviceID;                /**< PCI device ID of the device */
+    int          pciDomainID;                /**< PCI domain ID of the device */
+    int          tccDriver;                  /**< 1 if device is a Tesla device using TCC driver, 0 otherwise */
+    int          asyncEngineCount;           /**< Number of asynchronous engines */
+    int          unifiedAddressing;          /**< Device shares a unified address space with the host */
+    int          memoryClockRate;            /**< Peak memory clock frequency in kilohertz */
+    int          memoryBusWidth;             /**< Global memory bus width in bits */
+    int          l2CacheSize;                /**< Size of L2 cache in bytes */
+    int          persistingL2CacheMaxSize;   /**< Device's maximum l2 persisting lines capacity setting in bytes */
+    int          maxThreadsPerMultiProcessor;/**< Maximum resident threads per multiprocessor */
+    int          streamPrioritiesSupported;  /**< Device supports stream priorities */
+    int          globalL1CacheSupported;     /**< Device supports caching globals in L1 */
+    int          localL1CacheSupported;      /**< Device supports caching locals in L1 */
+    size_t       sharedMemPerMultiprocessor; /**< Shared memory available per multiprocessor in bytes */
+    int          regsPerMultiprocessor;      /**< 32-bit registers available per multiprocessor */
+    int          managedMemory;              /**< Device supports allocating managed memory on this system */
+    int          isMultiGpuBoard;            /**< Device is on a multi-GPU board */
+    int          multiGpuBoardGroupID;       /**< Unique identifier for a group of devices on the same multi-GPU board */
+    int          hostNativeAtomicSupported;  /**< Link between the device and the host supports native atomic operations */
+    int          singleToDoublePrecisionPerfRatio; /**< Ratio of single precision performance (in floating-point operations per second) to double precision performance */
+    int          pageableMemoryAccess;       /**< Device supports coherently accessing pageable memory without calling cudaHostRegister on it */
+    int          concurrentManagedAccess;    /**< Device can coherently access managed memory concurrently with the CPU */
+    int          computePreemptionSupported; /**< Device supports Compute Preemption */
+    int          canUseHostPointerForRegisteredMem; /**< Device can access host registered memory at the same virtual address as the CPU */
+    int          cooperativeLaunch;          /**< Device supports launching cooperative kernels via ::cudaLaunchCooperativeKernel */
+    int          cooperativeMultiDeviceLaunch; /**< Deprecated, cudaLaunchCooperativeKernelMultiDevice is deprecated. */
+    size_t       sharedMemPerBlockOptin;     /**< Per device maximum shared memory per block usable by special opt in */
+    int          pageableMemoryAccessUsesHostPageTables; /**< Device accesses pageable memory via the host's page tables */
+    int          directManagedMemAccessFromHost; /**< Host can directly access managed memory on the device without migration. */
+    int          maxBlocksPerMultiProcessor; /**< Maximum number of resident blocks per multiprocessor */
+    int          accessPolicyMaxWindowSize;  /**< The maximum value of ::cudaAccessPolicyWindow::num_bytes. */
+    size_t       reservedSharedMemPerBlock;  /**< Shared memory reserved by CUDA driver per block in bytes */
+  } cudaDeviceProp_t;
+
 typedef struct cudart_handle {
   void *handle;
   uint16_t verbose;
@@ -38,23 +130,17 @@ typedef struct cudart_handle {
   cudartReturn_t (*cudaGetDeviceCount)(int *);
   cudartReturn_t (*cudaDeviceGetAttribute)(int* value, cudartDeviceAttr_t attr, int device);
   cudartReturn_t (*cudaDriverGetVersion) (int *driverVersion);
+  cudartReturn_t (*cudaGetDeviceProperties) (cudaDeviceProp_t* prop, int device);
 } cudart_handle_t;
 
 typedef struct cudart_init_resp {
   char *err;  // If err is non-null handle is invalid
   cudart_handle_t ch;
+  int num_devices;
 } cudart_init_resp_t;
 
-typedef struct cudart_compute_capability {
-  char *err;
-  int major;
-  int minor;
-} cudart_compute_capability_t;
-
-
 void cudart_init(char *cudart_lib_path, cudart_init_resp_t *resp);
-void cudart_check_vram(cudart_handle_t ch, mem_info_t *resp);
-void cudart_compute_capability(cudart_handle_t th, cudart_compute_capability_t *cc);
+void cudart_check_vram(cudart_handle_t ch, int device_id, mem_info_t *resp);
 void cudart_release(cudart_handle_t ch);
 
 #endif  // __GPU_INFO_CUDART_H__

gpu/gpu_info_nvml.c

@@ -1,221 +0,0 @@
-#ifndef __APPLE__  // TODO - maybe consider nvidia support on intel macs?
-
-#include <string.h>
-
-#include "gpu_info_nvml.h"
-
-void nvml_init(char *nvml_lib_path, nvml_init_resp_t *resp) {
-  nvmlReturn_t ret;
-  resp->err = NULL;
-  const int buflen = 256;
-  char buf[buflen + 1];
-  int i;
-
-  struct lookup {
-    char *s;
-    void **p;
-  } l[] = {
-      {"nvmlInit_v2", (void *)&resp->ch.nvmlInit_v2},
-      {"nvmlShutdown", (void *)&resp->ch.nvmlShutdown},
-      {"nvmlDeviceGetHandleByIndex", (void *)&resp->ch.nvmlDeviceGetHandleByIndex},
-      {"nvmlDeviceGetMemoryInfo", (void *)&resp->ch.nvmlDeviceGetMemoryInfo},
-      {"nvmlDeviceGetCount_v2", (void *)&resp->ch.nvmlDeviceGetCount_v2},
-      {"nvmlDeviceGetCudaComputeCapability", (void *)&resp->ch.nvmlDeviceGetCudaComputeCapability},
-      {"nvmlSystemGetDriverVersion", (void *)&resp->ch.nvmlSystemGetDriverVersion},
-      {"nvmlDeviceGetName", (void *)&resp->ch.nvmlDeviceGetName},
-      {"nvmlDeviceGetSerial", (void *)&resp->ch.nvmlDeviceGetSerial},
-      {"nvmlDeviceGetVbiosVersion", (void *)&resp->ch.nvmlDeviceGetVbiosVersion},
-      {"nvmlDeviceGetBoardPartNumber", (void *)&resp->ch.nvmlDeviceGetBoardPartNumber},
-      {"nvmlDeviceGetBrand", (void *)&resp->ch.nvmlDeviceGetBrand},
-      {NULL, NULL},
-  };
-
-  resp->ch.handle = LOAD_LIBRARY(nvml_lib_path, RTLD_LAZY);
-  if (!resp->ch.handle) {
-    char *msg = LOAD_ERR();
-    LOG(resp->ch.verbose, "library %s load err: %s\n", nvml_lib_path, msg);
-    snprintf(buf, buflen,
-             "Unable to load %s library to query for Nvidia GPUs: %s",
-             nvml_lib_path, msg);
-    free(msg);
-    resp->err = strdup(buf);
-    return;
-  }
-
-  // TODO once we've squashed the remaining corner cases remove this log
-  LOG(resp->ch.verbose, "wiring nvidia management library functions in %s\n", nvml_lib_path);
-  
-  for (i = 0; l[i].s != NULL; i++) {
-    // TODO once we've squashed the remaining corner cases remove this log
-    LOG(resp->ch.verbose, "dlsym: %s\n", l[i].s);
-
-    *l[i].p = LOAD_SYMBOL(resp->ch.handle, l[i].s);
-    if (!l[i].p) {
-      resp->ch.handle = NULL;
-      char *msg = LOAD_ERR();
-      LOG(resp->ch.verbose, "dlerr: %s\n", msg);
-      UNLOAD_LIBRARY(resp->ch.handle);
-      snprintf(buf, buflen, "symbol lookup for %s failed: %s", l[i].s,
-               msg);
-      free(msg);
-      resp->err = strdup(buf);
-      return;
-    }
-  }
-
-  ret = (*resp->ch.nvmlInit_v2)();
-  if (ret != NVML_SUCCESS) {
-    LOG(resp->ch.verbose, "nvmlInit_v2 err: %d\n", ret);
-    UNLOAD_LIBRARY(resp->ch.handle);
-    resp->ch.handle = NULL;
-    snprintf(buf, buflen, "nvml vram init failure: %d", ret);
-    resp->err = strdup(buf);
-    return;
-  }
-
-  // Report driver version if we're in verbose mode, ignore errors
-  ret = (*resp->ch.nvmlSystemGetDriverVersion)(buf, buflen);
-  if (ret != NVML_SUCCESS) {
-    LOG(resp->ch.verbose, "nvmlSystemGetDriverVersion failed: %d\n", ret);
-  } else {
-    LOG(resp->ch.verbose, "CUDA driver version: %s\n", buf);
-  }
-}
-
-void nvml_check_vram(nvml_handle_t h, mem_info_t *resp) {
-  resp->err = NULL;
-  nvmlDevice_t device;
-  nvmlMemory_t memInfo = {0};
-  nvmlReturn_t ret;
-  const int buflen = 256;
-  char buf[buflen + 1];
-  int i;
-
-  if (h.handle == NULL) {
-    resp->err = strdup("nvml handle isn't initialized");
-    return;
-  }
-
-  ret = (*h.nvmlDeviceGetCount_v2)(&resp->count);
-  if (ret != NVML_SUCCESS) {
-    snprintf(buf, buflen, "unable to get device count: %d", ret);
-    resp->err = strdup(buf);
-    return;
-  }
-
-  resp->total = 0;
-  resp->free = 0;
-  for (i = 0; i < resp->count; i++) {
-    ret = (*h.nvmlDeviceGetHandleByIndex)(i, &device);
-    if (ret != NVML_SUCCESS) {
-      snprintf(buf, buflen, "unable to get device handle %d: %d", i, ret);
-      resp->err = strdup(buf);
-      return;
-    }
-
-    ret = (*h.nvmlDeviceGetMemoryInfo)(device, &memInfo);
-    if (ret != NVML_SUCCESS) {
-      snprintf(buf, buflen, "device memory info lookup failure %d: %d", i, ret);
-      resp->err = strdup(buf);
-      return;
-    }
-    if (h.verbose) {
-      nvmlBrandType_t brand = 0;
-      // When in verbose mode, report more information about
-      // the card we discover, but don't fail on error
-      ret = (*h.nvmlDeviceGetName)(device, buf, buflen);
-      if (ret != NVML_SUCCESS) {
-        LOG(h.verbose, "nvmlDeviceGetName failed: %d\n", ret);
-      } else {
-        LOG(h.verbose, "[%d] CUDA device name: %s\n", i, buf);
-      }
-      ret = (*h.nvmlDeviceGetBoardPartNumber)(device, buf, buflen);
-      if (ret != NVML_SUCCESS) {
-        LOG(h.verbose, "nvmlDeviceGetBoardPartNumber failed: %d\n", ret);
-      } else {
-        LOG(h.verbose, "[%d] CUDA part number: %s\n", i, buf);
-      }
-      ret = (*h.nvmlDeviceGetSerial)(device, buf, buflen);
-      if (ret != NVML_SUCCESS) {
-        LOG(h.verbose, "nvmlDeviceGetSerial failed: %d\n", ret);
-      } else {
-        LOG(h.verbose, "[%d] CUDA S/N: %s\n", i, buf);
-      }
-      ret = (*h.nvmlDeviceGetVbiosVersion)(device, buf, buflen);
-      if (ret != NVML_SUCCESS) {
-        LOG(h.verbose, "nvmlDeviceGetVbiosVersion failed: %d\n", ret);
-      } else {
-        LOG(h.verbose, "[%d] CUDA vbios version: %s\n", i, buf);
-      }
-      ret = (*h.nvmlDeviceGetBrand)(device, &brand);
-      if (ret != NVML_SUCCESS) {
-        LOG(h.verbose, "nvmlDeviceGetBrand failed: %d\n", ret);
-      } else {
-        LOG(h.verbose, "[%d] CUDA brand: %d\n", i, brand);
-      }
-    }
-
-    LOG(h.verbose, "[%d] CUDA totalMem %ld\n", i, memInfo.total);
-    LOG(h.verbose, "[%d] CUDA freeMem %ld\n", i, memInfo.free);
-
-    resp->total += memInfo.total;
-    resp->free += memInfo.free;
-  }
-}
-
-void nvml_compute_capability(nvml_handle_t h, nvml_compute_capability_t *resp) {
-  resp->err = NULL;
-  resp->major = 0;
-  resp->minor = 0;
-  nvmlDevice_t device;
-  int major = 0;
-  int minor = 0;
-  nvmlReturn_t ret;
-  const int buflen = 256;
-  char buf[buflen + 1];
-  int i;
-
-  if (h.handle == NULL) {
-    resp->err = strdup("nvml handle not initialized");
-    return;
-  }
-
-  unsigned int devices;
-  ret = (*h.nvmlDeviceGetCount_v2)(&devices);
-  if (ret != NVML_SUCCESS) {
-    snprintf(buf, buflen, "unable to get device count: %d", ret);
-    resp->err = strdup(buf);
-    return;
-  }
-
-  for (i = 0; i < devices; i++) {
-    ret = (*h.nvmlDeviceGetHandleByIndex)(i, &device);
-    if (ret != NVML_SUCCESS) {
-      snprintf(buf, buflen, "unable to get device handle %d: %d", i, ret);
-      resp->err = strdup(buf);
-      return;
-    }
-
-    ret = (*h.nvmlDeviceGetCudaComputeCapability)(device, &major, &minor);
-    if (ret != NVML_SUCCESS) {
-      snprintf(buf, buflen, "device compute capability lookup failure %d: %d", i, ret);
-      resp->err = strdup(buf);
-      return;
-    }
-    // Report the lowest major.minor we detect as that limits our compatibility
-    if (resp->major == 0 || resp->major > major ) {
-      resp->major = major;
-      resp->minor = minor;
-    } else if ( resp->major == major && resp->minor > minor ) {
-      resp->minor = minor;
-    }
-  }
-}
-
-void nvml_release(nvml_handle_t h) {
-  LOG(h.verbose, "releasing nvml library\n");
-  UNLOAD_LIBRARY(h.handle);
-  h.handle = NULL;
-}
-
-#endif  // __APPLE__

gpu/gpu_info_nvml.h

@@ -1,57 +0,0 @@
-#ifndef __APPLE__
-#ifndef __GPU_INFO_NVML_H__
-#define __GPU_INFO_NVML_H__
-#include "gpu_info.h"
-
-// Just enough typedef's to dlopen/dlsym for memory information
-typedef enum nvmlReturn_enum {
-  NVML_SUCCESS = 0,
-  // Other values omitted for now...
-} nvmlReturn_t;
-typedef void *nvmlDevice_t;  // Opaque is sufficient
-typedef struct nvmlMemory_st {
-  unsigned long long total;
-  unsigned long long free;
-  unsigned long long used;
-} nvmlMemory_t;
-
-typedef enum nvmlBrandType_enum
-{
-    NVML_BRAND_UNKNOWN          = 0,
-} nvmlBrandType_t;
-
-typedef struct nvml_handle {
-  void *handle;
-  uint16_t verbose;
-  nvmlReturn_t (*nvmlInit_v2)(void);
-  nvmlReturn_t (*nvmlShutdown)(void);
-  nvmlReturn_t (*nvmlDeviceGetHandleByIndex)(unsigned int, nvmlDevice_t *);
-  nvmlReturn_t (*nvmlDeviceGetMemoryInfo)(nvmlDevice_t, nvmlMemory_t *);
-  nvmlReturn_t (*nvmlDeviceGetCount_v2)(unsigned int *);
-  nvmlReturn_t (*nvmlDeviceGetCudaComputeCapability)(nvmlDevice_t, int* major, int* minor);
-  nvmlReturn_t (*nvmlSystemGetDriverVersion) (char* version, unsigned int  length);
-  nvmlReturn_t (*nvmlDeviceGetName) (nvmlDevice_t device, char* name, unsigned int  length);
-  nvmlReturn_t (*nvmlDeviceGetSerial) (nvmlDevice_t device, char* serial, unsigned int  length);
-  nvmlReturn_t (*nvmlDeviceGetVbiosVersion) (nvmlDevice_t device, char* version, unsigned int  length);
-  nvmlReturn_t (*nvmlDeviceGetBoardPartNumber) (nvmlDevice_t device, char* partNumber, unsigned int  length);
-  nvmlReturn_t (*nvmlDeviceGetBrand) (nvmlDevice_t device, nvmlBrandType_t* type);
-} nvml_handle_t;
-
-typedef struct nvml_init_resp {
-  char *err;  // If err is non-null handle is invalid
-  nvml_handle_t ch;
-} nvml_init_resp_t;
-
-typedef struct nvml_compute_capability {
-  char *err;
-  int major;
-  int minor;
-} nvml_compute_capability_t;
-
-void nvml_init(char *nvml_lib_path, nvml_init_resp_t *resp);
-void nvml_check_vram(nvml_handle_t ch, mem_info_t *resp);
-void nvml_compute_capability(nvml_handle_t ch, nvml_compute_capability_t *cc);
-void nvml_release(nvml_handle_t ch);
-
-#endif  // __GPU_INFO_NVML_H__
-#endif  // __APPLE__

gpu/gpu_test.go

@@ -9,23 +9,16 @@ import (
 
 func TestBasicGetGPUInfo(t *testing.T) {
 	info := GetGPUInfo()
-	assert.Contains(t, "cuda rocm cpu metal", info.Library)
-
-	switch runtime.GOOS {
-	case "darwin":
-		// TODO - remove this once MacOS returns some size for CPU
-		return
-	case "linux", "windows":
-		assert.Greater(t, info.TotalMemory, uint64(0))
-		assert.Greater(t, info.FreeMemory, uint64(0))
-		assert.Greater(t, info.DeviceCount, uint32(0))
-	default:
-		return
+	assert.Greater(t, len(info), 0)
+	assert.Contains(t, "cuda rocm cpu metal", info[0].Library)
+	if info[0].Library != "cpu" {
+		assert.Greater(t, info[0].TotalMemory, uint64(0))
+		assert.Greater(t, info[0].FreeMemory, uint64(0))
 	}
 }
 
 func TestCPUMemInfo(t *testing.T) {
-	info, err := getCPUMem()
+	info, err := GetCPUMem()
 	assert.NoError(t, err)
 	switch runtime.GOOS {
 	case "darwin":

gpu/types.go

@@ -3,7 +3,6 @@ package gpu
 type memInfo struct {
 	TotalMemory uint64 `json:"total_memory,omitempty"`
 	FreeMemory  uint64 `json:"free_memory,omitempty"`
-	DeviceCount uint32 `json:"device_count,omitempty"`
 }
 
 // Beginning of an `ollama info` command
@@ -17,11 +16,49 @@ type GpuInfo struct {
 	// MinimumMemory represents the minimum memory required to use the GPU
 	MinimumMemory uint64 `json:"-"`
 
-	// TODO add other useful attributes about the card here for discovery information
+	// Any extra PATH/LD_LIBRARY_PATH dependencies required for the Library to operate properly
+	DependencyPath string `json:"lib_path,omitempty"`
+
+	// GPU information
+	ID    string `json:"gpu_id"`          // string to use for selection of this specific GPU
+	Name  string `json:"name"`            // user friendly name if available
+	Major int    `json:"major,omitempty"` // Major compatibility version (CC or gfx)
+	Minor int    `json:"minor,omitempty"` // Minor compatibility version (CC or gfx)
+	Patch int    `json:"patch,omitempty"` // Patch compatibility only matters on AMD
+
+	// TODO other performance capability info to help in scheduling decisions
 }
 
-type Version struct {
-	Major uint
-	Minor uint
-	Patch uint
+type GpuInfoList []GpuInfo
+
+// Split up the set of gpu info's by Library and variant
+func (l GpuInfoList) ByLibrary() []GpuInfoList {
+	resp := []GpuInfoList{}
+	libs := []string{}
+	for _, info := range l {
+		found := false
+		requested := info.Library
+		if info.Variant != "" {
+			requested += "_" + info.Variant
+		}
+		for i, lib := range libs {
+			if lib == requested {
+				resp[i] = append(resp[i], info)
+				found = true
+				break
+			}
+		}
+		if !found {
+			libs = append(libs, info.Library)
+			resp = append(resp, []GpuInfo{info})
+		}
+	}
+	return resp
 }
+
+// Sort by Free Space
+type ByFreeMemory []GpuInfo
+
+func (a ByFreeMemory) Len() int           { return len(a) }
+func (a ByFreeMemory) Swap(i, j int)      { a[i], a[j] = a[j], a[i] }
+func (a ByFreeMemory) Less(i, j int) bool { return a[i].FreeMemory < a[j].FreeMemory }

integration/basic_test.go

@@ -4,11 +4,14 @@ package integration
 
 import (
 	"context"
-	"net/http"
+	"log/slog"
+	"os"
+	"runtime"
 	"testing"
 	"time"
 
 	"github.com/ollama/ollama/api"
+	"github.com/stretchr/testify/require"
 )
 
 func TestOrcaMiniBlueSky(t *testing.T) {
@@ -24,5 +27,44 @@ func TestOrcaMiniBlueSky(t *testing.T) {
 			"seed":        123,
 		},
 	}
-	GenerateTestHelper(ctx, t, &http.Client{}, req, []string{"rayleigh", "scattering"})
+	GenerateTestHelper(ctx, t, req, []string{"rayleigh", "scattering"})
+}
+
+func TestUnicodeModelDir(t *testing.T) {
+	// This is only useful for Windows with utf-16 characters, so skip this test for other platforms
+	if runtime.GOOS != "windows" {
+		t.Skip("Unicode test only applicable to windows")
+	}
+	// Only works for local testing
+	if os.Getenv("OLLAMA_TEST_EXISTING") != "" {
+		t.Skip("TestUnicodeModelDir only works for local testing, skipping")
+	}
+
+	modelDir, err := os.MkdirTemp("", "ollama_埃")
+	require.NoError(t, err)
+	defer os.RemoveAll(modelDir)
+	slog.Info("unicode", "OLLAMA_MODELS", modelDir)
+
+	oldModelsDir := os.Getenv("OLLAMA_MODELS")
+	if oldModelsDir == "" {
+		defer os.Unsetenv("OLLAMA_MODELS")
+	} else {
+		defer os.Setenv("OLLAMA_MODELS", oldModelsDir)
+	}
+	err = os.Setenv("OLLAMA_MODELS", modelDir)
+	require.NoError(t, err)
+
+	ctx, cancel := context.WithTimeout(context.Background(), 2*time.Minute)
+	defer cancel()
+
+	req := api.GenerateRequest{
+		Model:  "orca-mini",
+		Prompt: "why is the sky blue?",
+		Stream: &stream,
+		Options: map[string]interface{}{
+			"temperature": 0,
+			"seed":        123,
+		},
+	}
+	GenerateTestHelper(ctx, t, req, []string{"rayleigh", "scattering"})
 }

integration/concurrency_test.go

@@ -0,0 +1,225 @@
+//go:build integration
+
+package integration
+
+import (
+	"context"
+	"log/slog"
+	"os"
+	"strconv"
+	"sync"
+	"testing"
+	"time"
+
+	"github.com/ollama/ollama/api"
+	"github.com/stretchr/testify/require"
+)
+
+func TestMultiModelConcurrency(t *testing.T) {
+	var (
+		req = [2]api.GenerateRequest{
+			{
+				Model:  "orca-mini",
+				Prompt: "why is the ocean blue?",
+				Stream: &stream,
+				Options: map[string]interface{}{
+					"seed":        42,
+					"temperature": 0.0,
+				},
+			}, {
+				Model:  "tinydolphin",
+				Prompt: "what is the origin of the us thanksgiving holiday?",
+				Stream: &stream,
+				Options: map[string]interface{}{
+					"seed":        42,
+					"temperature": 0.0,
+				},
+			},
+		}
+		resp = [2][]string{
+			[]string{"sunlight"},
+			[]string{"england", "english", "massachusetts", "pilgrims"},
+		}
+	)
+	var wg sync.WaitGroup
+	wg.Add(len(req))
+	ctx, cancel := context.WithTimeout(context.Background(), time.Second*120)
+	defer cancel()
+	for i := 0; i < len(req); i++ {
+		go func(i int) {
+			defer wg.Done()
+			GenerateTestHelper(ctx, t, req[i], resp[i])
+		}(i)
+	}
+	wg.Wait()
+}
+
+func TestIntegrationConcurrentPredictOrcaMini(t *testing.T) {
+	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Minute) // GTX 750 2G card takes ~9 minutes
+	defer cancel()
+	client, _, cleanup := InitServerConnection(ctx, t)
+	defer cleanup()
+
+	req, resp := GenerateRequests()
+	// Get the server running (if applicable) warm the model up with a single initial request
+	DoGenerate(ctx, t, client, req[0], resp[0], 60*time.Second, 5*time.Second)
+
+	var wg sync.WaitGroup
+	wg.Add(len(req))
+	for i := 0; i < len(req); i++ {
+		go func(i int) {
+			defer wg.Done()
+			for j := 0; j < 5; j++ {
+				slog.Info("Starting", "req", i, "iter", j)
+				// On slower GPUs it can take a while to process the 4 concurrent requests
+				// so we allow a much longer initial timeout
+				DoGenerate(ctx, t, client, req[i], resp[i], 90*time.Second, 5*time.Second)
+			}
+		}(i)
+	}
+	wg.Wait()
+}
+
+// Stress the system if we know how much VRAM it has, and attempt to load more models than will fit
+func TestMultiModelStress(t *testing.T) {
+	vram := os.Getenv("OLLAMA_MAX_VRAM")
+	if vram == "" {
+		t.Skip("OLLAMA_MAX_VRAM not specified, can't pick the right models for the stress test")
+	}
+	max, err := strconv.ParseUint(vram, 10, 64)
+	require.NoError(t, err)
+	const MB = uint64(1024 * 1024)
+	type model struct {
+		name string
+		size uint64 // Approximate amount of VRAM they typically use when fully loaded in VRAM
+	}
+
+	smallModels := []model{
+		{
+			name: "orca-mini",
+			size: 2992 * MB,
+		},
+		{
+			name: "phi",
+			size: 2616 * MB,
+		},
+		{
+			name: "gemma:2b",
+			size: 2364 * MB,
+		},
+		{
+			name: "stable-code:3b",
+			size: 2608 * MB,
+		},
+		{
+			name: "starcoder2:3b",
+			size: 2166 * MB,
+		},
+	}
+	mediumModels := []model{
+		{
+			name: "llama2",
+			size: 5118 * MB,
+		},
+		{
+			name: "mistral",
+			size: 4620 * MB,
+		},
+		{
+			name: "orca-mini:7b",
+			size: 5118 * MB,
+		},
+		{
+			name: "dolphin-mistral",
+			size: 4620 * MB,
+		},
+		{
+			name: "gemma:7b",
+			size: 5000 * MB,
+		},
+		// TODO - uncomment this once #3565 is merged and this is rebased on it
+		// {
+		// 	name: "codellama:7b",
+		// 	size: 5118 * MB,
+		// },
+	}
+
+	// These seem to be too slow to be useful...
+	// largeModels := []model{
+	// 	{
+	// 		name: "llama2:13b",
+	// 		size: 7400 * MB,
+	// 	},
+	// 	{
+	// 		name: "codellama:13b",
+	// 		size: 7400 * MB,
+	// 	},
+	// 	{
+	// 		name: "orca-mini:13b",
+	// 		size: 7400 * MB,
+	// 	},
+	// 	{
+	// 		name: "gemma:7b",
+	// 		size: 5000 * MB,
+	// 	},
+	// 	{
+	// 		name: "starcoder2:15b",
+	// 		size: 9100 * MB,
+	// 	},
+	// }
+
+	var chosenModels []model
+	switch {
+	case max < 10000*MB:
+		slog.Info("selecting small models")
+		chosenModels = smallModels
+	// case max < 30000*MB:
+	default:
+		slog.Info("selecting medium models")
+		chosenModels = mediumModels
+		// default:
+		// 	slog.Info("selecting large models")
+		// 	chosenModels = largModels
+	}
+
+	req, resp := GenerateRequests()
+
+	for i := range req {
+		if i > len(chosenModels) {
+			break
+		}
+		req[i].Model = chosenModels[i].name
+	}
+
+	ctx, cancel := context.WithTimeout(context.Background(), 15*time.Minute) // TODO baseline -- 10m too short
+	defer cancel()
+	client, _, cleanup := InitServerConnection(ctx, t)
+	defer cleanup()
+
+	// Make sure all the models are pulled before we get started
+	for _, r := range req {
+		require.NoError(t, PullIfMissing(ctx, client, r.Model))
+	}
+
+	var wg sync.WaitGroup
+	consumed := uint64(256 * MB) // Assume some baseline usage
+	for i := 0; i < len(req); i++ {
+		// Always get at least 2 models, but dont' overshoot VRAM too much or we'll take too long
+		if i > 1 && consumed > max {
+			slog.Info("achieved target vram exhaustion", "count", i, "vramMB", max/1024/1024, "modelsMB", consumed/1024/1024)
+			break
+		}
+		consumed += chosenModels[i].size
+		slog.Info("target vram", "count", i, "vramMB", max/1024/1024, "modelsMB", consumed/1024/1024)
+
+		wg.Add(1)
+		go func(i int) {
+			defer wg.Done()
+			for j := 0; j < 3; j++ {
+				slog.Info("Starting", "req", i, "iter", j, "model", req[i].Model)
+				DoGenerate(ctx, t, client, req[i], resp[i], 90*time.Second, 5*time.Second)
+			}
+		}(i)
+	}
+	wg.Wait()
+}

integration/context_test.go

@@ -4,7 +4,6 @@ package integration
 
 import (
 	"context"
-	"net/http"
 	"testing"
 	"time"
 
@@ -25,5 +24,5 @@ func TestContextExhaustion(t *testing.T) {
 			"num_ctx":     128,
 		},
 	}
-	GenerateTestHelper(ctx, t, &http.Client{}, req, []string{"once", "upon", "lived"})
+	GenerateTestHelper(ctx, t, req, []string{"once", "upon", "lived"})
 }

integration/llm_image_test.go

@@ -5,7 +5,6 @@ package integration
 import (
 	"context"
 	"encoding/base64"
-	"net/http"
 	"testing"
 	"time"
 
@@ -29,10 +28,11 @@ func TestIntegrationMultimodal(t *testing.T) {
 		},
 	}
 
-	resp := "the ollamas"
+	// Note: sometimes it returns "the ollamas" sometimes "the ollams"
+	resp := "the ollam"
 	ctx, cancel := context.WithTimeout(context.Background(), 3*time.Minute)
 	defer cancel()
-	GenerateTestHelper(ctx, t, &http.Client{}, req, []string{resp})
+	GenerateTestHelper(ctx, t, req, []string{resp})
 }
 
 const imageEncoding = `iVBORw0KGgoAAAANSUhEUgAAANIAAAB4CAYAAACHHqzKAAAAAXNSR0IArs4c6QAAAIRlWElmTU0AKgAAAAgABQESAAMAAAABAAEAAAEaAAUAAAABAAAASgEb

integration/llm_test.go

@@ -4,8 +4,6 @@ package integration
 
 import (
 	"context"
-	"net/http"
-	"sync"
 	"testing"
 	"time"
 
@@ -45,25 +43,5 @@ var (
 func TestIntegrationSimpleOrcaMini(t *testing.T) {
 	ctx, cancel := context.WithTimeout(context.Background(), time.Second*120)
 	defer cancel()
-	GenerateTestHelper(ctx, t, &http.Client{}, req[0], resp[0])
+	GenerateTestHelper(ctx, t, req[0], resp[0])
 }
-
-// TODO
-// The server always loads a new runner and closes the old one, which forces serial execution
-// At present this test case fails with concurrency problems.  Eventually we should try to
-// get true concurrency working with n_parallel support in the backend
-func TestIntegrationConcurrentPredictOrcaMini(t *testing.T) {
-	var wg sync.WaitGroup
-	wg.Add(len(req))
-	ctx, cancel := context.WithTimeout(context.Background(), time.Second*120)
-	defer cancel()
-	for i := 0; i < len(req); i++ {
-		go func(i int) {
-			defer wg.Done()
-			GenerateTestHelper(ctx, t, &http.Client{}, req[i], resp[i])
-		}(i)
-	}
-	wg.Wait()
-}
-
-// TODO - create a parallel test with 2 different models once we support concurrency

integration/utils_test.go

@@ -5,13 +5,14 @@ package integration
 import (
 	"bytes"
 	"context"
-	"encoding/json"
+	"errors"
 	"fmt"
 	"io"
 	"log/slog"
 	"math/rand"
 	"net"
 	"net/http"
+	"net/url"
 	"os"
 	"path/filepath"
 	"runtime"
@@ -23,9 +24,13 @@ import (
 
 	"github.com/ollama/ollama/api"
 	"github.com/ollama/ollama/app/lifecycle"
-	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
 )
 
+func Init() {
+	lifecycle.InitLogging()
+}
+
 func FindPort() string {
 	port := 0
 	if a, err := net.ResolveTCPAddr("tcp", "localhost:0"); err == nil {
@@ -41,7 +46,7 @@ func FindPort() string {
 	return strconv.Itoa(port)
 }
 
-func GetTestEndpoint() (string, string) {
+func GetTestEndpoint() (*api.Client, string) {
 	defaultPort := "11434"
 	ollamaHost := os.Getenv("OLLAMA_HOST")
 
@@ -67,16 +72,20 @@ func GetTestEndpoint() (string, string) {
 		port = FindPort()
 	}
 
-	url := fmt.Sprintf("%s:%s", host, port)
-	slog.Info("server connection", "url", url)
-	return scheme, url
+	slog.Info("server connection", "host", host, "port", port)
+
+	return api.NewClient(
+		&url.URL{
+			Scheme: scheme,
+			Host:   net.JoinHostPort(host, port),
+		},
+		http.DefaultClient), fmt.Sprintf("%s:%s", host, port)
 }
 
-// TODO make fanicier, grab logs, etc.
 var serverMutex sync.Mutex
 var serverReady bool
 
-func StartServer(ctx context.Context, ollamaHost string) error {
+func startServer(ctx context.Context, ollamaHost string) error {
 	// Make sure the server has been built
 	CLIName, err := filepath.Abs("../ollama")
 	if err != nil {
@@ -98,7 +107,7 @@ func StartServer(ctx context.Context, ollamaHost string) error {
 
 	if tmp := os.Getenv("OLLAMA_HOST"); tmp != ollamaHost {
 		slog.Info("setting env", "OLLAMA_HOST", ollamaHost)
-		os.Setenv("OLLAMA_HOST", ollamaHost)
+		t.Setenv("OLLAMA_HOST", ollamaHost)
 	}
 
 	slog.Info("starting server", "url", ollamaHost)
@@ -125,67 +134,76 @@ func StartServer(ctx context.Context, ollamaHost string) error {
 	return nil
 }
 
-func PullIfMissing(ctx context.Context, client *http.Client, scheme, testEndpoint, modelName string) error {
+func PullIfMissing(ctx context.Context, client *api.Client, modelName string) error {
 	slog.Info("checking status of model", "model", modelName)
 	showReq := &api.ShowRequest{Name: modelName}
-	requestJSON, err := json.Marshal(showReq)
-	if err != nil {
-		return err
-	}
 
-	req, err := http.NewRequest("POST", scheme+"://"+testEndpoint+"/api/show", bytes.NewReader(requestJSON))
-	if err != nil {
+	showCtx, cancel := context.WithDeadlineCause(
+		ctx,
+		time.Now().Add(5*time.Second),
+		fmt.Errorf("show for existing model %s took too long", modelName),
+	)
+	defer cancel()
+	_, err := client.Show(showCtx, showReq)
+	var statusError api.StatusError
+	switch {
+	case errors.As(err, &statusError) && statusError.StatusCode == http.StatusNotFound:
+		break
+	case err != nil:
 		return err
-	}
-
-	// Make the request with the HTTP client
-	response, err := client.Do(req.WithContext(ctx))
-	if err != nil {
-		return err
-	}
-	defer response.Body.Close()
-	if response.StatusCode == 200 {
+	default:
 		slog.Info("model already present", "model", modelName)
 		return nil
 	}
-	slog.Info("model missing", "status", response.StatusCode)
+	slog.Info("model missing", "model", modelName)
 
+	stallDuration := 30 * time.Second // This includes checksum verification, which can take a while on larger models
+	stallTimer := time.NewTimer(stallDuration)
+	fn := func(resp api.ProgressResponse) error {
+		// fmt.Print(".")
+		if !stallTimer.Reset(stallDuration) {
+			return fmt.Errorf("stall was detected, aborting status reporting")
+		}
+		return nil
+	}
+
+	stream := true
 	pullReq := &api.PullRequest{Name: modelName, Stream: &stream}
-	requestJSON, err = json.Marshal(pullReq)
-	if err != nil {
-		return err
-	}
 
-	req, err = http.NewRequest("POST", scheme+"://"+testEndpoint+"/api/pull", bytes.NewReader(requestJSON))
-	if err != nil {
-		return err
-	}
-	slog.Info("pulling", "model", modelName)
+	var pullError error
 
-	response, err = client.Do(req.WithContext(ctx))
-	if err != nil {
-		return err
+	done := make(chan int)
+	go func() {
+		pullError = client.Pull(ctx, pullReq, fn)
+		done <- 0
+	}()
+
+	select {
+	case <-stallTimer.C:
+		return fmt.Errorf("download stalled")
+	case <-done:
+		return pullError
 	}
-	defer response.Body.Close()
-	if response.StatusCode != 200 {
-		return fmt.Errorf("failed to pull model") // TODO more details perhaps
-	}
-	slog.Info("model pulled", "model", modelName)
-	return nil
 }
 
 var serverProcMutex sync.Mutex
 
-func GenerateTestHelper(ctx context.Context, t *testing.T, client *http.Client, genReq api.GenerateRequest, anyResp []string) {
-
-	// TODO maybe stuff in an init routine?
-	lifecycle.InitLogging()
-
-	requestJSON, err := json.Marshal(genReq)
-	if err != nil {
-		t.Fatalf("Error serializing request: %v", err)
+// Returns an Client, the testEndpoint, and a cleanup function, fails the test on errors
+// Starts the server if needed
+func InitServerConnection(ctx context.Context, t *testing.T) (*api.Client, string, func()) {
+	client, testEndpoint := GetTestEndpoint()
+	if os.Getenv("OLLAMA_TEST_EXISTING") == "" {
+		serverProcMutex.Lock()
+		fp, err := os.CreateTemp("", "ollama-server-*.log")
+		if err != nil {
+			t.Fatalf("failed to generate log file: %s", err)
+		}
+		lifecycle.ServerLogFile = fp.Name()
+		fp.Close()
+		require.NoError(t, startServer(ctx, testEndpoint))
 	}
-	defer func() {
+
+	return client, testEndpoint, func() {
 		if os.Getenv("OLLAMA_TEST_EXISTING") == "" {
 			defer serverProcMutex.Unlock()
 			if t.Failed() {
@@ -203,63 +221,118 @@ func GenerateTestHelper(ctx context.Context, t *testing.T, client *http.Client,
 				os.Stderr.Write(data)
 				slog.Warn("END OF SERVER")
 			}
-			err = os.Remove(lifecycle.ServerLogFile)
+			err := os.Remove(lifecycle.ServerLogFile)
 			if err != nil && !os.IsNotExist(err) {
 				slog.Warn("failed to cleanup", "logfile", lifecycle.ServerLogFile, "error", err)
 			}
 		}
-	}()
-	scheme, testEndpoint := GetTestEndpoint()
-
-	if os.Getenv("OLLAMA_TEST_EXISTING") == "" {
-		serverProcMutex.Lock()
-		fp, err := os.CreateTemp("", "ollama-server-*.log")
-		if err != nil {
-			t.Fatalf("failed to generate log file: %s", err)
-		}
-		lifecycle.ServerLogFile = fp.Name()
-		fp.Close()
-		assert.NoError(t, StartServer(ctx, testEndpoint))
 	}
-
-	err = PullIfMissing(ctx, client, scheme, testEndpoint, genReq.Model)
-	if err != nil {
-		t.Fatalf("Error pulling model: %v", err)
-	}
-
-	// Make the request and get the response
-	req, err := http.NewRequest("POST", scheme+"://"+testEndpoint+"/api/generate", bytes.NewReader(requestJSON))
-	if err != nil {
-		t.Fatalf("Error creating request: %v", err)
-	}
-
-	// Set the content type for the request
-	req.Header.Set("Content-Type", "application/json")
-
-	// Make the request with the HTTP client
-	response, err := client.Do(req.WithContext(ctx))
-	if err != nil {
-		t.Fatalf("Error making request: %v", err)
-	}
-	defer response.Body.Close()
-	body, err := io.ReadAll(response.Body)
-	assert.NoError(t, err)
-	assert.Equal(t, response.StatusCode, 200, string(body))
-
-	// Verify the response is valid JSON
-	var payload api.GenerateResponse
-	err = json.Unmarshal(body, &payload)
-	if err != nil {
-		assert.NoError(t, err, body)
-	}
-
-	// Verify the response contains the expected data
-	atLeastOne := false
-	for _, resp := range anyResp {
-		if strings.Contains(strings.ToLower(payload.Response), resp) {
-			atLeastOne = true
-			break
-		}
-	}
-	assert.True(t, atLeastOne, "none of %v found in %s", anyResp, payload.Response)
+}
+
+func GenerateTestHelper(ctx context.Context, t *testing.T, genReq api.GenerateRequest, anyResp []string) {
+	client, _, cleanup := InitServerConnection(ctx, t)
+	defer cleanup()
+	require.NoError(t, PullIfMissing(ctx, client, genReq.Model))
+	DoGenerate(ctx, t, client, genReq, anyResp, 30*time.Second, 10*time.Second)
+}
+
+func DoGenerate(ctx context.Context, t *testing.T, client *api.Client, genReq api.GenerateRequest, anyResp []string, initialTimeout, streamTimeout time.Duration) {
+	stallTimer := time.NewTimer(initialTimeout)
+	var buf bytes.Buffer
+	fn := func(response api.GenerateResponse) error {
+		// fmt.Print(".")
+		buf.Write([]byte(response.Response))
+		if !stallTimer.Reset(streamTimeout) {
+			return fmt.Errorf("stall was detected while streaming response, aborting")
+		}
+		return nil
+	}
+
+	stream := true
+	genReq.Stream = &stream
+	done := make(chan int)
+	var genErr error
+	go func() {
+		genErr = client.Generate(ctx, &genReq, fn)
+		done <- 0
+	}()
+
+	select {
+	case <-stallTimer.C:
+		if buf.Len() == 0 {
+			t.Errorf("generate never started.  Timed out after :%s", initialTimeout.String())
+		} else {
+			t.Errorf("generate stalled.  Response so far:%s", buf.String())
+		}
+	case <-done:
+		require.NoError(t, genErr, "failed with %s request prompt %s ", genReq.Model, genReq.Prompt)
+		// Verify the response contains the expected data
+		response := buf.String()
+		atLeastOne := false
+		for _, resp := range anyResp {
+			if strings.Contains(strings.ToLower(response), resp) {
+				atLeastOne = true
+				break
+			}
+		}
+		require.True(t, atLeastOne, "none of %v found in %s", anyResp, response)
+		slog.Info("test pass", "model", genReq.Model, "prompt", genReq.Prompt, "contains", anyResp, "response", response)
+	case <-ctx.Done():
+		t.Error("outer test context done while waiting for generate")
+	}
+}
+
+// Generate a set of requests
+// By default each request uses orca-mini as the model
+func GenerateRequests() ([]api.GenerateRequest, [][]string) {
+	return []api.GenerateRequest{
+			{
+				Model:  "orca-mini",
+				Prompt: "why is the ocean blue?",
+				Stream: &stream,
+				Options: map[string]interface{}{
+					"seed":        42,
+					"temperature": 0.0,
+				},
+			}, {
+				Model:  "orca-mini",
+				Prompt: "why is the color of dirt brown?",
+				Stream: &stream,
+				Options: map[string]interface{}{
+					"seed":        42,
+					"temperature": 0.0,
+				},
+			}, {
+				Model:  "orca-mini",
+				Prompt: "what is the origin of the us thanksgiving holiday?",
+				Stream: &stream,
+				Options: map[string]interface{}{
+					"seed":        42,
+					"temperature": 0.0,
+				},
+			}, {
+				Model:  "orca-mini",
+				Prompt: "what is the origin of independence day?",
+				Stream: &stream,
+				Options: map[string]interface{}{
+					"seed":        42,
+					"temperature": 0.0,
+				},
+			}, {
+				Model:  "orca-mini",
+				Prompt: "what is the composition of air?",
+				Stream: &stream,
+				Options: map[string]interface{}{
+					"seed":        42,
+					"temperature": 0.0,
+				},
+			},
+		},
+		[][]string{
+			[]string{"sunlight"},
+			[]string{"soil", "organic", "earth", "black", "tan"},
+			[]string{"england", "english", "massachusetts", "pilgrims"},
+			[]string{"fourth", "july", "declaration", "independence"},
+			[]string{"nitrogen", "oxygen", "carbon", "dioxide"},
+		}
 }

llm/ext_server/server.cpp

@@ -1032,7 +1032,7 @@ struct llama_server_context
             slot.has_next_token = false;
         }
 
-        if (!slot.cache_tokens.empty() && result.tok == llama_token_eos(model))
+        if (!slot.cache_tokens.empty() && llama_token_is_eog(model, result.tok))
         {
             slot.stopped_eos = true;
             slot.has_next_token = false;
@@ -1144,12 +1144,15 @@ struct llama_server_context
 
         res.result_json = json
         {
-            {"content",    tkn.text_to_send},
             {"stop",       false},
             {"slot_id",    slot.id},
             {"multimodal", multimodal}
         };
 
+        if (!llama_token_is_eog(model, tkn.tok)) {
+            res.result_json["content"] = tkn.text_to_send;
+        }
+
         if (slot.sparams.n_probs > 0)
         {
             std::vector<completion_token_output> probs_output = {};
@@ -2644,18 +2647,18 @@ static void server_params_parse(int argc, char **argv, server_params &sparams,
             if (strncmp(sep, "int:", 4) == 0) {
                 sep += 4;
                 kvo.tag = LLAMA_KV_OVERRIDE_TYPE_INT;
-                kvo.int_value = std::atol(sep);
+                kvo.val_i64 = std::atol(sep);
             } else if (strncmp(sep, "float:", 6) == 0) {
                 sep += 6;
                 kvo.tag = LLAMA_KV_OVERRIDE_TYPE_FLOAT;
-                kvo.float_value = std::atof(sep);
+                kvo.val_f64 = std::atof(sep);
             } else if (strncmp(sep, "bool:", 5) == 0) {
                 sep += 5;
                 kvo.tag = LLAMA_KV_OVERRIDE_TYPE_BOOL;
                 if (std::strcmp(sep, "true") == 0) {
-                    kvo.bool_value = true;
+                    kvo.val_bool = true;
                 } else if (std::strcmp(sep, "false") == 0) {
-                    kvo.bool_value = false;
+                    kvo.val_bool = false;
                 } else {
                     fprintf(stderr, "error: Invalid boolean value for KV override: %s\n", argv[i]);
                     invalid_param = true;

llm/filetype.go

@@ -0,0 +1,165 @@
+package llm
+
+import "fmt"
+
+type fileType uint32
+
+const (
+	fileTypeF32 fileType = iota
+	fileTypeF16
+	fileTypeQ4_0
+	fileTypeQ4_1
+	fileTypeQ4_1_F16
+	fileTypeQ4_2 // unused
+	fileTypeQ4_3 // unused
+	fileTypeQ8_0
+	fileTypeQ5_0
+	fileTypeQ5_1
+	fileTypeQ2_K
+	fileTypeQ3_K_S
+	fileTypeQ3_K_M
+	fileTypeQ3_K_L
+	fileTypeQ4_K_S
+	fileTypeQ4_K_M
+	fileTypeQ5_K_S
+	fileTypeQ5_K_M
+	fileTypeQ6_K
+	fileTypeIQ2_XXS
+	fileTypeIQ2_XS
+	fileTypeQ2_K_S
+	fileTypeQ3_K_XS
+	fileTypeIQ3_XXS
+	fileTypeIQ1_S
+	fileTypeIQ4_NL
+	fileTypeIQ3_S
+	fileTypeIQ2_S
+	fileTypeIQ4_XS
+
+	fileTypeUnknown
+)
+
+func ParseFileType(s string) (fileType, error) {
+	switch s {
+	case "F32":
+		return fileTypeF32, nil
+	case "F16":
+		return fileTypeF16, nil
+	case "Q4_0":
+		return fileTypeQ4_0, nil
+	case "Q4_1":
+		return fileTypeQ4_1, nil
+	case "Q4_1_F16":
+		return fileTypeQ4_1_F16, nil
+	case "Q8_0":
+		return fileTypeQ8_0, nil
+	case "Q5_0":
+		return fileTypeQ5_0, nil
+	case "Q5_1":
+		return fileTypeQ5_1, nil
+	case "Q2_K":
+		return fileTypeQ2_K, nil
+	case "Q3_K_S":
+		return fileTypeQ3_K_S, nil
+	case "Q3_K_M":
+		return fileTypeQ3_K_M, nil
+	case "Q3_K_L":
+		return fileTypeQ3_K_L, nil
+	case "Q4_K_S":
+		return fileTypeQ4_K_S, nil
+	case "Q4_K_M":
+		return fileTypeQ4_K_M, nil
+	case "Q5_K_S":
+		return fileTypeQ5_K_S, nil
+	case "Q5_K_M":
+		return fileTypeQ5_K_M, nil
+	case "Q6_K":
+		return fileTypeQ6_K, nil
+	case "IQ2_XXS":
+		return fileTypeIQ2_XXS, nil
+	case "IQ2_XS":
+		return fileTypeIQ2_XS, nil
+	case "Q2_K_S":
+		return fileTypeQ2_K_S, nil
+	case "Q3_K_XS":
+		return fileTypeQ3_K_XS, nil
+	case "IQ1_S":
+		return fileTypeIQ1_S, nil
+	case "IQ4_NL":
+		return fileTypeIQ4_NL, nil
+	case "IQ3_S":
+		return fileTypeIQ3_S, nil
+	case "IQ2_S":
+		return fileTypeIQ2_S, nil
+	case "IQ4_XS":
+		return fileTypeIQ4_XS, nil
+	case "IQ3_XXS":
+		return fileTypeIQ3_XXS, nil
+	default:
+		return fileTypeUnknown, fmt.Errorf("unknown fileType: %s", s)
+	}
+}
+
+func (t fileType) String() string {
+	switch t {
+	case fileTypeF32:
+		return "F32"
+	case fileTypeF16:
+		return "F16"
+	case fileTypeQ4_0:
+		return "Q4_0"
+	case fileTypeQ4_1:
+		return "Q4_1"
+	case fileTypeQ4_1_F16:
+		return "Q4_1_F16"
+	case fileTypeQ8_0:
+		return "Q8_0"
+	case fileTypeQ5_0:
+		return "Q5_0"
+	case fileTypeQ5_1:
+		return "Q5_1"
+	case fileTypeQ2_K:
+		return "Q2_K"
+	case fileTypeQ3_K_S:
+		return "Q3_K_S"
+	case fileTypeQ3_K_M:
+		return "Q3_K_M"
+	case fileTypeQ3_K_L:
+		return "Q3_K_L"
+	case fileTypeQ4_K_S:
+		return "Q4_K_S"
+	case fileTypeQ4_K_M:
+		return "Q4_K_M"
+	case fileTypeQ5_K_S:
+		return "Q5_K_S"
+	case fileTypeQ5_K_M:
+		return "Q5_K_M"
+	case fileTypeQ6_K:
+		return "Q6_K"
+	case fileTypeIQ2_XXS:
+		return "IQ2_XXS"
+	case fileTypeIQ2_XS:
+		return "IQ2_XS"
+	case fileTypeQ2_K_S:
+		return "Q2_K_S"
+	case fileTypeQ3_K_XS:
+		return "Q3_K_XS"
+	case fileTypeIQ1_S:
+		return "IQ1_S"
+	case fileTypeIQ4_NL:
+		return "IQ4_NL"
+	case fileTypeIQ3_S:
+		return "IQ3_S"
+	case fileTypeIQ2_S:
+		return "IQ2_S"
+	case fileTypeIQ4_XS:
+		return "IQ4_XS"
+	case fileTypeIQ3_XXS:
+		return "IQ3_XXS"
+	default:
+		return "unknown"
+	}
+}
+
+func (t fileType) Value() uint32 {
+	return uint32(t)
+}

llm/generate/gen_common.sh

@@ -21,7 +21,7 @@ init_vars() {
         # TODO - add additional optimization flags...
         CMAKE_DEFS="-DCMAKE_BUILD_TYPE=Release -DLLAMA_SERVER_VERBOSE=off ${CMAKE_DEFS}"
     fi
-    case $(uname -s) in 
+    case $(uname -s) in
     "Darwin")
         LIB_EXT="dylib"
         WHOLE_ARCHIVE="-Wl,-force_load"

llm/generate/gen_linux.sh

@@ -165,14 +165,22 @@ if [ -d "${CUDA_LIB_DIR}" ]; then
     fi
     if [ "${ARCH}" == "arm64" ]; then
         echo "ARM CPU detected - disabling unsupported AVX instructions"
-        
+
         # ARM-based CPUs such as M1 and Tegra do not support AVX extensions.
         #
-        # CUDA compute < 6.0 lacks proper FP16 support on ARM. 
-        # Disabling has minimal performance effect while maintaining compatibility. 
+        # CUDA compute < 6.0 lacks proper FP16 support on ARM.
+        # Disabling has minimal performance effect while maintaining compatibility.
         ARM64_DEFS="-DLLAMA_AVX=off -DLLAMA_AVX2=off -DLLAMA_AVX512=off -DLLAMA_CUDA_F16=off"
     fi
-    CMAKE_DEFS="-DLLAMA_CUDA=on -DLLAMA_CUDA_FORCE_MMQ=on -DCMAKE_CUDA_ARCHITECTURES=${CMAKE_CUDA_ARCHITECTURES} ${COMMON_CMAKE_DEFS} ${CMAKE_DEFS} ${ARM64_DEFS}"
+    # Users building from source can tune the exact flags we pass to cmake for configuring llama.cpp
+    if [ -n "${OLLAMA_CUSTOM_CUDA_DEFS}" ]; then
+        echo "OLLAMA_CUSTOM_CUDA_DEFS=\"${OLLAMA_CUSTOM_CUDA_DEFS}\""
+        CMAKE_CUDA_DEFS="-DLLAMA_CUDA=on -DCMAKE_CUDA_ARCHITECTURES=${CMAKE_CUDA_ARCHITECTURES} ${OLLAMA_CUSTOM_CUDA_DEFS}"
+        echo "Building custom CUDA GPU"
+    else
+        CMAKE_CUDA_DEFS="-DLLAMA_CUDA=on -DLLAMA_CUDA_FORCE_MMQ=on -DCMAKE_CUDA_ARCHITECTURES=${CMAKE_CUDA_ARCHITECTURES}"
+    fi
+    CMAKE_DEFS="${COMMON_CMAKE_DEFS} ${CMAKE_DEFS} ${ARM64_DEFS} ${CMAKE_CUDA_DEFS}"
     BUILD_DIR="../build/linux/${ARCH}/cuda${CUDA_VARIANT}"
     EXTRA_LIBS="-L${CUDA_LIB_DIR} -lcudart -lcublas -lcublasLt -lcuda"
     build
@@ -217,6 +225,12 @@ if [ -d "${ROCM_PATH}" ]; then
     fi
     init_vars
     CMAKE_DEFS="${COMMON_CMAKE_DEFS} ${CMAKE_DEFS} -DLLAMA_HIPBLAS=on -DCMAKE_C_COMPILER=$ROCM_PATH/llvm/bin/clang -DCMAKE_CXX_COMPILER=$ROCM_PATH/llvm/bin/clang++ -DAMDGPU_TARGETS=$(amdGPUs) -DGPU_TARGETS=$(amdGPUs)"
+    # Users building from source can tune the exact flags we pass to cmake for configuring llama.cpp
+    if [ -n "${OLLAMA_CUSTOM_ROCM_DEFS}" ]; then
+        echo "OLLAMA_CUSTOM_ROCM_DEFS=\"${OLLAMA_CUSTOM_ROCM_DEFS}\""
+        CMAKE_DEFS="${CMAKE_DEFS} ${OLLAMA_CUSTOM_ROCM_DEFS}"
+        echo "Building custom ROCM GPU"
+    fi
     BUILD_DIR="../build/linux/${ARCH}/rocm${ROCM_VARIANT}"
     EXTRA_LIBS="-L${ROCM_PATH}/lib -L/opt/amdgpu/lib/x86_64-linux-gnu/ -Wl,-rpath,\$ORIGIN/../../rocm/ -lhipblas -lrocblas -lamdhip64 -lrocsolver -lamd_comgr -lhsa-runtime64 -lrocsparse -ldrm -ldrm_amdgpu"
     build

llm/generate/gen_windows.ps1

@@ -26,15 +26,25 @@ function amdGPUs {
     $GPU_LIST -join ';'
 }
 
+
 function init_vars {
-    $script:SRC_DIR = $(resolve-path "..\..\")
-    $script:llamacppDir = "../llama.cpp"
+    if (!$script:SRC_DIR) {
+        $script:SRC_DIR = $(resolve-path "..\..\")
+    }
+    if (!$script:llamacppDir) {
+        $script:llamacppDir = "../llama.cpp"
+    }
+    if (!$script:cmakeTargets) {
+        $script:cmakeTargets = @("ollama_llama_server")
+    }
     $script:cmakeDefs = @(
         "-DBUILD_SHARED_LIBS=on",
         "-DLLAMA_NATIVE=off"
         )
-    $script:cmakeTargets = @("ollama_llama_server")
-    $script:ARCH = "amd64" # arm not yet supported.
+    $script:commonCpuDefs = @("-DCMAKE_POSITION_INDEPENDENT_CODE=on")
+    $script:ARCH = $Env:PROCESSOR_ARCHITECTURE.ToLower()
+    $script:DIST_BASE = "${script:SRC_DIR}\dist\windows-${script:ARCH}\ollama_runners"
+    md "$script:DIST_BASE" -ea 0 > $null
     if ($env:CGO_CFLAGS -contains "-g") {
         $script:cmakeDefs += @("-DCMAKE_VERBOSE_MAKEFILE=on", "-DLLAMA_SERVER_VERBOSE=on", "-DCMAKE_BUILD_TYPE=RelWithDebInfo")
         $script:config = "RelWithDebInfo"
@@ -55,7 +65,6 @@ function init_vars {
     } else {
         $script:CUDA_LIB_DIR=$env:CUDA_LIB_DIR
     }
-    $script:GZIP=(get-command -ea 'silentlycontinue' gzip).path
     $script:DUMPBIN=(get-command -ea 'silentlycontinue' dumpbin).path
     if ($null -eq $env:CMAKE_CUDA_ARCHITECTURES) {
         $script:CMAKE_CUDA_ARCHITECTURES="50;52;61;70;75;80"
@@ -134,21 +143,18 @@ function sign {
     }
 }
 
-function compress {
-    if ($script:GZIP -eq $null) {
-        write-host "gzip not installed, not compressing files"
-        return
-    }
-    write-host "Compressing binaries..."
+function install {
+    write-host "Installing binaries to dist dir ${script:distDir}"
+    mkdir ${script:distDir} -ErrorAction SilentlyContinue
     $binaries = dir "${script:buildDir}/bin/*.exe"
     foreach ($file in $binaries) {
-        & "$script:GZIP" --best -f $file
+        copy-item -Path $file -Destination ${script:distDir} -Force
     }
 
-    write-host "Compressing dlls..."
+    write-host "Installing dlls to dist dir ${script:distDir}"
     $dlls = dir "${script:buildDir}/bin/*.dll"
     foreach ($file in $dlls) {
-        & "$script:GZIP" --best -f $file
+        copy-item -Path $file -Destination ${script:distDir} -Force
     }
 }
 
@@ -169,123 +175,195 @@ function cleanup {
     }
 }
 
-init_vars
-git_module_setup
-apply_patches
 
 # -DLLAMA_AVX -- 2011 Intel Sandy Bridge & AMD Bulldozer
 # -DLLAMA_AVX2 -- 2013 Intel Haswell & 2015 AMD Excavator / 2017 AMD Zen
 # -DLLAMA_FMA (FMA3) -- 2013 Intel Haswell & 2012 AMD Piledriver
 
-$script:commonCpuDefs = @("-DCMAKE_POSITION_INDEPENDENT_CODE=on")
 
-if ($null -eq ${env:OLLAMA_SKIP_CPU_GENERATE}) {
+function build_static() {
+    if ((-not "${env:OLLAMA_SKIP_STATIC_GENERATE}") -and ((-not "${env:OLLAMA_CPU_TARGET}") -or ("${env:OLLAMA_CPU_TARGET}" -eq "static"))) {
+        # GCC build for direct linking into the Go binary
+        init_vars
+        # cmake will silently fallback to msvc compilers if mingw isn't in the path, so detect and fail fast
+        # as we need this to be compiled by gcc for golang to be able to link with itx
+        write-host "Checking for MinGW..."
+        # error action ensures we exit on failure
+        get-command gcc
+        get-command mingw32-make
+        $oldTargets = $script:cmakeTargets
+        $script:cmakeTargets = @("llama", "ggml")
+        $script:cmakeDefs = @(
+            "-G", "MinGW Makefiles"
+            "-DCMAKE_C_COMPILER=gcc.exe",
+            "-DCMAKE_CXX_COMPILER=g++.exe",
+            "-DBUILD_SHARED_LIBS=off",
+            "-DLLAMA_NATIVE=off",
+            "-DLLAMA_AVX=off",
+            "-DLLAMA_AVX2=off",
+            "-DLLAMA_AVX512=off",
+            "-DLLAMA_F16C=off",
+            "-DLLAMA_FMA=off")
+        $script:buildDir="../build/windows/${script:ARCH}_static"
+        write-host "Building static library"
+        build
+        $script:cmakeTargets = $oldTargets
+    } else {
+        write-host "Skipping CPU generation step as requested"
+    }
+}
+
+function build_cpu($gen_arch) {
+    if ((-not "${env:OLLAMA_SKIP_CPU_GENERATE}" ) -and ((-not "${env:OLLAMA_CPU_TARGET}") -or ("${env:OLLAMA_CPU_TARGET}" -eq "cpu"))) {
+        # remaining llama.cpp builds use MSVC 
+        init_vars
+        $script:cmakeDefs = $script:commonCpuDefs + @("-A", $gen_arch, "-DLLAMA_AVX=off", "-DLLAMA_AVX2=off", "-DLLAMA_AVX512=off", "-DLLAMA_FMA=off", "-DLLAMA_F16C=off") + $script:cmakeDefs
+        $script:buildDir="../build/windows/${script:ARCH}/cpu"
+        $script:distDir="$script:DIST_BASE\cpu"
+        write-host "Building LCD CPU"
+        build
+        sign
+        install
+    } else {
+        write-host "Skipping CPU generation step as requested"
+    }
+}
+
+function build_cpu_avx() {
+    if ((-not "${env:OLLAMA_SKIP_CPU_GENERATE}" ) -and ((-not "${env:OLLAMA_CPU_TARGET}") -or ("${env:OLLAMA_CPU_TARGET}" -eq "cpu_avx"))) {
+        init_vars
+        $script:cmakeDefs = $script:commonCpuDefs + @("-A", "x64", "-DLLAMA_AVX=on", "-DLLAMA_AVX2=off", "-DLLAMA_AVX512=off", "-DLLAMA_FMA=off", "-DLLAMA_F16C=off") + $script:cmakeDefs
+        $script:buildDir="../build/windows/${script:ARCH}/cpu_avx"
+        $script:distDir="$script:DIST_BASE\cpu_avx"
+        write-host "Building AVX CPU"
+        build
+        sign
+        install
+    } else {
+        write-host "Skipping CPU AVX generation step as requested"
+    }
+}
+
+function build_cpu_avx2() {
+    if ((-not "${env:OLLAMA_SKIP_CPU_GENERATE}" ) -and ((-not "${env:OLLAMA_CPU_TARGET}") -or ("${env:OLLAMA_CPU_TARGET}" -eq "cpu_avx2"))) {
+        init_vars
+        $script:cmakeDefs = $script:commonCpuDefs + @("-A", "x64", "-DLLAMA_AVX=on", "-DLLAMA_AVX2=on", "-DLLAMA_AVX512=off", "-DLLAMA_FMA=on", "-DLLAMA_F16C=on") + $script:cmakeDefs
+        $script:buildDir="../build/windows/${script:ARCH}/cpu_avx2"
+        $script:distDir="$script:DIST_BASE\cpu_avx2"
+        write-host "Building AVX2 CPU"
+        build
+        sign
+        install
+    } else {
+        write-host "Skipping CPU AVX2 generation step as requested"
+    }
+}
+
+function build_cuda() {
+    if ((-not "${env:OLLAMA_SKIP_CUDA_GENERATE}") -and ("${script:CUDA_LIB_DIR}")) {
+        # Then build cuda as a dynamically loaded library
+        $nvcc = "$script:CUDA_LIB_DIR\nvcc.exe"
+        $script:CUDA_VERSION=(get-item ($nvcc | split-path | split-path)).Basename
+        if ($null -ne $script:CUDA_VERSION) {
+            $script:CUDA_VARIANT="_"+$script:CUDA_VERSION
+        }
+        init_vars
+        $script:buildDir="../build/windows/${script:ARCH}/cuda$script:CUDA_VARIANT"
+        $script:distDir="$script:DIST_BASE\cuda$script:CUDA_VARIANT"
+        $script:cmakeDefs += @("-A", "x64", "-DLLAMA_CUDA=ON", "-DLLAMA_AVX=on", "-DLLAMA_AVX2=off", "-DCUDAToolkit_INCLUDE_DIR=$script:CUDA_INCLUDE_DIR", "-DCMAKE_CUDA_ARCHITECTURES=${script:CMAKE_CUDA_ARCHITECTURES}")
+        if ($null -ne $env:OLLAMA_CUSTOM_CUDA_DEFS) {
+            write-host "OLLAMA_CUSTOM_CUDA_DEFS=`"${env:OLLAMA_CUSTOM_CUDA_DEFS}`""
+            $script:cmakeDefs +=@("${env:OLLAMA_CUSTOM_CUDA_DEFS}")
+            write-host "building custom CUDA GPU"
+        }
+        build
+        sign
+        install
+
+        write-host "copying CUDA dependencies to ${script:SRC_DIR}\dist\windows-${script:ARCH}\"
+        cp "${script:CUDA_LIB_DIR}\cudart64_*.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\"
+        cp "${script:CUDA_LIB_DIR}\cublas64_*.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\"
+        cp "${script:CUDA_LIB_DIR}\cublasLt64_*.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\"
+    } else {
+        write-host "Skipping CUDA generation step"
+    }
+}
+
+function build_rocm() {
+    if ((-not "${env:OLLAMA_SKIP_ROCM_GENERATE}") -and ("${env:HIP_PATH}")) {
+        $script:ROCM_VERSION=(get-item $env:HIP_PATH).Basename
+        if ($null -ne $script:ROCM_VERSION) {
+            $script:ROCM_VARIANT="_v"+$script:ROCM_VERSION
+        }
+
+        init_vars
+        $script:buildDir="../build/windows/${script:ARCH}/rocm$script:ROCM_VARIANT"
+        $script:distDir="$script:DIST_BASE\rocm$script:ROCM_VARIANT"
+        $script:cmakeDefs += @(
+            "-G", "Ninja", 
+            "-DCMAKE_C_COMPILER=clang.exe",
+            "-DCMAKE_CXX_COMPILER=clang++.exe",
+            "-DLLAMA_HIPBLAS=on",
+            "-DHIP_PLATFORM=amd",
+            "-DLLAMA_AVX=on",
+            "-DLLAMA_AVX2=off",
+            "-DCMAKE_POSITION_INDEPENDENT_CODE=on",
+            "-DAMDGPU_TARGETS=$(amdGPUs)",
+            "-DGPU_TARGETS=$(amdGPUs)"
+            )
+
+        # Make sure the ROCm binary dir is first in the path
+        $env:PATH="$env:HIP_PATH\bin;$env:PATH"
+
+        # We have to clobber the LIB var from the developer shell for clang to work properly
+        $env:LIB=""
+        if ($null -ne $env:OLLAMA_CUSTOM_ROCM_DEFS) {
+            write-host "OLLAMA_CUSTOM_ROCM_DEFS=`"${env:OLLAMA_CUSTOM_ROCM_DEFS}`""
+            $script:cmakeDefs += @("${env:OLLAMA_CUSTOM_ROCM_DEFS}")
+            write-host "building custom ROCM GPU"
+        }
+        write-host "Building ROCm"
+        build
+        # Ninja doesn't prefix with config name
+        ${script:config}=""
+        if ($null -ne $script:DUMPBIN) {
+            & "$script:DUMPBIN" /dependents "${script:buildDir}/bin/ollama_llama_server.exe" | select-string ".dll"
+        }
+        sign
+        install
+
+        # Assumes v5.7, may need adjustments for v6
+        rm -ea 0 -recurse -force -path "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\"
+        md "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\rocblas\library\" -ea 0 > $null
+        cp "${env:HIP_PATH}\bin\hipblas.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\"
+        cp "${env:HIP_PATH}\bin\rocblas.dll" "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\"
+        # amdhip64.dll dependency comes from the driver and must be installed on the host to use AMD GPUs
+        cp "${env:HIP_PATH}\bin\rocblas\library\*" "${script:SRC_DIR}\dist\windows-${script:ARCH}\rocm\rocblas\library\"
+    } else {
+        write-host "Skipping ROCm generation step"
+    }
+}
 
-# GCC build for direct linking into the Go binary
 init_vars
-# cmake will silently fallback to msvc compilers if mingw isn't in the path, so detect and fail fast
-# as we need this to be compiled by gcc for golang to be able to link with itx
-write-host "Checking for MinGW..."
-# error action ensures we exit on failure
-get-command gcc
-get-command mingw32-make
-$script:cmakeTargets = @("llama", "ggml")
-$script:cmakeDefs = @(
-    "-G", "MinGW Makefiles"
-    "-DCMAKE_C_COMPILER=gcc.exe",
-    "-DCMAKE_CXX_COMPILER=g++.exe",
-    "-DBUILD_SHARED_LIBS=off",
-    "-DLLAMA_NATIVE=off",
-    "-DLLAMA_AVX=off",
-    "-DLLAMA_AVX2=off",
-    "-DLLAMA_AVX512=off",
-    "-DLLAMA_F16C=off",
-    "-DLLAMA_FMA=off")
-$script:buildDir="../build/windows/${script:ARCH}_static"
-write-host "Building static library"
-build
+if ($($args.count) -eq 0) {
+    git_module_setup
+    apply_patches
+    build_static
+    if ($script:ARCH -eq "arm64") {
+        build_cpu("ARM64")
+    } else { # amd64
+        build_cpu("x64")
+        build_cpu_avx
+        build_cpu_avx2
+        build_cuda
+        build_rocm
+    }
 
-# remaining llama.cpp builds use MSVC 
-    init_vars
-    $script:cmakeDefs = $script:commonCpuDefs + @("-A", "x64", "-DLLAMA_AVX=off", "-DLLAMA_AVX2=off", "-DLLAMA_AVX512=off", "-DLLAMA_FMA=off", "-DLLAMA_F16C=off") + $script:cmakeDefs
-    $script:buildDir="../build/windows/${script:ARCH}/cpu"
-    write-host "Building LCD CPU"
-    build
-    sign
-    compress
-
-    init_vars
-    $script:cmakeDefs = $script:commonCpuDefs + @("-A", "x64", "-DLLAMA_AVX=on", "-DLLAMA_AVX2=off", "-DLLAMA_AVX512=off", "-DLLAMA_FMA=off", "-DLLAMA_F16C=off") + $script:cmakeDefs
-    $script:buildDir="../build/windows/${script:ARCH}/cpu_avx"
-    write-host "Building AVX CPU"
-    build
-    sign
-    compress
-
-    init_vars
-    $script:cmakeDefs = $script:commonCpuDefs + @("-A", "x64", "-DLLAMA_AVX=on", "-DLLAMA_AVX2=on", "-DLLAMA_AVX512=off", "-DLLAMA_FMA=on", "-DLLAMA_F16C=on") + $script:cmakeDefs
-    $script:buildDir="../build/windows/${script:ARCH}/cpu_avx2"
-    write-host "Building AVX2 CPU"
-    build
-    sign
-    compress
+    cleanup
+    write-host "`ngo generate completed.  LLM runners: $(get-childitem -path $script:DIST_BASE)"
 } else {
-    write-host "Skipping CPU generation step as requested"
-}
-
-if ($null -ne $script:CUDA_LIB_DIR) {
-    # Then build cuda as a dynamically loaded library
-    $nvcc = "$script:CUDA_LIB_DIR\nvcc.exe"
-    $script:CUDA_VERSION=(get-item ($nvcc | split-path | split-path)).Basename
-    if ($null -ne $script:CUDA_VERSION) {
-        $script:CUDA_VARIANT="_"+$script:CUDA_VERSION
-    }
-    init_vars
-    $script:buildDir="../build/windows/${script:ARCH}/cuda$script:CUDA_VARIANT"
-    $script:cmakeDefs += @("-A", "x64", "-DLLAMA_CUDA=ON", "-DLLAMA_AVX=on", "-DLLAMA_AVX2=off", "-DCUDAToolkit_INCLUDE_DIR=$script:CUDA_INCLUDE_DIR", "-DCMAKE_CUDA_ARCHITECTURES=${script:CMAKE_CUDA_ARCHITECTURES}")
-    build
-    sign
-    compress
-}
-
-if ($null -ne $env:HIP_PATH) {
-    $script:ROCM_VERSION=(get-item $env:HIP_PATH).Basename
-    if ($null -ne $script:ROCM_VERSION) {
-        $script:ROCM_VARIANT="_v"+$script:ROCM_VERSION
-    }
-
-    init_vars
-    $script:buildDir="../build/windows/${script:ARCH}/rocm$script:ROCM_VARIANT"
-    $script:cmakeDefs += @(
-        "-G", "Ninja", 
-        "-DCMAKE_C_COMPILER=clang.exe",
-        "-DCMAKE_CXX_COMPILER=clang++.exe",
-        "-DLLAMA_HIPBLAS=on",
-        "-DHIP_PLATFORM=amd",
-        "-DLLAMA_AVX=on",
-        "-DLLAMA_AVX2=off",
-        "-DCMAKE_POSITION_INDEPENDENT_CODE=on",
-        "-DAMDGPU_TARGETS=$(amdGPUs)",
-        "-DGPU_TARGETS=$(amdGPUs)"
-        )
-
-    # Make sure the ROCm binary dir is first in the path
-    $env:PATH="$env:HIP_PATH\bin;$env:PATH"
-
-    # We have to clobber the LIB var from the developer shell for clang to work properly
-    $env:LIB=""
-
-    write-host "Building ROCm"
-    build
-    # Ninja doesn't prefix with config name
-    ${script:config}=""
-    if ($null -ne $script:DUMPBIN) {
-        & "$script:DUMPBIN" /dependents "${script:buildDir}/bin/ollama_llama_server.exe" | select-string ".dll"
-    }
-    sign
-    compress
-}
-
-
-cleanup
-write-host "`ngo generate completed.  LLM runners: $(get-childitem -path ${script:SRC_DIR}\llm\build\windows\${script:ARCH})"
+    for ( $i = 0; $i -lt $args.count; $i++ ) {
+        write-host "performing $($args[$i])"
+        & $($args[$i])
+    } 
+}

llm/ggml.go

@@ -13,82 +13,6 @@ type GGML struct {
 	model
 }
 
-const (
-	fileTypeF32 uint32 = iota
-	fileTypeF16
-	fileTypeQ4_0
-	fileTypeQ4_1
-	fileTypeQ4_1_F16
-	fileTypeQ8_0 uint32 = iota + 2
-	fileTypeQ5_0
-	fileTypeQ5_1
-	fileTypeQ2_K
-	fileTypeQ3_K_S
-	fileTypeQ3_K_M
-	fileTypeQ3_K_L
-	fileTypeQ4_K_S
-	fileTypeQ4_K_M
-	fileTypeQ5_K_S
-	fileTypeQ5_K_M
-	fileTypeQ6_K
-	fileTypeIQ2_XXS
-	fileTypeIQ2_XS
-	fileTypeQ2_K_S
-	fileTypeQ3_K_XS
-	fileTypeIQ3_XXS
-)
-
-func fileType(fileType uint32) string {
-	switch fileType {
-	case fileTypeF32:
-		return "F32"
-	case fileTypeF16:
-		return "F16"
-	case fileTypeQ4_0:
-		return "Q4_0"
-	case fileTypeQ4_1:
-		return "Q4_1"
-	case fileTypeQ4_1_F16:
-		return "Q4_1_F16"
-	case fileTypeQ8_0:
-		return "Q8_0"
-	case fileTypeQ5_0:
-		return "Q5_0"
-	case fileTypeQ5_1:
-		return "Q5_1"
-	case fileTypeQ2_K:
-		return "Q2_K"
-	case fileTypeQ3_K_S:
-		return "Q3_K_S"
-	case fileTypeQ3_K_M:
-		return "Q3_K_M"
-	case fileTypeQ3_K_L:
-		return "Q3_K_L"
-	case fileTypeQ4_K_S:
-		return "Q4_K_S"
-	case fileTypeQ4_K_M:
-		return "Q4_K_M"
-	case fileTypeQ5_K_S:
-		return "Q5_K_S"
-	case fileTypeQ5_K_M:
-		return "Q5_K_M"
-	case fileTypeQ6_K:
-		return "Q6_K"
-	case fileTypeIQ2_XXS:
-		return "IQ2_XXS"
-	case fileTypeIQ2_XS:
-		return "IQ2_XS"
-	case fileTypeQ2_K_S:
-		return "Q2_K_S"
-	case fileTypeQ3_K_XS:
-		return "Q3_K_XS"
-	case fileTypeIQ3_XXS:
-		return "IQ3_XXS"
-	default:
-		return "unknown"
-	}
-}
-
 type model interface {
 	KV() KV
 	Tensors() Tensors
@@ -123,7 +47,7 @@ func (kv KV) ParameterCount() uint64 {
 
 func (kv KV) FileType() string {
 	if u64 := kv.u64("general.file_type"); u64 > 0 {
-		return fileType(uint32(u64))
+		return fileType(uint32(u64)).String()
 	}
 
 	return "unknown"
@@ -201,9 +125,9 @@ type Tensor struct {
 
 func (t Tensor) blockSize() uint64 {
 	switch {
-	case t.Kind < 2:
+	case t.Kind < 2 || (t.Kind > 23 && t.Kind < 29):
 		return 1
-	case t.Kind < 10:
+	case t.Kind < 10 || t.Kind == 20:
 		return 32
 	default:
 		return 256
@@ -248,6 +172,16 @@ func (t Tensor) typeSize() uint64 {
 		return 2 + 2*blockSize/8 + blockSize/32
 	case 18: // IQ3_XXS
 		return 2 + 3*blockSize/8
+	case 19: // IQ1_S
+		return 2 + blockSize/8 + blockSize/16
+	case 20: // IQ4_NL
+		return 2 + blockSize/2
+	case 21: // IQ3_S
+		return 2 + 2*blockSize/8 + blockSize/8 + blockSize/32 + 4
+	case 22: // IQ2_S
+		return 2 + blockSize/4 + blockSize/16
+	case 23: // IQ4_XS
+		return 4 + blockSize/2 + blockSize/64
 	default:
 		return 0
 	}
@@ -286,6 +220,23 @@ const (
 
 var ErrUnsupportedFormat = errors.New("unsupported model format")
 
+func DetectGGMLType(b []byte) string {
+	switch binary.LittleEndian.Uint32(b[:4]) {
+	case FILE_MAGIC_GGML:
+		return "ggml"
+	case FILE_MAGIC_GGMF:
+		return "ggmf"
+	case FILE_MAGIC_GGJT:
+		return "ggjt"
+	case FILE_MAGIC_GGLA:
+		return "ggla"
+	case FILE_MAGIC_GGUF_LE, FILE_MAGIC_GGUF_BE:
+		return "gguf"
+	default:
+		return ""
+	}
+}
+
 func DecodeGGML(rs io.ReadSeeker) (*GGML, int64, error) {
 	var magic uint32
 	if err := binary.Read(rs, binary.LittleEndian, &magic); err != nil {
@@ -343,7 +294,15 @@ func (llm GGML) GraphSize(context, batch uint64) (partialOffload, fullOffload ui
 			4*batch*(embedding+vocab)+embedding*vocab*105/128,
 		)
 
-		if ffnGateWeight, ok := layers["0"]["ffn_gate.0.weight"]; ok {
+		if ffnGateExpsWeight, ok := layers["blk.0"]["ffn_gate_exps.weight"]; ok {
+			// mixtral 8x22b
+			ff := uint64(llm.KV()["llama.feed_forward_length"].(uint32))
+			partialOffload = max(
+				3*ffnGateExpsWeight.size()+4*batch*(2*ff+headsKV+embedding+context+embedding/heads*headsKV),
+				4*(context*batch*heads+context*embedding/heads*headsKV+batch*1024+embedding/heads*headsKV*batch),
+			)
+		} else if ffnGateWeight, ok := layers["blk.0"]["ffn_gate.0.weight"]; ok {
+			// mixtral 8x7b
 			ffnGateWeight1 := ffnGateWeight.Shape[1]
 			fullOffload = 4 * batch * (2 + 3*embedding + context*(1+heads) + 2*headsKV + ffnGateWeight1)
 			partialOffload = max(

llm/gguf.go

@@ -190,8 +190,6 @@ func (llm *gguf) Decode(rs io.ReadSeeker) error {
 		llm.kv[k] = v
 	}
 
-	slog.Debug(fmt.Sprintf("general.architecture = %s", llm.kv["general.architecture"]))
-
 	// decode tensors
 	for i := 0; uint64(i) < llm.numTensor(); i++ {
 		name, err := readGGUFString(llm, rs)
@@ -465,11 +463,13 @@ var ggufKVOrder = map[string][]string{
 		"llama.embedding_length",
 		"llama.block_count",
 		"llama.feed_forward_length",
-		"llama.rope.dimension_count",
 		"llama.attention.head_count",
 		"llama.attention.head_count_kv",
 		"llama.attention.layer_norm_rms_epsilon",
 		"llama.rope.freq_base",
+		"llama.rope.dimension_count",
+		"llama.expert_count",
+		"llama.expert_used_count",
 		"gemma.context_length",
 		"gemma.embedding_length",
 		"gemma.block_count",
@@ -577,6 +577,8 @@ func (llm *gguf) Encode(ws io.WriteSeeker, kv KV, tensors []Tensor) error {
 					return err
 				}
 			}
+		default:
+			return fmt.Errorf("improper type for '%s'", k)
 		}
 		if err != nil {
 			return err
@@ -598,9 +600,11 @@ func (llm *gguf) Encode(ws io.WriteSeeker, kv KV, tensors []Tensor) error {
 			return err
 		}
 
-		dims := 1
-		if tensor.Shape[1] > 0 {
-			dims = 2
+		dims := 0
+		for cnt := 0; cnt < len(tensor.Shape); cnt++ {
+			if tensor.Shape[cnt] > 0 {
+				dims++
+			}
 		}
 
 		if err := binary.Write(ws, llm.ByteOrder, uint32(dims)); err != nil {

llm/llm.go

@@ -4,6 +4,7 @@ package llm
 // #cgo darwin,arm64 LDFLAGS: ${SRCDIR}/build/darwin/arm64_static/libllama.a -lstdc++
 // #cgo darwin,amd64 LDFLAGS: ${SRCDIR}/build/darwin/x86_64_static/libllama.a -lstdc++
 // #cgo windows,amd64 LDFLAGS: ${SRCDIR}/build/windows/amd64_static/libllama.a -static -lstdc++
+// #cgo windows,arm64 LDFLAGS: ${SRCDIR}/build/windows/arm64_static/libllama.a -static -lstdc++
 // #cgo linux,amd64 LDFLAGS: ${SRCDIR}/build/linux/x86_64_static/libllama.a -lstdc++
 // #cgo linux,arm64 LDFLAGS: ${SRCDIR}/build/linux/arm64_static/libllama.a -lstdc++
 // #include <stdlib.h>
@@ -19,7 +20,7 @@ func SystemInfo() string {
 	return C.GoString(C.llama_print_system_info())
 }
 
-func Quantize(infile, outfile, filetype string) error {
+func Quantize(infile, outfile string, ftype fileType) error {
 	cinfile := C.CString(infile)
 	defer C.free(unsafe.Pointer(cinfile))
 
@@ -28,58 +29,10 @@ func Quantize(infile, outfile, filetype string) error {
 
 	params := C.llama_model_quantize_default_params()
 	params.nthread = -1
+	params.ftype = ftype.Value()
 
-	switch filetype {
-	case "F32":
-		params.ftype = fileTypeF32
-	case "F16":
-		params.ftype = fileTypeF16
-	case "Q4_0":
-		params.ftype = fileTypeQ4_0
-	case "Q4_1":
-		params.ftype = fileTypeQ4_1
-	case "Q4_1_F16":
-		params.ftype = fileTypeQ4_1_F16
-	case "Q8_0":
-		params.ftype = fileTypeQ8_0
-	case "Q5_0":
-		params.ftype = fileTypeQ5_0
-	case "Q5_1":
-		params.ftype = fileTypeQ5_1
-	case "Q2_K":
-		params.ftype = fileTypeQ2_K
-	case "Q3_K_S":
-		params.ftype = fileTypeQ3_K_S
-	case "Q3_K_M":
-		params.ftype = fileTypeQ3_K_M
-	case "Q3_K_L":
-		params.ftype = fileTypeQ3_K_L
-	case "Q4_K_S":
-		params.ftype = fileTypeQ4_K_S
-	case "Q4_K_M":
-		params.ftype = fileTypeQ4_K_M
-	case "Q5_K_S":
-		params.ftype = fileTypeQ5_K_S
-	case "Q5_K_M":
-		params.ftype = fileTypeQ5_K_M
-	case "Q6_K":
-		params.ftype = fileTypeQ6_K
-	case "IQ2_XXS":
-		params.ftype = fileTypeIQ2_XXS
-	case "IQ2_XS":
-		params.ftype = fileTypeIQ2_XS
-	case "Q2_K_S":
-		params.ftype = fileTypeQ2_K_S
-	case "Q3_K_XS":
-		params.ftype = fileTypeQ3_K_XS
-	case "IQ3_XXS":
-		params.ftype = fileTypeIQ3_XXS
-	default:
-		return fmt.Errorf("unknown filetype: %s", filetype)
-	}
-
-	if retval := C.llama_model_quantize(cinfile, coutfile, &params); retval != 0 {
-		return fmt.Errorf("llama_model_quantize: %d", retval)
+	if rc := C.llama_model_quantize(cinfile, coutfile, &params); rc != 0 {
+		return fmt.Errorf("llama_model_quantize: %d", rc)
 	}
 
 	return nil

llm/llm_windows.go

@@ -2,5 +2,5 @@ package llm
 
 import "embed"
 
-//go:embed build/windows/*/*/bin/*
+// unused on windows
 var libEmbed embed.FS

llm/memory.go

@@ -0,0 +1,193 @@
+package llm
+
+import (
+	"fmt"
+	"log/slog"
+	"os"
+	"strconv"
+
+	"github.com/ollama/ollama/api"
+	"github.com/ollama/ollama/format"
+	"github.com/ollama/ollama/gpu"
+)
+
+// This algorithm looks for a complete fit to determine if we need to unload other models
+func PredictServerFit(allGpus gpu.GpuInfoList, ggml *GGML, adapters, projectors []string, opts api.Options) (bool, uint64) {
+	var estimatedVRAM uint64
+	if opts.NumCtx > int(ggml.KV().ContextLength()) {
+		slog.Warn("requested context length is greater than model max context length", "requested", opts.NumCtx, "model", ggml.KV().ContextLength())
+		opts.NumCtx = int(ggml.KV().ContextLength())
+	}
+
+	if opts.NumCtx < 4 {
+		opts.NumCtx = 4
+	}
+
+	// Split up the GPUs by type and try them
+	for _, gpus := range allGpus.ByLibrary() {
+		var layerCount int
+		layerCount, estimatedVRAM = EstimateGPULayers(gpus, ggml, projectors, opts)
+		if opts.NumGPU < 0 {
+			if layerCount > 0 && layerCount >= int(ggml.KV().BlockCount()+1) {
+				return true, estimatedVRAM
+			}
+		} else {
+			if layerCount > 0 && layerCount >= opts.NumGPU {
+				return true, estimatedVRAM
+			}
+		}
+	}
+	return false, estimatedVRAM
+}
+
+// Given a model and one or more GPU targets, predict how many layers and bytes we can load
+// The GPUs provided must all be the same Library
+func EstimateGPULayers(gpus []gpu.GpuInfo, ggml *GGML, projectors []string, opts api.Options) (int, uint64) {
+	if gpus[0].Library == "cpu" {
+		return 0, 0
+	}
+	var memoryAvailable uint64
+	for _, info := range gpus {
+		memoryAvailable += info.FreeMemory
+	}
+	userLimit := os.Getenv("OLLAMA_MAX_VRAM")
+	if userLimit != "" {
+		avail, err := strconv.ParseUint(userLimit, 10, 64)
+		if err != nil {
+			slog.Error("invalid setting, ignoring", "OLLAMA_MAX_VRAM", userLimit, "error", err)
+		} else {
+			slog.Info("user override memory limit", "OLLAMA_MAX_VRAM", avail, "actual", memoryAvailable)
+			memoryAvailable = avail
+		}
+	}
+
+	slog.Debug("evaluating", "library", gpus[0].Library, "gpu_count", len(gpus), "available", format.HumanBytes2(memoryAvailable))
+
+	// TODO - this is probably wrong, first GPU vs secondaries will have different overheads
+	memoryMinimum := gpus[0].MinimumMemory
+
+	for _, projector := range projectors {
+		memoryMinimum += projectorMemoryRequirements(projector)
+
+		// multimodal models require at least 2048 context
+		opts.NumCtx = max(opts.NumCtx, 2048)
+	}
+
+	// fp16 k,v = (1 (k) + 1 (v)) * sizeof(float16) * n_ctx * n_layer * n_embd / n_head * n_head_kv
+	var kv uint64 = 2 * 2 * uint64(opts.NumCtx) * ggml.KV().BlockCount() * ggml.KV().EmbeddingLength() / ggml.KV().HeadCount() * ggml.KV().HeadCountKV()
+
+	graphPartialOffload, graphFullOffload := ggml.GraphSize(uint64(opts.NumCtx), uint64(min(opts.NumCtx, opts.NumBatch)))
+	if graphPartialOffload == 0 {
+		graphPartialOffload = ggml.KV().GQA() * kv / 6
+	}
+
+	if graphFullOffload == 0 {
+		graphFullOffload = graphPartialOffload
+	}
+
+	graphFullOffload *= uint64(len(gpus))
+	graphPartialOffload *= uint64(len(gpus))
+
+	// on metal there's no partial offload overhead
+	if gpus[0].Library == "metal" {
+		graphPartialOffload = graphFullOffload
+	}
+
+	// memoryRequiredTotal represents the memory required for full GPU offloading (all layers)
+	memoryRequiredTotal := memoryMinimum + graphFullOffload
+
+	// memoryRequiredPartial represents the memory required for partial GPU offloading (n > 0, n < layers)
+	memoryRequiredPartial := memoryMinimum + graphPartialOffload
+
+	if memoryRequiredPartial > memoryAvailable {
+		slog.Debug("insufficient VRAM to load any model layers")
+		return 0, 0
+	}
+
+	layers := ggml.Tensors().Layers()
+
+	var memoryLayerOutput uint64
+	if layer, ok := layers["output_norm"]; ok {
+		memoryLayerOutput += layer.size()
+	}
+
+	if layer, ok := layers["output"]; ok {
+		memoryLayerOutput += layer.size()
+	} else if layer, ok := layers["token_embd"]; ok {
+		memoryLayerOutput += layer.size()
+	}
+
+	if gpus[0].Library == "metal" && opts.UseMMap {
+		// memory is preallocated for output tensors
+		memoryRequiredTotal += memoryLayerOutput
+		memoryRequiredPartial += memoryLayerOutput
+	}
+
+	var layerCount int
+	for i := 0; i < int(ggml.KV().BlockCount()); i++ {
+		memoryLayer := layers[fmt.Sprintf("blk.%d", i)].size()
+
+		// KV is proportional to the number of layers
+		memoryLayer += kv / ggml.KV().BlockCount()
+
+		memoryRequiredTotal += memoryLayer
+		if memoryAvailable > memoryRequiredPartial+memoryLayer {
+			memoryRequiredPartial += memoryLayer
+			layerCount++
+		}
+	}
+
+	if gpus[0].Library != "metal" || !opts.UseMMap {
+		// memory was not preallocated for output tensors
+		memoryRequiredTotal += memoryLayerOutput
+	}
+
+	if memoryAvailable > memoryRequiredTotal {
+		layerCount = int(ggml.KV().BlockCount()) + 1
+		memoryRequiredPartial = memoryRequiredTotal
+	}
+
+	memoryWeights := memoryRequiredTotal - memoryMinimum - graphFullOffload - kv
+
+	slog.Info(
+		"offload to gpu",
+		slog.Group(
+			"layers",
+			// actual number of layers offloaded
+			"real", opts.NumGPU,
+			// estimated number of layers that can be offloaded
+			"estimate", layerCount,
+		),
+		slog.Group(
+			"memory",
+			// memory available for offloading
+			"available", format.HumanBytes2(memoryAvailable),
+			slog.Group(
+				"required",
+				// memory required for full offloading
+				"full", format.HumanBytes2(memoryRequiredTotal),
+				// memory required to offload layers.estimate layers
+				"partial", format.HumanBytes2(memoryRequiredPartial),
+				// memory of KV cache
+				"kv", format.HumanBytes2(kv),
+			),
+			slog.Group(
+				"weights",
+				// memory of the weights
+				"total", format.HumanBytes2(memoryWeights),
+				// memory of repeating layers
+				"repeating", format.HumanBytes2(memoryWeights-memoryLayerOutput),
+				// memory of non-repeating layers
+				"nonrepeating", format.HumanBytes2(memoryLayerOutput),
+			),
+			slog.Group(
+				"graph",
+				// memory of graph when fully offloaded
+				"full", format.HumanBytes2(graphFullOffload),
+				// memory of graph when not fully offloaded
+				"partial", format.HumanBytes2(graphPartialOffload),
+			),
+		),
+	)
+	return layerCount, uint64(memoryRequiredPartial)
+}

llm/patches/02-clip-log.diff

@@ -0,0 +1,12 @@
+diff --git a/examples/llava/clip.cpp b/examples/llava/clip.cpp
+index e431c7f7..f077e688 100644
+--- a/examples/llava/clip.cpp
++++ b/examples/llava/clip.cpp
+@@ -3,6 +3,7 @@
+ // I'll gradually clean and extend it
+ // Note: Even when using identical normalized image inputs (see normalize_image_u8_to_f32()) we have a significant difference in resulting embeddings compared to pytorch
+ #include "clip.h"
++#include "common.h"
+ #include "log.h"
+ #include "ggml.h"
+ #include "ggml-alloc.h"

llm/patches/04-metal.diff

@@ -0,0 +1,45 @@
+diff --git a/ggml-metal.m b/ggml-metal.m
+index 0207b787..b5e9884b 100644
+--- a/ggml-metal.m
++++ b/ggml-metal.m
+@@ -1396,27 +1396,23 @@ static enum ggml_status ggml_metal_graph_compute(
+                         // to the matrix-vector kernel
+                         int ne11_mm_min = 1;
+ 
+-#if 0
+                         // the numbers below are measured on M2 Ultra for 7B and 13B models
+                         // these numbers do not translate to other devices or model sizes
+                         // TODO: need to find a better approach
+-                        if ([ctx->device.name isEqualToString:@"Apple M2 Ultra"]) {
+-                            switch (src0t) {
+-                                case GGML_TYPE_F16:  ne11_mm_min = 2;  break;
+-                                case GGML_TYPE_Q8_0: ne11_mm_min = 7;  break;
+-                                case GGML_TYPE_Q2_K: ne11_mm_min = 15; break;
+-                                case GGML_TYPE_Q3_K: ne11_mm_min = 7;  break;
+-                                case GGML_TYPE_Q4_0:
+-                                case GGML_TYPE_Q4_1: ne11_mm_min = 15; break;
+-                                case GGML_TYPE_Q4_K: ne11_mm_min = 11; break;
+-                                case GGML_TYPE_Q5_0:                          // not tested yet
+-                                case GGML_TYPE_Q5_1: ne11_mm_min = 13; break; // not tested yet
+-                                case GGML_TYPE_Q5_K: ne11_mm_min = 7;  break;
+-                                case GGML_TYPE_Q6_K: ne11_mm_min = 7;  break;
+-                                default:             ne11_mm_min = 1;  break;
+-                            }
++                        switch (src0t) {
++                            case GGML_TYPE_F16:  ne11_mm_min = 2;  break;
++                            case GGML_TYPE_Q8_0: ne11_mm_min = 7;  break;
++                            case GGML_TYPE_Q2_K: ne11_mm_min = 15; break;
++                            case GGML_TYPE_Q3_K: ne11_mm_min = 7;  break;
++                            case GGML_TYPE_Q4_0:
++                            case GGML_TYPE_Q4_1: ne11_mm_min = 15; break;
++                            case GGML_TYPE_Q4_K: ne11_mm_min = 11; break;
++                            case GGML_TYPE_Q5_0:                          // not tested yet
++                            case GGML_TYPE_Q5_1: ne11_mm_min = 13; break; // not tested yet
++                            case GGML_TYPE_Q5_K: ne11_mm_min = 7;  break;
++                            case GGML_TYPE_Q6_K: ne11_mm_min = 7;  break;
++                            default:             ne11_mm_min = 1;  break;
+                         }
+-#endif
+ 
+                         // for now the matrix-matrix multiplication kernel only works on A14+/M1+ SoCs
+                         // AMD GPU and older A-chips will reuse matrix-vector multiplication kernel

llm/payload.go

@@ -9,6 +9,7 @@ import (
 	"log/slog"
 	"os"
 	"path/filepath"
+	"runtime"
 	"strings"
 
 	"golang.org/x/exp/slices"
@@ -25,13 +26,15 @@ func Init() error {
 		return err
 	}
 
-	slog.Info("extracting embedded files", "dir", payloadsDir)
-	binGlob := "build/*/*/*/bin/*"
+	if runtime.GOOS != "windows" {
+		slog.Info("extracting embedded files", "dir", payloadsDir)
+		binGlob := "build/*/*/*/bin/*"
 
-	// extract server libraries
-	err = extractFiles(payloadsDir, binGlob)
-	if err != nil {
-		return fmt.Errorf("extract binaries: %v", err)
+		// extract server libraries
+		err = extractFiles(payloadsDir, binGlob)
+		if err != nil {
+			return fmt.Errorf("extract binaries: %v", err)
+		}
 	}
 
 	var variants []string
@@ -138,6 +141,23 @@ func serversForGpu(info gpu.GpuInfo) []string {
 	return servers
 }
 
+// Return the optimal server for this CPU architecture
+func serverForCpu() string {
+	if runtime.GOOS == "darwin" && runtime.GOARCH == "arm64" {
+		return "metal"
+	}
+	variant := gpu.GetCPUVariant()
+	availableServers := availableServers()
+	if variant != "" {
+		for cmp := range availableServers {
+			if cmp == "cpu_"+variant {
+				return cmp
+			}
+		}
+	}
+	return "cpu"
+}
+
 // extract extracts the embedded files to the target directory
 func extractFiles(targetDir string, glob string) error {
 	files, err := fs.Glob(libEmbed, glob)

llm/server.go

@@ -21,21 +21,43 @@ import (
 	"strings"
 	"time"
 
+	"golang.org/x/sync/semaphore"
+
 	"github.com/ollama/ollama/api"
 	"github.com/ollama/ollama/format"
 	"github.com/ollama/ollama/gpu"
 )
 
-// LlamaServer is an instance of the llama.cpp server
-type LlamaServer struct {
+type LlamaServer interface {
+	Ping(ctx context.Context) error
+	WaitUntilRunning(ctx context.Context) error
+	Completion(ctx context.Context, req CompletionRequest, fn func(CompletionResponse)) error
+	Embedding(ctx context.Context, prompt string) ([]float64, error)
+	Tokenize(ctx context.Context, content string) ([]int, error)
+	Detokenize(ctx context.Context, tokens []int) (string, error)
+	Close() error
+	EstimatedVRAM() uint64
+}
+
+// llmServer is an instance of the llama.cpp server
+type llmServer struct {
 	port    int
 	cmd     *exec.Cmd
 	done    chan error // Channel to signal when the process exits
 	status  *StatusWriter
 	options api.Options
+
+	// TODO - this should be broken down by GPU
+	estimatedVRAM uint64 // Estimated usage of VRAM by the loaded model
+
+	sem *semaphore.Weighted
 }
 
-func NewLlamaServer(model string, adapters, projectors []string, opts api.Options) (*LlamaServer, error) {
+func LoadModel(model string) (*GGML, error) {
+	if _, err := os.Stat(model); err != nil {
+		return nil, err
+	}
+
 	f, err := os.Open(model)
 	if err != nil {
 		return nil, err
@@ -43,144 +65,66 @@ func NewLlamaServer(model string, adapters, projectors []string, opts api.Option
 	defer f.Close()
 
 	ggml, _, err := DecodeGGML(f)
-	if err != nil {
-		return nil, err
-	}
+	return ggml, err
+}
 
+// NewLlamaServer will run a server for the given GPUs
+// The gpu list must be a single family.
+func NewLlamaServer(gpus gpu.GpuInfoList, model string, ggml *GGML, adapters, projectors []string, opts api.Options) (LlamaServer, error) {
+	var err error
 	if opts.NumCtx > int(ggml.KV().ContextLength()) {
-		slog.Warn("requested context length is greater than model max context length", "requested", opts.NumCtx, "model", ggml.KV().ContextLength())
-		opts.NumCtx = int(ggml.KV().ContextLength())
+		slog.Warn("requested context length is greater than the model's training context window size", "requested", opts.NumCtx, "training size", ggml.KV().ContextLength())
 	}
 
 	if opts.NumCtx < 4 {
 		opts.NumCtx = 4
 	}
 
-	memoryAvailable, _ := gpu.CheckVRAM()
-	info := gpu.GetGPUInfo()
+	cpuRunner := ""
+	var estimatedVRAM uint64
+	var systemMemory uint64
+	if (len(gpus) == 1 && gpus[0].Library == "cpu") || opts.NumGPU == 0 {
 
-	memoryMinimum := info.MinimumMemory
-	for _, projector := range projectors {
-		memoryMinimum += projectorMemoryRequirements(projector)
+		// TODO evaluate system memory to see if we should block the load, or force an unload of another CPU runner
 
-		// multimodal models require at least 2048 context
-		opts.NumCtx = max(opts.NumCtx, 2048)
-	}
+		cpuRunner = serverForCpu()
+	} else {
+		if gpus[0].Library == "metal" {
+			memInfo, err := gpu.GetCPUMem()
+			if err != nil {
+				slog.Error("failed to lookup system memory", "error", err)
+			} else {
+				systemMemory = memInfo.TotalMemory
+				slog.Debug("system memory", "total", format.HumanBytes2(systemMemory))
+			}
+		}
+		var layers int
+		layers, estimatedVRAM = EstimateGPULayers(gpus, ggml, projectors, opts)
 
-	// fp16 k,v = (1 (k) + 1 (v)) * sizeof(float16) * n_ctx * n_layer * n_embd / n_head * n_head_kv
-	var kv uint64 = 2 * 2 * uint64(opts.NumCtx) * ggml.KV().BlockCount() * ggml.KV().EmbeddingLength() / ggml.KV().HeadCount() * ggml.KV().HeadCountKV()
-
-	graphPartialOffload, graphFullOffload := ggml.GraphSize(uint64(opts.NumCtx), uint64(min(opts.NumCtx, opts.NumBatch)))
-	if graphPartialOffload == 0 {
-		graphPartialOffload = ggml.KV().GQA() * kv / 6
-	}
-
-	if graphFullOffload == 0 {
-		graphFullOffload = graphPartialOffload
-	}
-
-	graphFullOffload *= uint64(info.DeviceCount)
-	graphPartialOffload *= uint64(info.DeviceCount)
-
-	// memoryRequiredTotal represents the memory required for full GPU offloading (all layers)
-	memoryRequiredTotal := memoryMinimum + graphFullOffload
-
-	// memoryRequiredPartial represents the memory required for partial GPU offloading (n > 0, n < layers)
-	memoryRequiredPartial := memoryMinimum + graphPartialOffload
-
-	if info.Library != "metal" {
-		if memoryRequiredPartial > memoryAvailable {
-			info.Library = "cpu"
+		if gpus[0].Library == "metal" && estimatedVRAM > systemMemory {
+			// disable partial offloading when model is greater than total system memory as this
+			// can lead to locking up the system
+			opts.NumGPU = 0
+		} else if opts.NumGPU < 0 && layers > 0 && gpus[0].Library != "cpu" {
+			opts.NumGPU = layers
 		}
 	}
 
-	var layerCount int
-	layers := ggml.Tensors().Layers()
-	for i := 0; i < int(ggml.KV().BlockCount()); i++ {
-		memoryLayer := layers[fmt.Sprintf("blk.%d", i)].size()
-
-		// KV is proportional to the number of layers
-		memoryLayer += kv / ggml.KV().BlockCount()
-
-		memoryRequiredTotal += memoryLayer
-		if memoryAvailable > memoryRequiredPartial+memoryLayer {
-			memoryRequiredPartial += memoryLayer
-			layerCount++
-		}
-	}
-
-	var memoryLayerOutput uint64
-	for k, v := range layers {
-		if !strings.HasPrefix(k, "blk.") {
-			memoryLayerOutput += v.size()
-		}
-	}
-
-	memoryRequiredTotal += memoryLayerOutput
-
-	if info.Library == "metal" && memoryRequiredTotal > info.TotalMemory {
-		// disable partial offloading when model is greater than total system memory
-		opts.NumGPU = 0
-	} else if memoryAvailable > memoryRequiredTotal {
-		layerCount = int(ggml.KV().BlockCount()) + 1
-		memoryRequiredPartial = memoryRequiredTotal
-	}
-
-	if opts.NumGPU < 0 {
-		opts.NumGPU = layerCount
-	}
-
-	memoryWeights := memoryRequiredTotal - memoryMinimum - graphFullOffload - kv
-
-	slog.Info(
-		"offload to gpu",
-		slog.Group(
-			"layers",
-			// actual number of layers offloaded
-			"real", opts.NumGPU,
-			// estimated number of layers that can be offloaded
-			"estimate", layerCount,
-		),
-		slog.Group(
-			"memory",
-			// memory available for offloading
-			"available", format.HumanBytes2(memoryAvailable),
-			slog.Group(
-				"required",
-				// memory required for full offloading
-				"full", format.HumanBytes2(memoryRequiredTotal),
-				// memory required to offload layers.estimate layers
-				"partial", format.HumanBytes2(memoryRequiredPartial),
-				// memory of KV cache
-				"kv", format.HumanBytes2(kv),
-			),
-			slog.Group(
-				"weights",
-				// memory of the weights
-				"total", format.HumanBytes2(memoryWeights),
-				// memory of repeating layers
-				"repeating", format.HumanBytes2(memoryWeights-memoryLayerOutput),
-				// memory of non-repeating layers
-				"nonrepeating", format.HumanBytes2(memoryLayerOutput),
-			),
-			slog.Group(
-				"graph",
-				// memory of graph when fully offloaded
-				"full", format.HumanBytes2(graphFullOffload),
-				// memory of graph when not fully offloaded
-				"partial", format.HumanBytes2(graphPartialOffload),
-			),
-		),
-	)
+	// Loop through potential servers
+	finalErr := fmt.Errorf("no suitable llama servers found")
 
 	if len(adapters) > 1 {
 		return nil, errors.New("ollama supports only one lora adapter, but multiple were provided")
 	}
 
 	availableServers := availableServers()
-	servers := serversForGpu(info)
-
-	demandLib := os.Getenv("OLLAMA_LLM_LIBRARY")
+	var servers []string
+	if cpuRunner != "" {
+		servers = []string{cpuRunner}
+	} else {
+		servers = serversForGpu(gpus[0]) // All GPUs in the list are matching Library and Variant
+	}
+	demandLib := strings.Trim(os.Getenv("OLLAMA_LLM_LIBRARY"), "\"' ")
 	if demandLib != "" {
 		serverPath := availableServers[demandLib]
 		if serverPath == "" {
@@ -192,7 +136,7 @@ func NewLlamaServer(model string, adapters, projectors []string, opts api.Option
 	}
 
 	if len(servers) == 0 {
-		return nil, fmt.Errorf("no servers found for %v", info)
+		return nil, fmt.Errorf("no servers found for %v", gpus)
 	}
 
 	params := []string{
@@ -249,10 +193,26 @@ func NewLlamaServer(model string, adapters, projectors []string, opts api.Option
 		params = append(params, "--numa")
 	}
 
-	// Loop through potential servers
-	var finalErr error
+	// "--cont-batching", // TODO - doesn't seem to have any noticeable perf change for multiple requests
+	numParallel := 1
+	if onp := os.Getenv("OLLAMA_NUM_PARALLEL"); onp != "" {
+		numParallel, err = strconv.Atoi(onp)
+		if err != nil || numParallel <= 0 {
+			err = fmt.Errorf("invalid OLLAMA_NUM_PARALLEL=%s must be greater than zero - %w", onp, err)
+			slog.Error("misconfiguration", "error", err)
+			return nil, err
+		}
+	}
+	params = append(params, "--parallel", fmt.Sprintf("%d", numParallel))
+
 	for i := 0; i < len(servers); i++ {
 		dir := availableServers[servers[i]]
+		if dir == "" {
+			// Shouldn't happen
+			finalErr = fmt.Errorf("[%d] server %s not listed in available servers %v", i, servers[i], availableServers)
+			slog.Error("sever list inconsistent", "error", finalErr)
+			continue
+		}
 
 		// Find an availableServers  port, retry on each iterration in case the failure was a port conflict race
 		port := 0
@@ -275,30 +235,60 @@ func NewLlamaServer(model string, adapters, projectors []string, opts api.Option
 		}
 		// append the server directory to LD_LIBRARY_PATH/PATH
 		libraryPaths := []string{dir}
+
 		if libraryPath, ok := os.LookupEnv(pathEnv); ok {
 			// Append our runner directory to the path
 			// This will favor system libraries over our bundled library dependencies
 			libraryPaths = append(filepath.SplitList(libraryPath), libraryPaths...)
 		}
 
+		// Note: we always put the dependency path first
+		// since this was the exact version we verified for AMD GPUs
+		// and we favor what the user had in their path
+		if gpus[0].DependencyPath != "" {
+			// TODO refine for multi-gpu support
+			libraryPaths = append([]string{gpus[0].DependencyPath}, libraryPaths...)
+		}
+
 		server := filepath.Join(dir, "ollama_llama_server")
 		if runtime.GOOS == "windows" {
 			server = server + ".exe"
 		}
 
-		s := &LlamaServer{
-			port:    port,
-			cmd:     exec.Command(server, finalParams...),
-			status:  NewStatusWriter(os.Stderr),
-			options: opts,
+		// Detect tmp cleaners wiping out the file
+		_, err := os.Stat(server)
+		if errors.Is(err, os.ErrNotExist) {
+			slog.Warn("llama server disappeared, reinitializing payloads", "path", server, "error", err)
+			err = Init()
+			if err != nil {
+				slog.Warn("failed to reinitialize payloads", "error", err)
+				return nil, err
+			}
 		}
+
+		s := &llmServer{
+			port:          port,
+			cmd:           exec.Command(server, finalParams...),
+			status:        NewStatusWriter(os.Stderr),
+			options:       opts,
+			estimatedVRAM: estimatedVRAM,
+			sem:           semaphore.NewWeighted(int64(numParallel)),
+		}
+
 		libEnv := fmt.Sprintf("%s=%s", pathEnv, strings.Join(libraryPaths, string(filepath.ListSeparator)))
-		slog.Debug(libEnv)
 		s.cmd.Env = append(os.Environ(), libEnv)
 		s.cmd.Stdout = os.Stdout
 		s.cmd.Stderr = s.status
 
+		// TODO - multiple GPU selection logic...
+		key, val := gpu.GpuInfoList(gpus).GetVisibleDevicesEnv()
+		if key != "" {
+			s.cmd.Env = append(s.cmd.Env, key+"="+val)
+		}
+
 		slog.Info("starting llama server", "cmd", s.cmd.String())
+		// Log at debug as the environment is inherited and might contain sensitive information
+		slog.Debug("subprocess", "environment", s.cmd.Env)
 
 		if err = s.cmd.Start(); err != nil {
 			msg := ""
@@ -310,12 +300,13 @@ func NewLlamaServer(model string, adapters, projectors []string, opts api.Option
 			continue
 		}
 
-		// reap subprocess when it exits
-		go func() {
-			// Exit status managed via getServerStatus
-			_ = s.cmd.Wait()
-		}()
-
+		// TODO - make sure this is all wired up correctly
+		// if err = s.WaitUntilRunning(); err != nil {
+		// 	slog.Error("error starting llama server", "server", servers[i], "error", err)
+		// 	s.Close()
+		// 	finalErr = err
+		// 	continue
+		// }
 		return s, nil
 	}
 
@@ -353,6 +344,21 @@ const ( // iota is reset to 0
 	ServerStatusError
 )
 
+func (s ServerStatus) ToString() string {
+	switch s {
+	case ServerStatusReady:
+		return "llm server ready"
+	case ServerStatusNoSlotsAvaialble:
+		return "llm busy - no slots available"
+	case ServerStatusLoadingModel:
+		return "llm server loading model"
+	case ServerStatusNotResponding:
+		return "llm server not responding"
+	default:
+		return "llm server error"
+	}
+}
+
 type ServerStatusResp struct {
 	Status          string `json:"status"`
 	SlotsIdle       int    `json:"slots_idle"`
@@ -360,7 +366,7 @@ type ServerStatusResp struct {
 	Error           string `json:"error"`
 }
 
-func (s *LlamaServer) getServerStatus(ctx context.Context) (ServerStatus, error) {
+func (s *llmServer) getServerStatus(ctx context.Context) (ServerStatus, error) {
 	// Fail fast if its exited
 	if s.cmd.ProcessState != nil {
 		msg := ""
@@ -407,7 +413,7 @@ func (s *LlamaServer) getServerStatus(ctx context.Context) (ServerStatus, error)
 	}
 }
 
-func (s *LlamaServer) Ping(ctx context.Context) error {
+func (s *llmServer) Ping(ctx context.Context) error {
 	_, err := s.getServerStatus(ctx)
 	if err != nil {
 		slog.Debug("server unhealthy", "error", err)
@@ -416,7 +422,7 @@ func (s *LlamaServer) Ping(ctx context.Context) error {
 	return nil
 }
 
-func (s *LlamaServer) WaitUntilRunning() error {
+func (s *llmServer) WaitUntilRunning(ctx context.Context) error {
 	start := time.Now()
 	// TODO we need to wire up a better way to detect hangs during model load and startup of the server
 	expiresAt := time.Now().Add(10 * time.Minute) // be generous with timeout, large models can take a while to load
@@ -427,6 +433,9 @@ func (s *LlamaServer) WaitUntilRunning() error {
 	var lastStatus ServerStatus = -1
 	for {
 		select {
+		case <-ctx.Done():
+			slog.Info("context expired before server started")
+			return fmt.Errorf("timed out waiting for llama runner to start: %w", ctx.Err())
 		case err := <-s.done:
 			msg := ""
 			if s.status != nil && s.status.LastErrMsg != "" {
@@ -450,9 +459,9 @@ func (s *LlamaServer) WaitUntilRunning() error {
 				return fmt.Errorf("llama runner process no longer running: %d %s", s.cmd.ProcessState.ExitCode(), msg)
 			}
 
-			ctx, cancel := context.WithTimeout(context.Background(), 200*time.Millisecond)
+			c, cancel := context.WithTimeout(ctx, 200*time.Millisecond)
 			defer cancel()
-			status, err := s.getServerStatus(ctx)
+			status, err := s.getServerStatus(c)
 			if err != nil && lastStatus != status {
 				slog.Debug("server not yet available", "error", err)
 				lastStatus = status
@@ -509,13 +518,10 @@ type ImageData struct {
 }
 
 type completion struct {
-	Content      string `json:"content"`
-	Model        string `json:"model"`
-	Prompt       string `json:"prompt"`
-	Stop         bool   `json:"stop"`
-	StoppedEos   bool   `json:"stopped_eos"`
-	StoppedWord  bool   `json:"stopped_word"`
-	StoppedLimit bool   `json:"stopped_limit"`
+	Content string `json:"content"`
+	Model   string `json:"model"`
+	Prompt  string `json:"prompt"`
+	Stop    bool   `json:"stop"`
 
 	Timings struct {
 		PredictedN  int     `json:"predicted_n"`
@@ -535,14 +541,25 @@ type CompletionRequest struct {
 type CompletionResponse struct {
 	Content            string
 	Done               bool
-	DoneReason         string
 	PromptEvalCount    int
 	PromptEvalDuration time.Duration
 	EvalCount          int
 	EvalDuration       time.Duration
 }
 
-func (s *LlamaServer) Completion(ctx context.Context, req CompletionRequest, fn func(CompletionResponse)) error {
+func (s *llmServer) Completion(ctx context.Context, req CompletionRequest, fn func(CompletionResponse)) error {
+	if err := s.sem.Acquire(ctx, 1); err != nil {
+		slog.Error("Failed to acquire semaphore", "error", err)
+		return err
+	}
+	defer s.sem.Release(1)
+
+	// only allow maximum 10 "context shifts" to avoid infinite generation
+	if req.Options.NumPredict < 0 || req.Options.NumPredict > 10*s.options.NumCtx {
+		req.Options.NumPredict = 10 * s.options.NumCtx
+		slog.Debug("setting token limit to 10x num_ctx", "num_ctx", s.options.NumCtx, "num_predict", req.Options.NumPredict)
+	}
+
 	request := map[string]any{
 		"prompt":            req.Prompt,
 		"stream":            true,
@@ -573,7 +590,7 @@ func (s *LlamaServer) Completion(ctx context.Context, req CompletionRequest, fn
 	if err != nil {
 		return err
 	} else if status != ServerStatusReady {
-		return fmt.Errorf("unexpected server status: %d", status)
+		return fmt.Errorf("unexpected server status: %s", status.ToString())
 	}
 
 	if req.Format == "json" {
@@ -652,8 +669,6 @@ func (s *LlamaServer) Completion(ctx context.Context, req CompletionRequest, fn
 					return fmt.Errorf("error parsing llm response stream: %s", line)
 				}
 
-				fmt.Println("c", string(evt))
-
 				var c completion
 				if err := json.Unmarshal(evt, &c); err != nil {
 					return fmt.Errorf("error unmarshaling llm prediction response: %v", err)
@@ -680,18 +695,8 @@ func (s *LlamaServer) Completion(ctx context.Context, req CompletionRequest, fn
 				}
 
 				if c.Stop {
-					var doneReason string
-					switch {
-					case c.StoppedEos:
-						doneReason = "stop"
-					case c.StoppedWord:
-						doneReason = "stop"
-					case c.StoppedLimit:
-						doneReason = "limit"
-					}
 					fn(CompletionResponse{
 						Done:               true,
-						DoneReason:         doneReason,
 						PromptEvalCount:    c.Timings.PromptN,
 						PromptEvalDuration: parseDurationMs(c.Timings.PromptMS),
 						EvalCount:          c.Timings.PredictedN,
@@ -732,13 +737,18 @@ type EmbeddingResponse struct {
 	Embedding []float64 `json:"embedding"`
 }
 
-func (s *LlamaServer) Embedding(ctx context.Context, prompt string) ([]float64, error) {
+func (s *llmServer) Embedding(ctx context.Context, prompt string) ([]float64, error) {
+	if err := s.sem.Acquire(ctx, 1); err != nil {
+		slog.Error("Failed to acquire semaphore", "error", err)
+		return nil, err
+	}
+	defer s.sem.Release(1)
 	// Make sure the server is ready
 	status, err := s.getServerStatus(ctx)
 	if err != nil {
 		return nil, err
 	} else if status != ServerStatusReady {
-		return nil, fmt.Errorf("unexpected server status: %d", status)
+		return nil, fmt.Errorf("unexpected server status: %s", status.ToString())
 	}
 
 	data, err := json.Marshal(TokenizeRequest{Content: prompt})
@@ -784,13 +794,13 @@ type TokenizeResponse struct {
 	Tokens []int `json:"tokens"`
 }
 
-func (s *LlamaServer) Tokenize(ctx context.Context, content string) ([]int, error) {
+func (s *llmServer) Tokenize(ctx context.Context, content string) ([]int, error) {
 	// Make sure the server is ready
 	status, err := s.getServerStatus(ctx)
 	if err != nil {
 		return nil, err
-	} else if status != ServerStatusReady {
-		return nil, fmt.Errorf("unexpected server status: %d", status)
+	} else if status != ServerStatusReady && status != ServerStatusNoSlotsAvaialble {
+		return nil, fmt.Errorf("unexpected server status: %s", status.ToString())
 	}
 
 	data, err := json.Marshal(TokenizeRequest{Content: content})
@@ -836,13 +846,13 @@ type DetokenizeResponse struct {
 	Content string `json:"content"`
 }
 
-func (s *LlamaServer) Detokenize(ctx context.Context, tokens []int) (string, error) {
+func (s *llmServer) Detokenize(ctx context.Context, tokens []int) (string, error) {
 	// Make sure the server is ready
 	status, err := s.getServerStatus(ctx)
 	if err != nil {
 		return "", err
-	} else if status != ServerStatusReady {
-		return "", fmt.Errorf("unexpected server status: %d", status)
+	} else if status != ServerStatusReady && status != ServerStatusNoSlotsAvaialble {
+		return "", fmt.Errorf("unexpected server status: %s", status.ToString())
 	}
 
 	data, err := json.Marshal(DetokenizeRequest{Tokens: tokens})
@@ -880,15 +890,25 @@ func (s *LlamaServer) Detokenize(ctx context.Context, tokens []int) (string, err
 	return decoded.Content, nil
 }
 
-func (s *LlamaServer) Close() error {
+func (s *llmServer) Close() error {
 	if s.cmd != nil {
 		slog.Debug("stopping llama server")
-		return s.cmd.Process.Kill()
+		if err := s.cmd.Process.Kill(); err != nil {
+			return err
+		}
+
+		_ = s.cmd.Wait()
+
+		slog.Debug("llama server stopped")
 	}
 
 	return nil
 }
 
+func (s *llmServer) EstimatedVRAM() uint64 {
+	return s.estimatedVRAM
+}
+
 func parseDurationMs(ms float64) time.Duration {
 	dur, err := time.ParseDuration(fmt.Sprintf("%fms", ms))
 	if err != nil {

macapp/src/app.tsx

@@ -19,7 +19,7 @@ export default function () {
   const [step, setStep] = useState<Step>(Step.WELCOME)
   const [commandCopied, setCommandCopied] = useState<boolean>(false)
 
-  const command = 'ollama run llama2'
+  const command = 'ollama run llama3'
 
   return (
     <div className='drag'>

parser/parser.go

@@ -1,132 +0,0 @@
-package parser
-
-import (
-	"bufio"
-	"bytes"
-	"errors"
-	"fmt"
-	"io"
-	"log/slog"
-	"slices"
-)
-
-type Command struct {
-	Name string
-	Args string
-}
-
-func (c *Command) Reset() {
-	c.Name = ""
-	c.Args = ""
-}
-
-func Parse(reader io.Reader) ([]Command, error) {
-	var commands []Command
-	var command, modelCommand Command
-
-	scanner := bufio.NewScanner(reader)
-	scanner.Buffer(make([]byte, 0, bufio.MaxScanTokenSize), bufio.MaxScanTokenSize)
-	scanner.Split(scanModelfile)
-	for scanner.Scan() {
-		line := scanner.Bytes()
-
-		fields := bytes.SplitN(line, []byte(" "), 2)
-		if len(fields) == 0 || len(fields[0]) == 0 {
-			continue
-		}
-
-		switch string(bytes.ToUpper(fields[0])) {
-		case "FROM":
-			command.Name = "model"
-			command.Args = string(bytes.TrimSpace(fields[1]))
-			// copy command for validation
-			modelCommand = command
-		case "ADAPTER":
-			command.Name = string(bytes.ToLower(fields[0]))
-			command.Args = string(bytes.TrimSpace(fields[1]))
-		case "LICENSE", "TEMPLATE", "SYSTEM", "PROMPT":
-			command.Name = string(bytes.ToLower(fields[0]))
-			command.Args = string(fields[1])
-		case "PARAMETER":
-			fields = bytes.SplitN(fields[1], []byte(" "), 2)
-			if len(fields) < 2 {
-				return nil, fmt.Errorf("missing value for %s", fields)
-			}
-
-			command.Name = string(fields[0])
-			command.Args = string(bytes.TrimSpace(fields[1]))
-		case "EMBED":
-			return nil, fmt.Errorf("deprecated command: EMBED is no longer supported, use the /embed API endpoint instead")
-		case "MESSAGE":
-			command.Name = string(bytes.ToLower(fields[0]))
-			fields = bytes.SplitN(fields[1], []byte(" "), 2)
-			if len(fields) < 2 {
-				return nil, fmt.Errorf("should be in the format <role> <message>")
-			}
-			if !slices.Contains([]string{"system", "user", "assistant"}, string(bytes.ToLower(fields[0]))) {
-				return nil, fmt.Errorf("role must be one of \"system\", \"user\", or \"assistant\"")
-			}
-			command.Args = fmt.Sprintf("%s: %s", string(bytes.ToLower(fields[0])), string(fields[1]))
-		default:
-			if !bytes.HasPrefix(fields[0], []byte("#")) {
-				// log a warning for unknown commands
-				slog.Warn(fmt.Sprintf("Unknown command: %s", fields[0]))
-			}
-			continue
-		}
-
-		commands = append(commands, command)
-		command.Reset()
-	}
-
-	if modelCommand.Args == "" {
-		return nil, errors.New("no FROM line for the model was specified")
-	}
-
-	return commands, scanner.Err()
-}
-
-func scanModelfile(data []byte, atEOF bool) (advance int, token []byte, err error) {
-	advance, token, err = scan([]byte(`"""`), []byte(`"""`), data, atEOF)
-	if err != nil {
-		return 0, nil, err
-	}
-
-	if advance > 0 && token != nil {
-		return advance, token, nil
-	}
-
-	advance, token, err = scan([]byte(`"`), []byte(`"`), data, atEOF)
-	if err != nil {
-		return 0, nil, err
-	}
-
-	if advance > 0 && token != nil {
-		return advance, token, nil
-	}
-
-	return bufio.ScanLines(data, atEOF)
-}
-
-func scan(openBytes, closeBytes, data []byte, atEOF bool) (advance int, token []byte, err error) {
-	newline := bytes.IndexByte(data, '\n')
-
-	if start := bytes.Index(data, openBytes); start >= 0 && start < newline {
-		end := bytes.Index(data[start+len(openBytes):], closeBytes)
-		if end < 0 {
-			if atEOF {
-				return 0, nil, fmt.Errorf("unterminated %s: expecting %s", openBytes, closeBytes)
-			} else {
-				return 0, nil, nil
-			}
-		}
-
-		n := start + len(openBytes) + end + len(closeBytes)
-
-		newData := data[:start]
-		newData = append(newData, data[start+len(openBytes):n-len(closeBytes)]...)
-		return n, newData, nil
-	}
-
-	return 0, nil, nil
-}

parser/parser_test.go

@@ -1,98 +0,0 @@
-package parser
-
-import (
-	"strings"
-	"testing"
-
-	"github.com/stretchr/testify/assert"
-)
-
-func Test_Parser(t *testing.T) {
-
-	input := `
-FROM model1
-ADAPTER adapter1
-LICENSE MIT
-PARAMETER param1 value1
-PARAMETER param2 value2
-TEMPLATE template1
-`
-
-	reader := strings.NewReader(input)
-
-	commands, err := Parse(reader)
-	assert.Nil(t, err)
-
-	expectedCommands := []Command{
-		{Name: "model", Args: "model1"},
-		{Name: "adapter", Args: "adapter1"},
-		{Name: "license", Args: "MIT"},
-		{Name: "param1", Args: "value1"},
-		{Name: "param2", Args: "value2"},
-		{Name: "template", Args: "template1"},
-	}
-
-	assert.Equal(t, expectedCommands, commands)
-}
-
-func Test_Parser_NoFromLine(t *testing.T) {
-
-	input := `
-PARAMETER param1 value1
-PARAMETER param2 value2
-`
-
-	reader := strings.NewReader(input)
-
-	_, err := Parse(reader)
-	assert.ErrorContains(t, err, "no FROM line")
-}
-
-func Test_Parser_MissingValue(t *testing.T) {
-
-	input := `
-FROM foo
-PARAMETER param1
-`
-
-	reader := strings.NewReader(input)
-
-	_, err := Parse(reader)
-	assert.ErrorContains(t, err, "missing value for [param1]")
-
-}
-
-func Test_Parser_Messages(t *testing.T) {
-
-	input := `
-FROM foo
-MESSAGE system You are a Parser. Always Parse things.
-MESSAGE user Hey there!
-MESSAGE assistant Hello, I want to parse all the things!
-`
-
-	reader := strings.NewReader(input)
-	commands, err := Parse(reader)
-	assert.Nil(t, err)
-
-	expectedCommands := []Command{
-		{Name: "model", Args: "foo"},
-		{Name: "message", Args: "system: You are a Parser. Always Parse things."},
-		{Name: "message", Args: "user: Hey there!"},
-		{Name: "message", Args: "assistant: Hello, I want to parse all the things!"},
-	}
-
-	assert.Equal(t, expectedCommands, commands)
-}
-
-func Test_Parser_Messages_BadRole(t *testing.T) {
-
-	input := `
-FROM foo
-MESSAGE badguy I'm a bad guy!
-`
-
-	reader := strings.NewReader(input)
-	_, err := Parse(reader)
-	assert.ErrorContains(t, err, "role must be one of \"system\", \"user\", or \"assistant\"")
-}

scripts/build_windows.ps1

@@ -7,6 +7,8 @@
 $ErrorActionPreference = "Stop"
 
 function checkEnv() {
+    $script:TARGET_ARCH=$Env:PROCESSOR_ARCHITECTURE.ToLower()
+    Write-host "Building for ${script:TARGET_ARCH}"
     write-host "Locating required tools and paths"
     $script:SRC_DIR=$PWD
     if (!$env:VCToolsRedistDir) {
@@ -30,7 +32,7 @@ function checkEnv() {
     
     $script:INNO_SETUP_DIR=(get-item "C:\Program Files*\Inno Setup*\")[0]
 
-    $script:DEPS_DIR="${script:SRC_DIR}\dist\windeps"
+    $script:DEPS_DIR="${script:SRC_DIR}\dist\windows-${script:TARGET_ARCH}"
     $env:CGO_ENABLED="1"
     echo "Checking version"
     if (!$env:VERSION) {
@@ -81,8 +83,8 @@ function buildOllama() {
             /csp "Google Cloud KMS Provider" /kc ${env:KEY_CONTAINER} ollama.exe
         if ($LASTEXITCODE -ne 0) { exit($LASTEXITCODE)}
     }
-    New-Item -ItemType Directory -Path .\dist -Force
-    cp .\ollama.exe .\dist\ollama-windows-amd64.exe
+    New-Item -ItemType Directory -Path .\dist\windows-${script:TARGET_ARCH}\ -Force
+    cp .\ollama.exe .\dist\windows-${script:TARGET_ARCH}\
 }
 
 function buildApp() {
@@ -101,7 +103,6 @@ function buildApp() {
 function gatherDependencies() {
     write-host "Gathering runtime dependencies"
     cd "${script:SRC_DIR}"
-    rm -ea 0 -recurse -force -path "${script:DEPS_DIR}"
     md "${script:DEPS_DIR}" -ea 0 > $null
 
     # TODO - this varies based on host build system and MSVC version - drive from dumpbin output
@@ -110,9 +111,6 @@ function gatherDependencies() {
     cp "${env:VCToolsRedistDir}\x64\Microsoft.VC*.CRT\vcruntime140.dll" "${script:DEPS_DIR}\"
     cp "${env:VCToolsRedistDir}\x64\Microsoft.VC*.CRT\vcruntime140_1.dll" "${script:DEPS_DIR}\"
 
-    cp "${script:NVIDIA_DIR}\cudart64_*.dll" "${script:DEPS_DIR}\"
-    cp "${script:NVIDIA_DIR}\cublas64_*.dll" "${script:DEPS_DIR}\"
-    cp "${script:NVIDIA_DIR}\cublasLt64_*.dll" "${script:DEPS_DIR}\"
 
     cp "${script:SRC_DIR}\app\ollama_welcome.ps1" "${script:SRC_DIR}\dist\"
     if ("${env:KEY_CONTAINER}") {
@@ -124,7 +122,6 @@ function gatherDependencies() {
             if ($LASTEXITCODE -ne 0) { exit($LASTEXITCODE)}
         }
     }
-
 }
 
 function buildInstaller() {
@@ -132,19 +129,25 @@ function buildInstaller() {
     cd "${script:SRC_DIR}\app"
     $env:PKG_VERSION=$script:PKG_VERSION
     if ("${env:KEY_CONTAINER}") {
-        & "${script:INNO_SETUP_DIR}\ISCC.exe" /SMySignTool="${script:SignTool} sign /fd sha256 /t http://timestamp.digicert.com /f ${script:OLLAMA_CERT} /csp `$qGoogle Cloud KMS Provider`$q /kc ${env:KEY_CONTAINER} `$f" .\ollama.iss
+        & "${script:INNO_SETUP_DIR}\ISCC.exe" /DARCH=$script:TARGET_ARCH /SMySignTool="${script:SignTool} sign /fd sha256 /t http://timestamp.digicert.com /f ${script:OLLAMA_CERT} /csp `$qGoogle Cloud KMS Provider`$q /kc ${env:KEY_CONTAINER} `$f" .\ollama.iss
     } else {
-        & "${script:INNO_SETUP_DIR}\ISCC.exe" .\ollama.iss
+        & "${script:INNO_SETUP_DIR}\ISCC.exe" /DARCH=$script:TARGET_ARCH .\ollama.iss
     }
     if ($LASTEXITCODE -ne 0) { exit($LASTEXITCODE)}
 }
 
+function distZip() {
+    write-host "Generating stand-alone distribution zip file ${script:SRC_DIR}\dist\ollama-windows-${script:TARGET_ARCH}.zip"
+    Compress-Archive -Path "${script:SRC_DIR}\dist\windows-${script:TARGET_ARCH}\*" -DestinationPath "${script:SRC_DIR}\dist\ollama-windows-${script:TARGET_ARCH}.zip" -Force
+}
+
 try {
     checkEnv
     buildOllama
     buildApp
     gatherDependencies
     buildInstaller
+    distZip
 } catch {
     write-host "Build Failed"
     write-host $_

server/images.go

@@ -1,16 +1,16 @@
 package server
 
 import (
-	"archive/zip"
 	"bytes"
+	"cmp"
 	"context"
 	"crypto/sha256"
+	"encoding/base64"
 	"encoding/hex"
 	"encoding/json"
 	"errors"
 	"fmt"
 	"io"
-	"io/fs"
 	"log"
 	"log/slog"
 	"net/http"
@@ -20,15 +20,15 @@ import (
 	"runtime"
 	"strconv"
 	"strings"
-	"text/template"
 
 	"golang.org/x/exp/slices"
 
 	"github.com/ollama/ollama/api"
-	"github.com/ollama/ollama/convert"
+	"github.com/ollama/ollama/auth"
 	"github.com/ollama/ollama/format"
 	"github.com/ollama/ollama/llm"
-	"github.com/ollama/ollama/parser"
+	"github.com/ollama/ollama/types/errtypes"
+	"github.com/ollama/ollama/types/model"
 	"github.com/ollama/ollama/version"
 )
 
@@ -60,6 +60,76 @@ func (m *Model) IsEmbedding() bool {
 	return slices.Contains(m.Config.ModelFamilies, "bert") || slices.Contains(m.Config.ModelFamilies, "nomic-bert")
 }
 
+func (m *Model) String() string {
+	var modelfile model.File
+
+	modelfile.Commands = append(modelfile.Commands, model.Command{
+		Name: "model",
+		Args: m.ModelPath,
+	})
+
+	if m.Template != "" {
+		modelfile.Commands = append(modelfile.Commands, model.Command{
+			Name: "template",
+			Args: m.Template,
+		})
+	}
+
+	if m.System != "" {
+		modelfile.Commands = append(modelfile.Commands, model.Command{
+			Name: "system",
+			Args: m.System,
+		})
+	}
+
+	for _, adapter := range m.AdapterPaths {
+		modelfile.Commands = append(modelfile.Commands, model.Command{
+			Name: "adapter",
+			Args: adapter,
+		})
+	}
+
+	for _, projector := range m.ProjectorPaths {
+		modelfile.Commands = append(modelfile.Commands, model.Command{
+			Name: "projector",
+			Args: projector,
+		})
+	}
+
+	for k, v := range m.Options {
+		switch v := v.(type) {
+		case []any:
+			for _, s := range v {
+				modelfile.Commands = append(modelfile.Commands, model.Command{
+					Name: k,
+					Args: fmt.Sprintf("%v", s),
+				})
+			}
+		default:
+			modelfile.Commands = append(modelfile.Commands, model.Command{
+				Name: k,
+				Args: fmt.Sprintf("%v", v),
+			})
+		}
+	}
+
+	for _, license := range m.License {
+		modelfile.Commands = append(modelfile.Commands, model.Command{
+			Name: "license",
+			Args: license,
+		})
+	}
+
+	for _, msg := range m.Messages {
+		modelfile.Commands = append(modelfile.Commands, model.Command{
+			Name: "message",
+			Args: fmt.Sprintf("%s %s", msg.Role, msg.Content),
+		})
+	}
+
+	return modelfile.String()
+}
+
 type Message struct {
 	Role    string `json:"role"`
 	Content string `json:"content"`
@@ -85,36 +155,6 @@ type ConfigV2 struct {
 	RootFS       RootFS `json:"rootfs"`
 }
 
-func (c *ConfigV2) SetModelFormat(format string) {
-	if c.ModelFormat == "" {
-		c.ModelFormat = format
-	}
-}
-
-func (c *ConfigV2) SetModelFamily(families ...string) {
-	for _, family := range families {
-		if c.ModelFamily == "" {
-			c.ModelFamily = family
-		}
-
-		if !slices.Contains(c.ModelFamilies, family) {
-			c.ModelFamilies = append(c.ModelFamilies, family)
-		}
-	}
-}
-
-func (c *ConfigV2) SetModelType(modelType string) {
-	if c.ModelType == "" {
-		c.ModelType = modelType
-	}
-}
-
-func (c *ConfigV2) SetFileType(fileType string) {
-	if c.FileType == "" {
-		c.FileType = fileType
-	}
-}
-
 type RootFS struct {
 	Type    string   `json:"type"`
 	DiffIDs []string `json:"diff_ids"`
@@ -259,7 +299,7 @@ func GetModel(name string) (*Model, error) {
 	return model, nil
 }
 
-func realpath(mfDir, from string) string {
+func realpath(rel, from string) string {
 	abspath, err := filepath.Abs(from)
 	if err != nil {
 		return from
@@ -276,22 +316,15 @@ func realpath(mfDir, from string) string {
 		return filepath.Join(home, from[2:])
 	}
 
-	if _, err := os.Stat(filepath.Join(mfDir, from)); err == nil {
+	if _, err := os.Stat(filepath.Join(rel, from)); err == nil {
 		// this is a file relative to the Modelfile
-		return filepath.Join(mfDir, from)
+		return filepath.Join(rel, from)
 	}
 
 	return abspath
 }
 
-func CreateModel(ctx context.Context, name, modelFileDir, quantization string, commands []parser.Command, fn func(resp api.ProgressResponse)) error {
-	deleteMap := make(map[string]struct{})
-	if manifest, _, err := GetManifest(ParseModelPath(name)); err == nil {
-		for _, layer := range append(manifest.Layers, manifest.Config) {
-			deleteMap[layer.Digest] = struct{}{}
-		}
-	}
-
+func CreateModel(ctx context.Context, name, modelFileDir, quantization string, modelfile *model.File, fn func(resp api.ProgressResponse)) (err error) {
 	config := ConfigV2{
 		OS:           "linux",
 		Architecture: "amd64",
@@ -300,250 +333,181 @@ func CreateModel(ctx context.Context, name, modelFileDir, quantization string, c
 		},
 	}
 
-	var layers Layers
-	messages := []string{}
+	var messages []*api.Message
+	parameters := make(map[string]any)
 
-	params := make(map[string][]string)
-	fromParams := make(map[string]any)
-
-	for _, c := range commands {
+	var layers []*Layer
+	for _, c := range modelfile.Commands {
 		mediatype := fmt.Sprintf("application/vnd.ollama.image.%s", c.Name)
 
 		switch c.Name {
-		case "model":
-			if strings.HasPrefix(c.Args, "@") {
-				blobPath, err := GetBlobsPath(strings.TrimPrefix(c.Args, "@"))
+		case "model", "adapter":
+			var baseLayers []*layerWithGGML
+			if name := model.ParseName(c.Args); name.IsValid() {
+				baseLayers, err = parseFromModel(ctx, name, fn)
+				if err != nil {
+					return err
+				}
+			} else if strings.HasPrefix(c.Args, "@") {
+				blobpath, err := GetBlobsPath(strings.TrimPrefix(c.Args, "@"))
 				if err != nil {
 					return err
 				}
 
-				c.Args = blobPath
-			}
-
-			pathName := realpath(modelFileDir, c.Args)
-
-			ggufName, err := convertModel(name, pathName, fn)
-			if err != nil {
-				var pathErr *fs.PathError
-				switch {
-				case errors.Is(err, zip.ErrFormat):
-					// it's not a safetensor archive
-				case errors.As(err, &pathErr):
-					// it's not a file on disk, could be a model reference
-				default:
-					return err
-				}
-			}
-
-			if ggufName != "" {
-				pathName = ggufName
-				defer os.RemoveAll(ggufName)
-
-				if quantization != "" {
-					quantization = strings.ToUpper(quantization)
-					fn(api.ProgressResponse{Status: fmt.Sprintf("quantizing %s model to %s", "F16", quantization)})
-					tempfile, err := os.CreateTemp(filepath.Dir(ggufName), quantization)
-					if err != nil {
-						return err
-					}
-					defer os.RemoveAll(tempfile.Name())
-
-					if err := llm.Quantize(ggufName, tempfile.Name(), quantization); err != nil {
-						return err
-					}
-
-					if err := tempfile.Close(); err != nil {
-						return err
-					}
-
-					pathName = tempfile.Name()
-				}
-			}
-
-			bin, err := os.Open(pathName)
-			if err != nil {
-				// not a file on disk so must be a model reference
-				modelpath := ParseModelPath(c.Args)
-				manifest, _, err := GetManifest(modelpath)
-				switch {
-				case errors.Is(err, os.ErrNotExist):
-					fn(api.ProgressResponse{Status: "pulling model"})
-					if err := PullModel(ctx, c.Args, &registryOptions{}, fn); err != nil {
-						return err
-					}
-
-					manifest, _, err = GetManifest(modelpath)
-					if err != nil {
-						return err
-					}
-				case err != nil:
-					return err
-				}
-
-				fn(api.ProgressResponse{Status: "reading model metadata"})
-				fromConfigPath, err := GetBlobsPath(manifest.Config.Digest)
+				blob, err := os.Open(blobpath)
 				if err != nil {
 					return err
 				}
+				defer blob.Close()
 
-				fromConfigFile, err := os.Open(fromConfigPath)
+				baseLayers, err = parseFromFile(ctx, blob, fn)
 				if err != nil {
 					return err
 				}
-				defer fromConfigFile.Close()
+			} else if file, err := os.Open(realpath(modelFileDir, c.Args)); err == nil {
+				defer file.Close()
 
-				var fromConfig ConfigV2
-				if err := json.NewDecoder(fromConfigFile).Decode(&fromConfig); err != nil {
+				baseLayers, err = parseFromFile(ctx, file, fn)
+				if err != nil {
 					return err
 				}
+			} else {
+				return fmt.Errorf("invalid model reference: %s", c.Args)
+			}
 
-				// if the model is still not in gguf format, error out
-				if fromConfig.ModelFormat != "gguf" {
-					return fmt.Errorf("%s is not in gguf format, this base model is not compatible with this version of ollama", c.Args)
-				}
-
-				config.SetModelFormat(fromConfig.ModelFormat)
-				config.SetModelFamily(append(fromConfig.ModelFamilies, fromConfig.ModelFamily)...)
-				config.SetModelType(fromConfig.ModelType)
-				config.SetFileType(fromConfig.FileType)
-
-				for _, layer := range manifest.Layers {
-					deleteMap[layer.Digest] = struct{}{}
-					if layer.MediaType == "application/vnd.ollama.image.params" {
-						fromParamsPath, err := GetBlobsPath(layer.Digest)
-						if err != nil {
-							return err
-						}
-
-						fromParamsFile, err := os.Open(fromParamsPath)
-						if err != nil {
-							return err
-						}
-						defer fromParamsFile.Close()
-
-						if err := json.NewDecoder(fromParamsFile).Decode(&fromParams); err != nil {
-							return err
-						}
-					}
-
-					layer, err := NewLayerFromLayer(layer.Digest, layer.MediaType, modelpath.GetShortTagname())
+			for _, baseLayer := range baseLayers {
+				if quantization != "" &&
+					baseLayer.MediaType == "application/vnd.ollama.image.model" &&
+					baseLayer.GGML != nil &&
+					baseLayer.GGML.Name() == "gguf" {
+					ftype, err := llm.ParseFileType(quantization)
 					if err != nil {
 						return err
 					}
 
-					layers.Add(layer)
+					filetype := baseLayer.GGML.KV().FileType()
+					if !slices.Contains([]string{"F16", "F32"}, filetype) {
+						return errors.New("quantization is only supported for F16 and F32 models")
+					}
+
+					fn(api.ProgressResponse{Status: fmt.Sprintf("quantizing %s model to %s", filetype, quantization)})
+
+					blob, err := GetBlobsPath(baseLayer.Digest)
+					if err != nil {
+						return err
+					}
+
+					temp, err := os.CreateTemp(filepath.Dir(blob), quantization)
+					if err != nil {
+						return err
+					}
+					defer temp.Close()
+					defer os.Remove(temp.Name())
+
+					if err := llm.Quantize(blob, temp.Name(), ftype); err != nil {
+						return err
+					}
+
+					baseLayer.Layer, err = NewLayer(temp, baseLayer.Layer.MediaType)
+					if err != nil {
+						return err
+					}
 				}
 
-				deleteMap[manifest.Config.Digest] = struct{}{}
-				continue
+				if baseLayer.GGML != nil {
+					config.ModelFormat = cmp.Or(config.ModelFormat, baseLayer.GGML.Name())
+					config.ModelFamily = cmp.Or(config.ModelFamily, baseLayer.GGML.KV().Architecture())
+					config.ModelType = cmp.Or(config.ModelType, format.HumanNumber(baseLayer.GGML.KV().ParameterCount()))
+					config.FileType = cmp.Or(config.FileType, baseLayer.GGML.KV().FileType())
+					config.ModelFamilies = append(config.ModelFamilies, baseLayer.GGML.KV().Architecture())
+				}
+
+				layers = append(layers, baseLayer.Layer)
 			}
-			defer bin.Close()
-
-			var offset int64
-			for {
-				fn(api.ProgressResponse{Status: "creating model layer"})
-				if _, err := bin.Seek(offset, io.SeekStart); err != nil {
-					return err
-				}
-
-				ggml, size, err := llm.DecodeGGML(bin)
-				if errors.Is(err, io.EOF) {
-					break
-				} else if errors.Is(err, llm.ErrUnsupportedFormat) {
-					return fmt.Errorf("model binary specified in FROM field is not a valid gguf format model, %w", err)
-				} else if err != nil {
-					return err
-				}
-
-				config.SetModelFormat(ggml.Name())
-				config.SetModelFamily(ggml.KV().Architecture())
-				config.SetModelType(format.HumanNumber(ggml.KV().ParameterCount()))
-				config.SetFileType(ggml.KV().FileType())
-
-				mediatype := mediatype
-				if ggml.KV().Architecture() == "clip" {
-					mediatype = "application/vnd.ollama.image.projector"
-				}
-
-				sr := io.NewSectionReader(bin, offset, size)
-				layer, err := NewLayer(sr, mediatype)
-				if err != nil {
-					return err
-				}
-
-				layers.Add(layer)
-
-				offset += size
-			}
-		case "adapter":
-			if strings.HasPrefix(c.Args, "@") {
-				blobPath, err := GetBlobsPath(strings.TrimPrefix(c.Args, "@"))
-				if err != nil {
-					return err
-				}
-
-				c.Args = blobPath
-			}
-
-			fn(api.ProgressResponse{Status: "creating adapter layer"})
-			bin, err := os.Open(realpath(modelFileDir, c.Args))
-			if err != nil {
-				return err
-			}
-			defer bin.Close()
-
-			_, size, err := llm.DecodeGGML(bin)
+		case "license", "template", "system":
+			blob := strings.NewReader(c.Args)
+			layer, err := NewLayer(blob, mediatype)
 			if err != nil {
 				return err
 			}
 
-			sr := io.NewSectionReader(bin, 0, size)
-			layer, err := NewLayer(sr, mediatype)
-			if err != nil {
-				return err
+			if c.Name != "license" {
+				// replace
+				layers = slices.DeleteFunc(layers, func(layer *Layer) bool {
+					return layer.MediaType == mediatype
+				})
 			}
 
-			layers.Add(layer)
-		case "license":
-			fn(api.ProgressResponse{Status: "creating license layer"})
-
-			bin := strings.NewReader(c.Args)
-			layer, err := NewLayer(bin, mediatype)
-			if err != nil {
-				return err
-			}
-
-			layers.Add(layer)
-		case "template", "system":
-			fn(api.ProgressResponse{Status: fmt.Sprintf("creating %s layer", c.Name)})
-
-			bin := strings.NewReader(c.Args)
-			layer, err := NewLayer(bin, mediatype)
-			if err != nil {
-				return err
-			}
-
-			layers.Replace(layer)
+			layers = append(layers, layer)
 		case "message":
-			messages = append(messages, c.Args)
+			role, content, ok := strings.Cut(c.Args, ": ")
+			if !ok {
+				return fmt.Errorf("invalid message: %s", c.Args)
+			}
+
+			messages = append(messages, &api.Message{Role: role, Content: content})
 		default:
-			params[c.Name] = append(params[c.Name], c.Args)
+			ps, err := api.FormatParams(map[string][]string{c.Name: {c.Args}})
+			if err != nil {
+				return err
+			}
+
+			for k, v := range ps {
+				if ks, ok := parameters[k].([]string); ok {
+					parameters[k] = append(ks, v.([]string)...)
+				} else if vs, ok := v.([]string); ok {
+					parameters[k] = vs
+				} else {
+					parameters[k] = v
+				}
+			}
 		}
 	}
 
-	if len(messages) > 0 {
-		fn(api.ProgressResponse{Status: "creating parameters layer"})
+	var err2 error
+	layers = slices.DeleteFunc(layers, func(layer *Layer) bool {
+		switch layer.MediaType {
+		case "application/vnd.ollama.image.message":
+			// if there are new messages, remove the inherited ones
+			if len(messages) > 0 {
+				return true
+			}
 
-		msgs := make([]api.Message, 0)
+			return false
+		case "application/vnd.ollama.image.params":
+			// merge inherited parameters with new ones
+			r, err := layer.Open()
+			if err != nil {
+				err2 = err
+				return false
+			}
+			defer r.Close()
 
-		for _, m := range messages {
-			// todo: handle images
-			msg := strings.SplitN(m, ": ", 2)
-			msgs = append(msgs, api.Message{Role: msg[0], Content: msg[1]})
+			var ps map[string]any
+			if err := json.NewDecoder(r).Decode(&ps); err != nil {
+				err2 = err
+				return false
+			}
+
+			for k, v := range ps {
+				if _, ok := parameters[k]; !ok {
+					parameters[k] = v
+				}
+			}
+
+			return true
+		default:
+			return false
 		}
+	})
 
+	if err2 != nil {
+		return err2
+	}
+
+	if len(messages) > 0 {
 		var b bytes.Buffer
-		if err := json.NewEncoder(&b).Encode(msgs); err != nil {
+		if err := json.NewEncoder(&b).Encode(messages); err != nil {
 			return err
 		}
 
@@ -552,39 +516,25 @@ func CreateModel(ctx context.Context, name, modelFileDir, quantization string, c
 			return err
 		}
 
-		layers.Replace(layer)
+		layers = append(layers, layer)
 	}
 
-	if len(params) > 0 {
-		fn(api.ProgressResponse{Status: "creating parameters layer"})
-
-		formattedParams, err := api.FormatParams(params)
-		if err != nil {
-			return err
-		}
-
-		for k, v := range fromParams {
-			if _, ok := formattedParams[k]; !ok {
-				formattedParams[k] = v
-			}
-		}
-
+	if len(parameters) > 0 {
 		var b bytes.Buffer
-		if err := json.NewEncoder(&b).Encode(formattedParams); err != nil {
+		if err := json.NewEncoder(&b).Encode(parameters); err != nil {
 			return err
 		}
 
-		fn(api.ProgressResponse{Status: "creating config layer"})
 		layer, err := NewLayer(&b, "application/vnd.ollama.image.params")
 		if err != nil {
 			return err
 		}
 
-		layers.Replace(layer)
+		layers = append(layers, layer)
 	}
 
-	digests := make([]string, len(layers.items))
-	for i, layer := range layers.items {
+	digests := make([]string, len(layers))
+	for i, layer := range layers {
 		digests[i] = layer.Digest
 	}
 
@@ -595,36 +545,38 @@ func CreateModel(ctx context.Context, name, modelFileDir, quantization string, c
 		return err
 	}
 
-	configLayer, err := NewLayer(&b, "application/vnd.docker.container.image.v1+json")
+	layer, err := NewLayer(&b, "application/vnd.docker.container.image.v1+json")
 	if err != nil {
 		return err
 	}
 
-	delete(deleteMap, configLayer.Digest)
+	for _, layer := range append(layers, layer) {
+		if layer.status != "" {
+			fn(api.ProgressResponse{Status: layer.status})
+		}
+	}
 
-	for _, layer := range append(layers.items, configLayer) {
-		committed, err := layer.Commit()
-		if err != nil {
-			return err
+	unref := make(map[string]struct{})
+	if manifest, _, err := GetManifest(ParseModelPath(name)); err == nil {
+		for _, layer := range manifest.Layers {
+			if !slices.Contains(digests, layer.Digest) {
+				unref[layer.Digest] = struct{}{}
+			}
 		}
 
-		status := "writing layer"
-		if !committed {
-			status = "using already created layer"
+		if manifest.Config.Digest != layer.Digest {
+			unref[manifest.Config.Digest] = struct{}{}
 		}
-
-		fn(api.ProgressResponse{Status: fmt.Sprintf("%s %s", status, layer.Digest)})
-
-		delete(deleteMap, layer.Digest)
 	}
 
 	fn(api.ProgressResponse{Status: "writing manifest"})
-	if err := WriteManifest(name, configLayer, layers.items); err != nil {
+	if err := WriteManifest(name, layer, layers); err != nil {
 		return err
 	}
 
-	if noprune := os.Getenv("OLLAMA_NOPRUNE"); noprune == "" {
-		if err := deleteUnusedLayers(nil, deleteMap, false); err != nil {
+	if os.Getenv("OLLAMA_NOPRUNE") == "" && len(unref) > 0 {
+		fn(api.ProgressResponse{Status: "removing unused layers"})
+		if err := deleteUnusedLayers(nil, unref, false); err != nil {
 			return err
 		}
 	}
@@ -633,104 +585,43 @@ func CreateModel(ctx context.Context, name, modelFileDir, quantization string, c
 	return nil
 }
 
-func convertModel(name, path string, fn func(resp api.ProgressResponse)) (string, error) {
-	r, err := zip.OpenReader(path)
-	if err != nil {
-		return "", err
+func CopyModel(src, dst model.Name) error {
+	if !dst.IsFullyQualified() {
+		return model.Unqualified(dst)
 	}
-	defer r.Close()
-
-	tempDir, err := os.MkdirTemp("", "ollama-convert")
-	if err != nil {
-		return "", err
-	}
-	defer os.RemoveAll(tempDir)
-
-	fn(api.ProgressResponse{Status: "unpacking model metadata"})
-	for _, f := range r.File {
-		fpath := filepath.Join(tempDir, f.Name)
-		outFile, err := os.OpenFile(fpath, os.O_WRONLY|os.O_CREATE|os.O_TRUNC, f.Mode())
-		if err != nil {
-			return "", err
-		}
-
-		rc, err := f.Open()
-		if err != nil {
-			return "", err
-		}
-
-		_, err = io.Copy(outFile, rc)
-		if err != nil {
-			return "", err
-		}
-
-		outFile.Close()
-		rc.Close()
+	if !src.IsFullyQualified() {
+		return model.Unqualified(src)
 	}
 
-	mf, err := convert.GetModelFormat(tempDir)
-	if err != nil {
-		return "", err
+	if src.Filepath() == dst.Filepath() {
+		return nil
 	}
 
-	params, err := mf.GetParams(tempDir)
-	if err != nil {
-		return "", err
-	}
-
-	mArch, err := mf.GetModelArch(name, tempDir, params)
-	if err != nil {
-		return "", err
-	}
-
-	fn(api.ProgressResponse{Status: "processing tensors"})
-	if err := mArch.GetTensors(); err != nil {
-		return "", err
-	}
-
-	if err := mArch.LoadVocab(); err != nil {
-		return "", err
-	}
-
-	fn(api.ProgressResponse{Status: "converting model"})
-	path, err = mArch.WriteGGUF()
-	if err != nil {
-		return "", err
-	}
-
-	return path, nil
-}
-
-func CopyModel(src, dest string) error {
-	srcModelPath := ParseModelPath(src)
-	srcPath, err := srcModelPath.GetManifestPath()
+	manifests, err := GetManifestPath()
 	if err != nil {
 		return err
 	}
 
-	destModelPath := ParseModelPath(dest)
-	destPath, err := destModelPath.GetManifestPath()
-	if err != nil {
-		return err
-	}
-	if err := os.MkdirAll(filepath.Dir(destPath), 0o755); err != nil {
+	dstpath := filepath.Join(manifests, dst.Filepath())
+	if err := os.MkdirAll(filepath.Dir(dstpath), 0o755); err != nil {
 		return err
 	}
 
-	// copy the file
-	input, err := os.ReadFile(srcPath)
+	srcpath := filepath.Join(manifests, src.Filepath())
+	srcfile, err := os.Open(srcpath)
 	if err != nil {
-		fmt.Println("Error reading file:", err)
 		return err
 	}
+	defer srcfile.Close()
 
-	err = os.WriteFile(destPath, input, 0o644)
+	dstfile, err := os.Create(dstpath)
 	if err != nil {
-		fmt.Println("Error reading file:", err)
 		return err
 	}
+	defer dstfile.Close()
 
-	return nil
+	_, err = io.Copy(dstfile, srcfile)
+	return err
 }
 
 func deleteUnusedLayers(skipModelPath *ModelPath, deleteMap map[string]struct{}, dryRun bool) error {
@@ -890,67 +781,6 @@ func DeleteModel(name string) error {
 	return nil
 }
 
-func ShowModelfile(model *Model) (string, error) {
-	var mt struct {
-		*Model
-		From       string
-		Parameters map[string][]any
-	}
-
-	mt.Parameters = make(map[string][]any)
-	for k, v := range model.Options {
-		if s, ok := v.([]any); ok {
-			mt.Parameters[k] = s
-			continue
-		}
-
-		mt.Parameters[k] = []any{v}
-	}
-
-	mt.Model = model
-	mt.From = model.ModelPath
-
-	if model.ParentModel != "" {
-		mt.From = model.ParentModel
-	}
-
-	modelFile := `# Modelfile generated by "ollama show"
-# To build a new Modelfile based on this one, replace the FROM line with:
-# FROM {{ .ShortName }}
-
-FROM {{ .From }}
-TEMPLATE """{{ .Template }}"""
-
-{{- if .System }}
-SYSTEM """{{ .System }}"""
-{{- end }}
-
-{{- range $adapter := .AdapterPaths }}
-ADAPTER {{ $adapter }}
-{{- end }}
-
-{{- range $k, $v := .Parameters }}
-{{- range $parameter := $v }}
-PARAMETER {{ $k }} {{ printf "%#v" $parameter }}
-{{- end }}
-{{- end }}`
-
-	tmpl, err := template.New("").Parse(modelFile)
-	if err != nil {
-		slog.Info(fmt.Sprintf("error parsing template: %q", err))
-		return "", err
-	}
-
-	var buf bytes.Buffer
-
-	if err = tmpl.Execute(&buf, mt); err != nil {
-		slog.Info(fmt.Sprintf("error executing template: %q", err))
-		return "", err
-	}
-
-	return buf.String(), nil
-}
-
 func PushModel(ctx context.Context, name string, regOpts *registryOptions, fn func(api.ProgressResponse)) error {
 	mp := ParseModelPath(name)
 	fn(api.ProgressResponse{Status: "retrieving manifest"})
@@ -972,9 +802,6 @@ func PushModel(ctx context.Context, name string, regOpts *registryOptions, fn fu
 	for _, layer := range layers {
 		if err := uploadBlob(ctx, mp, layer, regOpts, fn); err != nil {
 			slog.Info(fmt.Sprintf("error uploading blob: %v", err))
-			if errors.Is(err, errUnauthorized) {
-				return fmt.Errorf("unable to push %s, make sure this namespace exists and you are authorized to push to it", ParseModelPath(name).GetNamespaceRepository())
-			}
 			return err
 		}
 	}
@@ -1137,9 +964,40 @@ func GetSHA256Digest(r io.Reader) (string, int64) {
 	return fmt.Sprintf("sha256:%x", h.Sum(nil)), n
 }
 
-var errUnauthorized = errors.New("unauthorized")
+var errUnauthorized = fmt.Errorf("unauthorized: access denied")
+
+// getTokenSubject returns the subject of a JWT token, it does not validate the token
+func getTokenSubject(token string) string {
+	parts := strings.Split(token, ".")
+	if len(parts) != 3 {
+		slog.Error("jwt token does not contain 3 parts")
+		return ""
+	}
+
+	payload := parts[1]
+	payloadBytes, err := base64.RawURLEncoding.DecodeString(payload)
+	if err != nil {
+		slog.Error(fmt.Sprintf("failed to decode jwt payload: %v", err))
+		return ""
+	}
+
+	var payloadMap map[string]interface{}
+	if err := json.Unmarshal(payloadBytes, &payloadMap); err != nil {
+		slog.Error(fmt.Sprintf("failed to unmarshal payload JSON: %v", err))
+		return ""
+	}
+
+	sub, ok := payloadMap["sub"]
+	if !ok {
+		slog.Error("jwt does not contain 'sub' field")
+		return ""
+	}
+
+	return fmt.Sprintf("%s", sub)
+}
 
 func makeRequestWithRetry(ctx context.Context, method string, requestURL *url.URL, headers http.Header, body io.ReadSeeker, regOpts *registryOptions) (*http.Response, error) {
+	anonymous := true // access will default to anonymous if no user is found associated with the public key
 	for i := 0; i < 2; i++ {
 		resp, err := makeRequest(ctx, method, requestURL, headers, body, regOpts)
 		if err != nil {
@@ -1158,6 +1016,7 @@ func makeRequestWithRetry(ctx context.Context, method string, requestURL *url.UR
 			if err != nil {
 				return nil, err
 			}
+			anonymous = getTokenSubject(token) == "anonymous"
 			regOpts.Token = token
 			if body != nil {
 				_, err = body.Seek(0, io.SeekStart)
@@ -1178,6 +1037,16 @@ func makeRequestWithRetry(ctx context.Context, method string, requestURL *url.UR
 		}
 	}
 
+	if anonymous {
+		// no user is associated with the public key, and the request requires non-anonymous access
+		pubKey, nestedErr := auth.GetPublicKey()
+		if nestedErr != nil {
+			slog.Error(fmt.Sprintf("couldn't get public key: %v", nestedErr))
+			return nil, errUnauthorized
+		}
+		return nil, &errtypes.UnknownOllamaKey{Key: pubKey}
+	}
+	// user is associated with the public key, but is not authorized to make the request
 	return nil, errUnauthorized
 }
 

server/layer.go

@@ -5,39 +5,14 @@ import (
 	"fmt"
 	"io"
 	"os"
-	"strings"
-
-	"golang.org/x/exp/slices"
 )
 
-type Layers struct {
-	items []*Layer
-}
-
-func (ls *Layers) Add(layer *Layer) {
-	if layer.Size > 0 {
-		ls.items = append(ls.items, layer)
-	}
-}
-
-func (ls *Layers) Replace(layer *Layer) {
-	if layer.Size > 0 {
-		mediatype := layer.MediaType
-		layers := slices.DeleteFunc(ls.items, func(l *Layer) bool {
-			return l.MediaType == mediatype
-		})
-
-		ls.items = append(layers, layer)
-	}
-}
-
 type Layer struct {
 	MediaType string `json:"mediaType"`
 	Digest    string `json:"digest"`
 	Size      int64  `json:"size"`
 	From      string `json:"from,omitempty"`
-
-	tempFileName string
+	status    string
 }
 
 func NewLayer(r io.Reader, mediatype string) (*Layer, error) {
@@ -46,14 +21,12 @@ func NewLayer(r io.Reader, mediatype string) (*Layer, error) {
 		return nil, err
 	}
 
-	const delimiter = "-"
-
-	pattern := strings.Join([]string{"sha256", "*-partial"}, delimiter)
-	temp, err := os.CreateTemp(blobs, pattern)
+	temp, err := os.CreateTemp(blobs, "sha256-")
 	if err != nil {
 		return nil, err
 	}
 	defer temp.Close()
+	defer os.Remove(temp.Name())
 
 	sha256sum := sha256.New()
 	n, err := io.Copy(io.MultiWriter(temp, sha256sum), r)
@@ -61,11 +34,29 @@ func NewLayer(r io.Reader, mediatype string) (*Layer, error) {
 		return nil, err
 	}
 
+	if err := temp.Close(); err != nil {
+		return nil, err
+	}
+
+	digest := fmt.Sprintf("sha256:%x", sha256sum.Sum(nil))
+	blob, err := GetBlobsPath(digest)
+	if err != nil {
+		return nil, err
+	}
+
+	status := "using existing layer"
+	if _, err := os.Stat(blob); err != nil {
+		status = "creating new layer"
+		if err := os.Rename(temp.Name(), blob); err != nil {
+			return nil, err
+		}
+	}
+
 	return &Layer{
-		MediaType:    mediatype,
-		Digest:       fmt.Sprintf("sha256:%x", sha256sum.Sum(nil)),
-		Size:         n,
-		tempFileName: temp.Name(),
+		MediaType: mediatype,
+		Digest:    digest,
+		Size:      n,
+		status:    fmt.Sprintf("%s %s", status, digest),
 	}, nil
 }
 
@@ -85,21 +76,15 @@ func NewLayerFromLayer(digest, mediatype, from string) (*Layer, error) {
 		Digest:    digest,
 		Size:      fi.Size(),
 		From:      from,
+		status:    fmt.Sprintf("using existing layer %s", digest),
 	}, nil
 }
 
-func (l *Layer) Commit() (bool, error) {
-	// always remove temp
-	defer os.Remove(l.tempFileName)
-
+func (l *Layer) Open() (io.ReadCloser, error) {
 	blob, err := GetBlobsPath(l.Digest)
 	if err != nil {
-		return false, err
+		return nil, err
 	}
 
-	if _, err := os.Stat(blob); err != nil {
-		return true, os.Rename(l.tempFileName, blob)
-	}
-
-	return false, nil
+	return os.Open(blob)
 }

server/model.go

@@ -0,0 +1,259 @@
+package server
+
+import (
+	"archive/zip"
+	"bytes"
+	"context"
+	"errors"
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+	"path/filepath"
+
+	"github.com/ollama/ollama/api"
+	"github.com/ollama/ollama/convert"
+	"github.com/ollama/ollama/llm"
+	"github.com/ollama/ollama/types/model"
+)
+
+type layerWithGGML struct {
+	*Layer
+	*llm.GGML
+}
+
+func parseFromModel(ctx context.Context, name model.Name, fn func(api.ProgressResponse)) (layers []*layerWithGGML, err error) {
+	modelpath := ParseModelPath(name.String())
+	manifest, _, err := GetManifest(modelpath)
+	switch {
+	case errors.Is(err, os.ErrNotExist):
+		if err := PullModel(ctx, name.String(), &registryOptions{}, fn); err != nil {
+			return nil, err
+		}
+
+		modelpath = ParseModelPath(name.String())
+		manifest, _, err = GetManifest(modelpath)
+		if err != nil {
+			return nil, err
+		}
+	case err != nil:
+		return nil, err
+	}
+
+	for _, layer := range manifest.Layers {
+		layer, err := NewLayerFromLayer(layer.Digest, layer.MediaType, modelpath.GetShortTagname())
+		if err != nil {
+			return nil, err
+		}
+
+		switch layer.MediaType {
+		case "application/vnd.ollama.image.model",
+			"application/vnd.ollama.image.projector",
+			"application/vnd.ollama.image.adapter":
+			blobpath, err := GetBlobsPath(layer.Digest)
+			if err != nil {
+				return nil, err
+			}
+
+			blob, err := os.Open(blobpath)
+			if err != nil {
+				return nil, err
+			}
+			defer blob.Close()
+
+			ggml, _, err := llm.DecodeGGML(blob)
+			if err != nil {
+				return nil, err
+			}
+
+			layers = append(layers, &layerWithGGML{layer, ggml})
+		default:
+			layers = append(layers, &layerWithGGML{layer, nil})
+		}
+
+	}
+
+	return layers, nil
+}
+
+func parseFromZipFile(_ context.Context, file *os.File, fn func(api.ProgressResponse)) (layers []*layerWithGGML, err error) {
+	stat, err := file.Stat()
+	if err != nil {
+		return nil, err
+	}
+
+	r, err := zip.NewReader(file, stat.Size())
+	if err != nil {
+		return nil, err
+	}
+
+	tempdir, err := os.MkdirTemp(filepath.Dir(file.Name()), "")
+	if err != nil {
+		return nil, err
+	}
+	defer os.RemoveAll(tempdir)
+
+	fn(api.ProgressResponse{Status: "unpacking model metadata"})
+	for _, f := range r.File {
+		// TODO(mxyng): this should not write out all files to disk
+		outfile, err := os.Create(filepath.Join(tempdir, f.Name))
+		if err != nil {
+			return nil, err
+		}
+
+		infile, err := f.Open()
+		if err != nil {
+			return nil, err
+		}
+
+		if _, err = io.Copy(outfile, infile); err != nil {
+			return nil, err
+		}
+
+		if err := outfile.Close(); err != nil {
+			return nil, err
+		}
+
+		if err := infile.Close(); err != nil {
+			return nil, err
+		}
+	}
+
+	mf, err := convert.GetModelFormat(tempdir)
+	if err != nil {
+		return nil, err
+	}
+
+	params, err := mf.GetParams(tempdir)
+	if err != nil {
+		return nil, err
+	}
+
+	mArch, err := mf.GetModelArch("", tempdir, params)
+	if err != nil {
+		return nil, err
+	}
+
+	fn(api.ProgressResponse{Status: "processing tensors"})
+	if err := mArch.GetTensors(); err != nil {
+		return nil, err
+	}
+
+	if err := mArch.LoadVocab(); err != nil {
+		return nil, err
+	}
+
+	fn(api.ProgressResponse{Status: "converting model"})
+
+	// TODO(mxyng): this should write directly into a layer
+	// e.g. NewLayer(arch.Reader(), "application/vnd.ollama.image.model")
+	temp, err := os.CreateTemp(tempdir, "fp16")
+	if err != nil {
+		return nil, err
+	}
+	defer temp.Close()
+	defer os.Remove(temp.Name())
+
+	if err = mArch.WriteGGUF(temp); err != nil {
+		return nil, err
+	}
+
+	if _, err := temp.Seek(0, io.SeekStart); err != nil {
+		return nil, err
+	}
+
+	layer, err := NewLayer(temp, "application/vnd.ollama.image.model")
+	if err != nil {
+		return nil, fmt.Errorf("aaa: %w", err)
+	}
+
+	blobpath, err := GetBlobsPath(layer.Digest)
+	if err != nil {
+		return nil, err
+	}
+
+	bin, err := os.Open(blobpath)
+	if err != nil {
+		return nil, err
+	}
+	defer bin.Close()
+
+	ggml, _, err := llm.DecodeGGML(bin)
+	if err != nil {
+		return nil, err
+	}
+
+	layer, err = NewLayerFromLayer(layer.Digest, layer.MediaType, "")
+	if err != nil {
+		return nil, err
+	}
+
+	layers = append(layers, &layerWithGGML{layer, ggml})
+	return layers, nil
+}
+
+func parseFromFile(ctx context.Context, file *os.File, fn func(api.ProgressResponse)) (layers []*layerWithGGML, err error) {
+	sr := io.NewSectionReader(file, 0, 512)
+	contentType, err := detectContentType(sr)
+	if err != nil {
+		return nil, err
+	}
+
+	switch contentType {
+	case "gguf", "ggla":
+		// noop
+	case "application/zip":
+		return parseFromZipFile(ctx, file, fn)
+	default:
+		return nil, fmt.Errorf("unsupported content type: %s", contentType)
+	}
+
+	stat, err := file.Stat()
+	if err != nil {
+		return nil, err
+	}
+
+	var offset int64
+	for offset < stat.Size() {
+		ggml, n, err := llm.DecodeGGML(file)
+		if errors.Is(err, io.EOF) {
+			break
+		} else if err != nil {
+			return nil, err
+		}
+
+		mediatype := "application/vnd.ollama.image.model"
+		if ggml.Name() == "ggla" {
+			mediatype = "application/vnd.ollama.image.adapter"
+		} else if ggml.KV().Architecture() == "clip" {
+			mediatype = "application/vnd.ollama.image.projector"
+		}
+
+		layer, err := NewLayer(io.NewSectionReader(file, offset, n), mediatype)
+		if err != nil {
+			return nil, err
+		}
+
+		layers = append(layers, &layerWithGGML{layer, ggml})
+		offset = n
+	}
+
+	return layers, nil
+}
+
+func detectContentType(r io.Reader) (string, error) {
+	var b bytes.Buffer
+	if _, err := io.Copy(&b, r); err != nil {
+		return "", err
+	}
+
+	if contentType := llm.DetectGGMLType(b.Bytes()); contentType != "" {
+		return contentType, nil
+	}
+
+	if contentType := http.DetectContentType(b.Bytes()); contentType != "application/octet-stream" {
+		return contentType, nil
+	}
+
+	return "unknown", nil
+}

server/prompt.go

@@ -91,7 +91,7 @@ func countTokens(tmpl string, system string, prompt string, response string, enc
 }
 
 // ChatPrompt builds up a prompt from a series of messages, truncating based on context window size
-func ChatPrompt(tmpl string, messages []api.Message, window int, encode func(string) ([]int, error)) (string, int, error) {
+func ChatPrompt(tmpl string, messages []api.Message, window int, encode func(string) ([]int, error)) (string, error) {
 	type prompt struct {
 		System   string
 		Prompt   string
@@ -138,7 +138,7 @@ func ChatPrompt(tmpl string, messages []api.Message, window int, encode func(str
 
 			p.Response = msg.Content
 		default:
-			return "", 0, fmt.Errorf("invalid role: %s, role must be one of [system, user, assistant]", msg.Role)
+			return "", fmt.Errorf("invalid role: %s, role must be one of [system, user, assistant]", msg.Role)
 		}
 	}
 
@@ -151,7 +151,7 @@ func ChatPrompt(tmpl string, messages []api.Message, window int, encode func(str
 	for i, p := range prompts {
 		tokens, err := countTokens(tmpl, p.System, p.Prompt, p.Response, encode)
 		if err != nil {
-			return "", 0, err
+			return "", err
 		}
 
 		prompts[i].tokens = tokens + len(prompts[i].images)*768
@@ -160,17 +160,15 @@ func ChatPrompt(tmpl string, messages []api.Message, window int, encode func(str
 	// truncate images and prompts starting from the beginning of the list
 	// until either one prompt remains or the total tokens fits the context window
 	// TODO (jmorganca): this doesn't account for the context window room required for the response
-	var required int
 	for {
-		required = 0
+		var required int
 		for _, p := range prompts {
 			required += p.tokens
 		}
 
 		required += 1 // for bos token
 
-		// leave ~1024 tokens for generation
-		if required <= max(1024, window/2) {
+		if required <= window {
 			slog.Debug("prompt now fits in context window", "required", required, "window", window)
 			break
 		}
@@ -196,7 +194,7 @@ func ChatPrompt(tmpl string, messages []api.Message, window int, encode func(str
 
 				tokens, err := countTokens(tmpl, prompts[0].System, prompts[0].Prompt, prompts[0].Response, encode)
 				if err != nil {
-					return "", 0, err
+					return "", err
 				}
 
 				prompts[0].tokens = tokens + len(prompts[0].images)*768
@@ -214,10 +212,10 @@ func ChatPrompt(tmpl string, messages []api.Message, window int, encode func(str
 		// last prompt should leave the response unrendered (for completion)
 		rendered, err := Prompt(tmpl, p.System, p.Prompt, p.Response, i == len(prompts)-1)
 		if err != nil {
-			return "", 0, err
+			return "", err
 		}
 		sb.WriteString(rendered)
 	}
 
-	return sb.String(), required, nil
+	return sb.String(), nil
 }

server/prompt_test.go

@@ -192,7 +192,7 @@ func TestChatPrompt(t *testing.T) {
 
 	for _, tc := range tests {
 		t.Run(tc.name, func(t *testing.T) {
-			got, _, err := ChatPrompt(tc.template, tc.messages, tc.window, encode)
+			got, err := ChatPrompt(tc.template, tc.messages, tc.window, encode)
 			if err != nil {
 				t.Errorf("error = %v", err)
 			}

server/routes.go

@@ -1,6 +1,7 @@
 package server
 
 import (
+	"cmp"
 	"context"
 	"encoding/json"
 	"errors"
@@ -15,11 +16,8 @@ import (
 	"os"
 	"os/signal"
 	"path/filepath"
-	"reflect"
-	"runtime"
 	"strconv"
 	"strings"
-	"sync"
 	"syscall"
 	"time"
 
@@ -31,14 +29,15 @@ import (
 	"github.com/ollama/ollama/gpu"
 	"github.com/ollama/ollama/llm"
 	"github.com/ollama/ollama/openai"
-	"github.com/ollama/ollama/parser"
+	"github.com/ollama/ollama/types/model"
 	"github.com/ollama/ollama/version"
 )
 
 var mode string = gin.DebugMode
 
 type Server struct {
-	addr net.Addr
+	addr  net.Addr
+	sched *Scheduler
 }
 
 func init() {
@@ -53,88 +52,8 @@ func init() {
 	gin.SetMode(mode)
 }
 
-var loaded struct {
-	mu sync.Mutex
-
-	llama *llm.LlamaServer
-
-	expireTimer *time.Timer
-
-	model      string
-	adapters   []string
-	projectors []string
-	*api.Options
-}
-
 var defaultSessionDuration = 5 * time.Minute
 
-func unload() {
-	if loaded.llama != nil {
-		loaded.llama.Close()
-	}
-
-	loaded.llama = nil
-	loaded.model = ""
-	loaded.adapters = nil
-	loaded.projectors = nil
-	loaded.Options = nil
-}
-
-// load a model into memory if it is not already loaded, it is up to the caller to lock loaded.mu before calling this function
-func load(c *gin.Context, model *Model, opts api.Options, sessionDuration time.Duration) error {
-	ctx, cancel := context.WithTimeout(c, 10*time.Second)
-	defer cancel()
-
-	needLoad := loaded.llama == nil || // is there a model loaded?
-		loaded.model != model.ModelPath || // has the base model changed?
-		!reflect.DeepEqual(loaded.adapters, model.AdapterPaths) || // have the adapters changed?
-		!reflect.DeepEqual(loaded.projectors, model.ProjectorPaths) || // have the adapters changed?
-		!reflect.DeepEqual(loaded.Options.Runner, opts.Runner) || // have the runner options changed?
-		loaded.llama.Ping(ctx) != nil
-
-	if needLoad {
-		if loaded.llama != nil {
-			slog.Info("changing loaded model")
-			unload()
-		}
-
-		llama, err := llm.NewLlamaServer(model.ModelPath, model.AdapterPaths, model.ProjectorPaths, opts)
-		if err != nil {
-			// some older models are not compatible with newer versions of llama.cpp
-			// show a generalized compatibility error until there is a better way to
-			// check for model compatibility
-			if errors.Is(llm.ErrUnsupportedFormat, err) || strings.Contains(err.Error(), "failed to load model") {
-				err = fmt.Errorf("%v: this model may be incompatible with your version of Ollama. If you previously pulled this model, try updating it by running `ollama pull %s`", err, model.ShortName)
-			}
-
-			return err
-		}
-
-		loaded.model = model.ModelPath
-		loaded.adapters = model.AdapterPaths
-		loaded.projectors = model.ProjectorPaths
-		loaded.llama = llama
-		loaded.Options = &opts
-
-		if err = llama.WaitUntilRunning(); err != nil {
-			slog.Error("error loading llama server", "error", err)
-			unload()
-			return err
-		}
-	}
-
-	if loaded.expireTimer == nil {
-		loaded.expireTimer = time.AfterFunc(sessionDuration, func() {
-			loaded.mu.Lock()
-			defer loaded.mu.Unlock()
-			unload()
-		})
-	}
-
-	loaded.expireTimer.Reset(sessionDuration)
-	return nil
-}
-
 func modelOptions(model *Model, requestOpts map[string]interface{}) (api.Options, error) {
 	opts := api.DefaultOptions()
 	if err := opts.FromMap(model.Options); err != nil {
@@ -154,9 +73,7 @@ func isSupportedImageType(image []byte) bool {
 	return slices.Contains(allowedTypes, contentType)
 }
 
-func GenerateHandler(c *gin.Context) {
-	loaded.mu.Lock()
-	defer loaded.mu.Unlock()
+func (s *Server) GenerateHandler(c *gin.Context) {
 
 	checkpointStart := time.Now()
 	var req api.GenerateRequest
@@ -224,7 +141,16 @@ func GenerateHandler(c *gin.Context) {
 		sessionDuration = req.KeepAlive.Duration
 	}
 
-	if err := load(c, model, opts, sessionDuration); err != nil {
+	rCh, eCh := s.sched.GetRunner(c.Request.Context(), model, opts, sessionDuration)
+	var runner *runnerRef
+	select {
+	case runner = <-rCh:
+	case err = <-eCh:
+		if errors.Is(err, context.Canceled) {
+			c.JSON(499, gin.H{"error": "request canceled"})
+			return
+		}
+
 		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
 		return
 	}
@@ -234,10 +160,9 @@ func GenerateHandler(c *gin.Context) {
 	// of `raw` mode so we need to check for it too
 	if req.Prompt == "" && req.Template == "" && req.System == "" {
 		c.JSON(http.StatusOK, api.GenerateResponse{
-			CreatedAt:  time.Now().UTC(),
-			Model:      req.Model,
-			Done:       true,
-			DoneReason: "load",
+			CreatedAt: time.Now().UTC(),
+			Model:     req.Model,
+			Done:      true,
 		})
 		return
 	}
@@ -276,7 +201,7 @@ func GenerateHandler(c *gin.Context) {
 
 		sb.Reset()
 		if req.Context != nil {
-			prev, err := loaded.llama.Detokenize(c.Request.Context(), req.Context)
+			prev, err := runner.llama.Detokenize(c.Request.Context(), req.Context)
 			if err != nil {
 				c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
 				return
@@ -290,14 +215,6 @@ func GenerateHandler(c *gin.Context) {
 		prompt = sb.String()
 	}
 
-	tokens, err := loaded.llama.Tokenize(c.Request.Context(), prompt)
-	if err != nil {
-		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
-		return
-	}
-
-	opts.NumPredict = max(opts.NumCtx-len(tokens), 0)
-
 	slog.Debug("generate handler", "prompt", prompt)
 
 	ch := make(chan any)
@@ -306,9 +223,6 @@ func GenerateHandler(c *gin.Context) {
 		defer close(ch)
 
 		fn := func(r llm.CompletionResponse) {
-			// Update model expiration
-			loaded.expireTimer.Reset(sessionDuration)
-
 			// Build up the full response
 			if _, err := generated.WriteString(r.Content); err != nil {
 				ch <- gin.H{"error": err.Error()}
@@ -316,11 +230,10 @@ func GenerateHandler(c *gin.Context) {
 			}
 
 			resp := api.GenerateResponse{
-				Model:      req.Model,
-				CreatedAt:  time.Now().UTC(),
-				Done:       r.Done,
-				DoneReason: r.DoneReason,
-				Response:   r.Content,
+				Model:     req.Model,
+				CreatedAt: time.Now().UTC(),
+				Done:      r.Done,
+				Response:  r.Content,
 				Metrics: api.Metrics{
 					PromptEvalCount:    r.PromptEvalCount,
 					PromptEvalDuration: r.PromptEvalDuration,
@@ -341,7 +254,7 @@ func GenerateHandler(c *gin.Context) {
 					}
 
 					// TODO (jmorganca): encode() should not strip special tokens
-					tokens, err := loaded.llama.Tokenize(c.Request.Context(), p)
+					tokens, err := runner.llama.Tokenize(c.Request.Context(), p)
 					if err != nil {
 						ch <- gin.H{"error": err.Error()}
 						return
@@ -369,7 +282,7 @@ func GenerateHandler(c *gin.Context) {
 			Images:  images,
 			Options: opts,
 		}
-		if err := loaded.llama.Completion(c.Request.Context(), req, fn); err != nil {
+		if err := runner.llama.Completion(c.Request.Context(), req, fn); err != nil {
 			ch <- gin.H{"error": err.Error()}
 		}
 	}()
@@ -431,10 +344,7 @@ func getDefaultSessionDuration() time.Duration {
 	return defaultSessionDuration
 }
 
-func EmbeddingsHandler(c *gin.Context) {
-	loaded.mu.Lock()
-	defer loaded.mu.Unlock()
-
+func (s *Server) EmbeddingsHandler(c *gin.Context) {
 	var req api.EmbeddingRequest
 	err := c.ShouldBindJSON(&req)
 	switch {
@@ -479,7 +389,16 @@ func EmbeddingsHandler(c *gin.Context) {
 		sessionDuration = req.KeepAlive.Duration
 	}
 
-	if err := load(c, model, opts, sessionDuration); err != nil {
+	rCh, eCh := s.sched.GetRunner(c.Request.Context(), model, opts, sessionDuration)
+	var runner *runnerRef
+	select {
+	case runner = <-rCh:
+	case err = <-eCh:
+		if errors.Is(err, context.Canceled) {
+			c.JSON(499, gin.H{"error": "request canceled"})
+			return
+		}
+
 		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
 		return
 	}
@@ -490,7 +409,7 @@ func EmbeddingsHandler(c *gin.Context) {
 		return
 	}
 
-	embedding, err := loaded.llama.Embedding(c.Request.Context(), req.Prompt)
+	embedding, err := runner.llama.Embedding(c.Request.Context(), req.Prompt)
 	if err != nil {
 		slog.Info(fmt.Sprintf("embedding generation failed: %v", err))
 		c.JSON(http.StatusInternalServerError, gin.H{"error": "failed to generate embedding"})
@@ -503,7 +422,7 @@ func EmbeddingsHandler(c *gin.Context) {
 	c.JSON(http.StatusOK, resp)
 }
 
-func PullModelHandler(c *gin.Context) {
+func (s *Server) PullModelHandler(c *gin.Context) {
 	var req api.PullRequest
 	err := c.ShouldBindJSON(&req)
 	switch {
@@ -552,7 +471,7 @@ func PullModelHandler(c *gin.Context) {
 	streamResponse(c, ch)
 }
 
-func PushModelHandler(c *gin.Context) {
+func (s *Server) PushModelHandler(c *gin.Context) {
 	var req api.PushRequest
 	err := c.ShouldBindJSON(&req)
 	switch {
@@ -601,30 +520,19 @@ func PushModelHandler(c *gin.Context) {
 	streamResponse(c, ch)
 }
 
-func CreateModelHandler(c *gin.Context) {
+func (s *Server) CreateModelHandler(c *gin.Context) {
 	var req api.CreateRequest
-	err := c.ShouldBindJSON(&req)
-	switch {
-	case errors.Is(err, io.EOF):
+	if err := c.ShouldBindJSON(&req); errors.Is(err, io.EOF) {
 		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": "missing request body"})
 		return
-	case err != nil:
+	} else if err != nil {
 		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
 
-	var model string
-	if req.Model != "" {
-		model = req.Model
-	} else if req.Name != "" {
-		model = req.Name
-	} else {
-		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": "model is required"})
-		return
-	}
-
-	if err := ParseModelPath(model).Validate(); err != nil {
-		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+	name := model.ParseName(cmp.Or(req.Model, req.Name))
+	if !name.IsValid() {
+		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": "invalid model name"})
 		return
 	}
 
@@ -633,19 +541,19 @@ func CreateModelHandler(c *gin.Context) {
 		return
 	}
 
-	var modelfile io.Reader = strings.NewReader(req.Modelfile)
+	var r io.Reader = strings.NewReader(req.Modelfile)
 	if req.Path != "" && req.Modelfile == "" {
-		mf, err := os.Open(req.Path)
+		f, err := os.Open(req.Path)
 		if err != nil {
 			c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": fmt.Sprintf("error reading modelfile: %s", err)})
 			return
 		}
-		defer mf.Close()
+		defer f.Close()
 
-		modelfile = mf
+		r = f
 	}
 
-	commands, err := parser.Parse(modelfile)
+	modelfile, err := model.ParseFile(r)
 	if err != nil {
 		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
@@ -661,7 +569,7 @@ func CreateModelHandler(c *gin.Context) {
 		ctx, cancel := context.WithCancel(c.Request.Context())
 		defer cancel()
 
-		if err := CreateModel(ctx, model, filepath.Dir(req.Path), req.Quantization, commands, fn); err != nil {
+		if err := CreateModel(ctx, name.String(), filepath.Dir(req.Path), strings.ToUpper(req.Quantization), modelfile, fn); err != nil {
 			ch <- gin.H{"error": err.Error()}
 		}
 	}()
@@ -674,7 +582,7 @@ func CreateModelHandler(c *gin.Context) {
 	streamResponse(c, ch)
 }
 
-func DeleteModelHandler(c *gin.Context) {
+func (s *Server) DeleteModelHandler(c *gin.Context) {
 	var req api.DeleteRequest
 	err := c.ShouldBindJSON(&req)
 	switch {
@@ -719,7 +627,7 @@ func DeleteModelHandler(c *gin.Context) {
 	c.JSON(http.StatusOK, nil)
 }
 
-func ShowModelHandler(c *gin.Context) {
+func (s *Server) ShowModelHandler(c *gin.Context) {
 	var req api.ShowRequest
 	err := c.ShouldBindJSON(&req)
 	switch {
@@ -809,17 +717,17 @@ func GetModelInfo(req api.ShowRequest) (*api.ShowResponse, error) {
 		}
 	}
 
-	mf, err := ShowModelfile(model)
-	if err != nil {
-		return nil, err
-	}
-
-	resp.Modelfile = mf
+	var sb strings.Builder
+	fmt.Fprintln(&sb, "# Modelfile generate by \"ollama show\"")
+	fmt.Fprintln(&sb, "# To build a new Modelfile based on this, replace FROM with:")
+	fmt.Fprintf(&sb, "# FROM %s\n\n", model.ShortName)
+	fmt.Fprint(&sb, model.String())
+	resp.Modelfile = sb.String()
 
 	return resp, nil
 }
 
-func ListModelsHandler(c *gin.Context) {
+func (s *Server) ListModelsHandler(c *gin.Context) {
 	models := make([]api.ModelResponse, 0)
 	manifestsPath, err := GetManifestPath()
 	if err != nil {
@@ -879,39 +787,36 @@ func ListModelsHandler(c *gin.Context) {
 	c.JSON(http.StatusOK, api.ListResponse{Models: models})
 }
 
-func CopyModelHandler(c *gin.Context) {
-	var req api.CopyRequest
-	err := c.ShouldBindJSON(&req)
-	switch {
-	case errors.Is(err, io.EOF):
+func (s *Server) CopyModelHandler(c *gin.Context) {
+	var r api.CopyRequest
+	if err := c.ShouldBindJSON(&r); errors.Is(err, io.EOF) {
 		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": "missing request body"})
 		return
-	case err != nil:
+	} else if err != nil {
 		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
 
-	if req.Source == "" || req.Destination == "" {
-		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": "source add destination are required"})
+	src := model.ParseName(r.Source)
+	if !src.IsValid() {
+		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": fmt.Sprintf("source %q is invalid", r.Source)})
 		return
 	}
 
-	if err := ParseModelPath(req.Destination).Validate(); err != nil {
-		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": err.Error()})
+	dst := model.ParseName(r.Destination)
+	if !dst.IsValid() {
+		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": fmt.Sprintf("destination %q is invalid", r.Source)})
 		return
 	}
 
-	if err := CopyModel(req.Source, req.Destination); err != nil {
-		if os.IsNotExist(err) {
-			c.JSON(http.StatusNotFound, gin.H{"error": fmt.Sprintf("model '%s' not found", req.Source)})
-		} else {
-			c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
-		}
-		return
+	if err := CopyModel(src, dst); errors.Is(err, os.ErrNotExist) {
+		c.JSON(http.StatusNotFound, gin.H{"error": fmt.Sprintf("model %q not found", r.Source)})
+	} else if err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
 	}
 }
 
-func HeadBlobHandler(c *gin.Context) {
+func (s *Server) HeadBlobHandler(c *gin.Context) {
 	path, err := GetBlobsPath(c.Param("digest"))
 	if err != nil {
 		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": err.Error()})
@@ -926,7 +831,7 @@ func HeadBlobHandler(c *gin.Context) {
 	c.Status(http.StatusOK)
 }
 
-func CreateBlobHandler(c *gin.Context) {
+func (s *Server) CreateBlobHandler(c *gin.Context) {
 	path, err := GetBlobsPath(c.Param("digest"))
 	if err != nil {
 		c.AbortWithStatusJSON(http.StatusBadRequest, gin.H{"error": err.Error()})
@@ -956,11 +861,6 @@ func CreateBlobHandler(c *gin.Context) {
 		return
 	}
 
-	if _, err := layer.Commit(); err != nil {
-		c.AbortWithStatusJSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
-		return
-	}
-
 	c.Status(http.StatusCreated)
 }
 
@@ -1073,27 +973,27 @@ func (s *Server) GenerateRoutes() http.Handler {
 		allowedHostsMiddleware(s.addr),
 	)
 
-	r.POST("/api/pull", PullModelHandler)
-	r.POST("/api/generate", GenerateHandler)
-	r.POST("/api/chat", ChatHandler)
-	r.POST("/api/embeddings", EmbeddingsHandler)
-	r.POST("/api/create", CreateModelHandler)
-	r.POST("/api/push", PushModelHandler)
-	r.POST("/api/copy", CopyModelHandler)
-	r.DELETE("/api/delete", DeleteModelHandler)
-	r.POST("/api/show", ShowModelHandler)
-	r.POST("/api/blobs/:digest", CreateBlobHandler)
-	r.HEAD("/api/blobs/:digest", HeadBlobHandler)
+	r.POST("/api/pull", s.PullModelHandler)
+	r.POST("/api/generate", s.GenerateHandler)
+	r.POST("/api/chat", s.ChatHandler)
+	r.POST("/api/embeddings", s.EmbeddingsHandler)
+	r.POST("/api/create", s.CreateModelHandler)
+	r.POST("/api/push", s.PushModelHandler)
+	r.POST("/api/copy", s.CopyModelHandler)
+	r.DELETE("/api/delete", s.DeleteModelHandler)
+	r.POST("/api/show", s.ShowModelHandler)
+	r.POST("/api/blobs/:digest", s.CreateBlobHandler)
+	r.HEAD("/api/blobs/:digest", s.HeadBlobHandler)
 
 	// Compatibility endpoints
-	r.POST("/v1/chat/completions", openai.Middleware(), ChatHandler)
+	r.POST("/v1/chat/completions", openai.Middleware(), s.ChatHandler)
 
 	for _, method := range []string{http.MethodGet, http.MethodHead} {
 		r.Handle(method, "/", func(c *gin.Context) {
 			c.String(http.StatusOK, "Ollama is running")
 		})
 
-		r.Handle(method, "/api/tags", ListModelsHandler)
+		r.Handle(method, "/api/tags", s.ListModelsHandler)
 		r.Handle(method, "/api/version", func(c *gin.Context) {
 			c.JSON(http.StatusOK, gin.H{"version": version.Version})
 		})
@@ -1147,7 +1047,9 @@ func Serve(ln net.Listener) error {
 		}
 	}
 
-	s := &Server{addr: ln.Addr()}
+	ctx, done := context.WithCancel(context.Background())
+	sched := InitScheduler(ctx)
+	s := &Server{addr: ln.Addr(), sched: sched}
 	r := s.GenerateRoutes()
 
 	slog.Info(fmt.Sprintf("Listening on %s (version %s)", ln.Addr(), version.Version))
@@ -1160,7 +1062,8 @@ func Serve(ln net.Listener) error {
 	signal.Notify(signals, syscall.SIGINT, syscall.SIGTERM)
 	go func() {
 		<-signals
-		unload()
+		done()
+		sched.unloadAllRunners()
 		gpu.Cleanup()
 		os.Exit(0)
 	}()
@@ -1168,12 +1071,12 @@ func Serve(ln net.Listener) error {
 	if err := llm.Init(); err != nil {
 		return fmt.Errorf("unable to initialize llm library %w", err)
 	}
-	if runtime.GOOS == "linux" { // TODO - windows too
-		// check compatibility to log warnings
-		if _, err := gpu.CheckVRAM(); err != nil {
-			slog.Info(err.Error())
-		}
-	}
+
+	s.sched.Run(ctx)
+
+	// At startup we retrieve GPU information so we can get log messages before loading a model
+	// This will log warnings to the log in case we have problems with detected GPUs
+	_ = gpu.GetGPUInfo()
 
 	return srvr.Serve(ln)
 }
@@ -1229,23 +1132,20 @@ func streamResponse(c *gin.Context, ch chan any) {
 }
 
 // ChatPrompt builds up a prompt from a series of messages for the currently `loaded` model
-func chatPrompt(ctx context.Context, template string, messages []api.Message, numCtx int) (string, int, error) {
+func chatPrompt(ctx context.Context, runner *runnerRef, template string, messages []api.Message, numCtx int) (string, error) {
 	encode := func(s string) ([]int, error) {
-		return loaded.llama.Tokenize(ctx, s)
+		return runner.llama.Tokenize(ctx, s)
 	}
 
-	prompt, tokens, err := ChatPrompt(template, messages, numCtx, encode)
+	prompt, err := ChatPrompt(template, messages, numCtx, encode)
 	if err != nil {
-		return "", 0, err
+		return "", err
 	}
 
-	return prompt, tokens, nil
+	return prompt, nil
 }
 
-func ChatHandler(c *gin.Context) {
-	loaded.mu.Lock()
-	defer loaded.mu.Unlock()
-
+func (s *Server) ChatHandler(c *gin.Context) {
 	checkpointStart := time.Now()
 
 	var req api.ChatRequest
@@ -1302,7 +1202,16 @@ func ChatHandler(c *gin.Context) {
 		sessionDuration = req.KeepAlive.Duration
 	}
 
-	if err := load(c, model, opts, sessionDuration); err != nil {
+	rCh, eCh := s.sched.GetRunner(c.Request.Context(), model, opts, sessionDuration)
+	var runner *runnerRef
+	select {
+	case runner = <-rCh:
+	case err = <-eCh:
+		if errors.Is(err, context.Canceled) {
+			c.JSON(499, gin.H{"error": "request canceled"})
+			return
+		}
+
 		c.JSON(http.StatusInternalServerError, gin.H{"error": err.Error()})
 		return
 	}
@@ -1319,22 +1228,19 @@ func ChatHandler(c *gin.Context) {
 		}, req.Messages...)
 	}
 
-	prompt, tokens, err := chatPrompt(c.Request.Context(), model.Template, req.Messages, opts.NumCtx)
+	prompt, err := chatPrompt(c.Request.Context(), runner, model.Template, req.Messages, opts.NumCtx)
 	if err != nil {
 		c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
 		return
 	}
 
-	opts.NumPredict = max(opts.NumCtx-tokens, 0)
-
 	// an empty request loads the model
 	if len(req.Messages) == 0 || prompt == "" {
 		resp := api.ChatResponse{
-			CreatedAt:  time.Now().UTC(),
-			Model:      req.Model,
-			Done:       true,
-			DoneReason: "load",
-			Message:    api.Message{Role: "assistant"},
+			CreatedAt: time.Now().UTC(),
+			Model:     req.Model,
+			Done:      true,
+			Message:   api.Message{Role: "assistant"},
 		}
 		c.JSON(http.StatusOK, resp)
 		return
@@ -1365,15 +1271,12 @@ func ChatHandler(c *gin.Context) {
 		defer close(ch)
 
 		fn := func(r llm.CompletionResponse) {
-			// Update model expiration
-			loaded.expireTimer.Reset(sessionDuration)
 
 			resp := api.ChatResponse{
-				Model:      req.Model,
-				CreatedAt:  time.Now().UTC(),
-				Message:    api.Message{Role: "assistant", Content: r.Content},
-				Done:       r.Done,
-				DoneReason: r.DoneReason,
+				Model:     req.Model,
+				CreatedAt: time.Now().UTC(),
+				Message:   api.Message{Role: "assistant", Content: r.Content},
+				Done:      r.Done,
 				Metrics: api.Metrics{
 					PromptEvalCount:    r.PromptEvalCount,
 					PromptEvalDuration: r.PromptEvalDuration,
@@ -1390,7 +1293,7 @@ func ChatHandler(c *gin.Context) {
 			ch <- resp
 		}
 
-		if err := loaded.llama.Completion(c.Request.Context(), llm.CompletionRequest{
+		if err := runner.llama.Completion(c.Request.Context(), llm.CompletionRequest{
 			Prompt:  prompt,
 			Format:  req.Format,
 			Images:  images,

server/routes_test.go

@@ -17,7 +17,7 @@ import (
 	"github.com/stretchr/testify/assert"
 
 	"github.com/ollama/ollama/api"
-	"github.com/ollama/ollama/parser"
+	"github.com/ollama/ollama/types/model"
 	"github.com/ollama/ollama/version"
 )
 
@@ -55,13 +55,13 @@ func Test_Routes(t *testing.T) {
 	createTestModel := func(t *testing.T, name string) {
 		fname := createTestFile(t, "ollama-model")
 
-		modelfile := strings.NewReader(fmt.Sprintf("FROM %s\nPARAMETER seed 42\nPARAMETER top_p 0.9\nPARAMETER stop foo\nPARAMETER stop bar", fname))
-		commands, err := parser.Parse(modelfile)
+		r := strings.NewReader(fmt.Sprintf("FROM %s\nPARAMETER seed 42\nPARAMETER top_p 0.9\nPARAMETER stop foo\nPARAMETER stop bar", fname))
+		modelfile, err := model.ParseFile(r)
 		assert.Nil(t, err)
 		fn := func(resp api.ProgressResponse) {
 			t.Logf("Status: %s", resp.Status)
 		}
-		err = CreateModel(context.TODO(), name, "", "", commands, fn)
+		err = CreateModel(context.TODO(), name, "", "", modelfile, fn)
 		assert.Nil(t, err)
 	}
 
@@ -124,14 +124,12 @@ func Test_Routes(t *testing.T) {
 			Method: http.MethodPost,
 			Path:   "/api/create",
 			Setup: func(t *testing.T, req *http.Request) {
-				f, err := os.CreateTemp(t.TempDir(), "ollama-model")
-				assert.Nil(t, err)
-				defer f.Close()
+				fname := createTestFile(t, "ollama-model")
 
 				stream := false
 				createReq := api.CreateRequest{
 					Name:      "t-bone",
-					Modelfile: fmt.Sprintf("FROM %s", f.Name()),
+					Modelfile: fmt.Sprintf("FROM %s", fname),
 					Stream:    &stream,
 				}
 				jsonData, err := json.Marshal(createReq)
@@ -216,28 +214,25 @@ func Test_Routes(t *testing.T) {
 	httpSrv := httptest.NewServer(router)
 	t.Cleanup(httpSrv.Close)
 
-	workDir, err := os.MkdirTemp("", "ollama-test")
-	assert.Nil(t, err)
-	defer os.RemoveAll(workDir)
-	os.Setenv("OLLAMA_MODELS", workDir)
+	t.Setenv("OLLAMA_MODELS", t.TempDir())
 
 	for _, tc := range testCases {
-		t.Logf("Running Test: [%s]", tc.Name)
-		u := httpSrv.URL + tc.Path
-		req, err := http.NewRequestWithContext(context.TODO(), tc.Method, u, nil)
-		assert.Nil(t, err)
+		t.Run(tc.Name, func(t *testing.T) {
+			u := httpSrv.URL + tc.Path
+			req, err := http.NewRequestWithContext(context.TODO(), tc.Method, u, nil)
+			assert.Nil(t, err)
 
-		if tc.Setup != nil {
-			tc.Setup(t, req)
-		}
+			if tc.Setup != nil {
+				tc.Setup(t, req)
+			}
 
-		resp, err := httpSrv.Client().Do(req)
-		assert.Nil(t, err)
-		defer resp.Body.Close()
-
-		if tc.Expected != nil {
-			tc.Expected(t, resp)
-		}
+			resp, err := httpSrv.Client().Do(req)
+			assert.Nil(t, err)
+			defer resp.Body.Close()
 
+			if tc.Expected != nil {
+				tc.Expected(t, resp)
+			}
+		})
 	}
 }

server/sched.go

@@ -0,0 +1,563 @@
+package server
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"log/slog"
+	"os"
+	"reflect"
+	"sort"
+	"strconv"
+	"strings"
+	"sync"
+	"time"
+
+	"github.com/ollama/ollama/api"
+	"github.com/ollama/ollama/format"
+	"github.com/ollama/ollama/gpu"
+	"github.com/ollama/ollama/llm"
+	"golang.org/x/exp/slices"
+)
+
+type LlmRequest struct {
+	ctx             context.Context //nolint:containedctx
+	model           *Model
+	opts            api.Options
+	sessionDuration time.Duration
+	successCh       chan *runnerRef
+	errCh           chan error
+}
+
+type Scheduler struct {
+	pendingReqCh  chan *LlmRequest
+	finishedReqCh chan *LlmRequest
+	expiredCh     chan *runnerRef
+	unloadedCh    chan interface{}
+
+	loaded   map[string]*runnerRef
+	loadedMu sync.Mutex
+
+	loadFn      func(req *LlmRequest, ggml *llm.GGML, gpus gpu.GpuInfoList)
+	newServerFn func(gpus gpu.GpuInfoList, model string, ggml *llm.GGML, adapters []string, projectors []string, opts api.Options) (llm.LlamaServer, error)
+	getGpuFn    func() gpu.GpuInfoList
+}
+
+// TODO set this to zero after a release or two, to enable multiple models by default
+var loadedMax = 1          // Maximum runners; < 1 maps to as many as will fit in VRAM (unlimited for CPU runners)
+var maxQueuedRequests = 10 // TODO configurable
+var numParallel = 1
+
+func InitScheduler(ctx context.Context) *Scheduler {
+	maxRunners := os.Getenv("OLLAMA_MAX_LOADED_MODELS")
+	if maxRunners != "" {
+		m, err := strconv.Atoi(maxRunners)
+		if err != nil {
+			slog.Error("invalid setting", "OLLAMA_MAX_LOADED_MODELS", maxRunners, "error", err)
+		} else {
+			loadedMax = m
+		}
+	}
+	if onp := os.Getenv("OLLAMA_NUM_PARALLEL"); onp != "" {
+		p, err := strconv.Atoi(onp)
+		if err != nil || p <= 0 {
+			slog.Error("invalid parallel setting, must be greater than zero", "OLLAMA_NUM_PARALLEL", onp, "error", err)
+		} else {
+			numParallel = p
+		}
+	}
+
+	sched := &Scheduler{
+		pendingReqCh:  make(chan *LlmRequest, maxQueuedRequests),
+		finishedReqCh: make(chan *LlmRequest, maxQueuedRequests),
+		expiredCh:     make(chan *runnerRef, maxQueuedRequests),
+		unloadedCh:    make(chan interface{}, maxQueuedRequests),
+		loaded:        make(map[string]*runnerRef),
+		newServerFn:   llm.NewLlamaServer,
+		getGpuFn:      gpu.GetGPUInfo,
+	}
+	sched.loadFn = sched.load
+	return sched
+}
+
+// context must be canceled to decrement ref count and release the runner
+func (s *Scheduler) GetRunner(c context.Context, model *Model, opts api.Options, sessionDuration time.Duration) (chan *runnerRef, chan error) {
+	req := &LlmRequest{
+		ctx:             c,
+		model:           model,
+		opts:            opts,
+		sessionDuration: sessionDuration,
+		successCh:       make(chan *runnerRef),
+		errCh:           make(chan error, 1),
+	}
+	// context split across parallel threads
+	opts.NumCtx = opts.NumCtx * numParallel
+	select {
+	case s.pendingReqCh <- req:
+	default:
+		req.errCh <- fmt.Errorf("server busy, please try again.  maximum pending requests exceeded")
+	}
+	return req.successCh, req.errCh
+}
+
+// Returns immediately, spawns go routines for the scheduler which will shutdown when ctx is done
+func (s *Scheduler) Run(ctx context.Context) {
+	slog.Debug("starting llm scheduler")
+	go func() {
+		s.processPending(ctx)
+	}()
+
+	go func() {
+		s.processCompleted(ctx)
+	}()
+}
+
+func (s *Scheduler) processPending(ctx context.Context) {
+	for {
+		select {
+		case <-ctx.Done():
+			slog.Debug("shutting down scheduler pending loop")
+			return
+		case pending := <-s.pendingReqCh:
+			// Block other requests until we get this pending request running
+			for {
+				var runnerToExpire *runnerRef
+				s.loadedMu.Lock()
+				runner := s.loaded[pending.model.ModelPath]
+				loadedCount := len(s.loaded)
+				s.loadedMu.Unlock()
+				if runner != nil {
+					if runner.needsReload(ctx, pending) {
+						runnerToExpire = runner
+					} else {
+						// Runner is usable, return it
+						pending.useLoadedRunner(runner, s.finishedReqCh)
+						break
+					}
+				} else if loadedMax > 0 && loadedCount >= loadedMax {
+					slog.Debug("max runners achieved, unloading one to make room", "runner_count", loadedCount)
+					runnerToExpire = s.findRunnerToUnload(pending)
+				} else {
+					// Either no models are loaded or below loadedMax
+					// Get a refreshed GPU list
+					gpus := s.getGpuFn()
+
+					// Load model for fitting
+					ggml, err := llm.LoadModel(pending.model.ModelPath)
+					if err != nil {
+						pending.errCh <- err
+						break
+					}
+
+					// If we're CPU only mode, just limit by loadedMax above
+					// TODO handle system memory exhaustion
+					if (len(gpus) == 1 && gpus[0].Library == "cpu") || pending.opts.NumGPU == 0 {
+						slog.Debug("cpu mode with existing models, loading")
+						s.loadFn(pending, ggml, gpus)
+						break
+					}
+
+					// No models loaded. Load the model but prefer the best fit.
+					if loadedCount == 0 {
+						slog.Debug("loading first model", "model", pending.model.ModelPath)
+						g := pickBestFitGPUs(pending, ggml, gpus)
+						if g != nil {
+							gpus = g
+						}
+						s.loadFn(pending, ggml, gpus)
+						break
+					}
+
+					// More than one loaded model, so we have to see if the new one fits
+					// Update free memory from currently loaded models
+					s.updateFreeSpace(gpus)
+					gpus = pickBestFitGPUs(pending, ggml, gpus)
+					if gpus != nil {
+						slog.Debug("new model fits with existing models, loading")
+						s.loadFn(pending, ggml, gpus)
+						break
+					}
+					runnerToExpire = s.findRunnerToUnload(pending)
+				}
+
+				if runnerToExpire == nil {
+					// Shouildn't happen
+					slog.Error("runner to expire was nil!")
+					continue
+				}
+				// Trigger an expiration to unload once it's done
+				runnerToExpire.refMu.Lock()
+				slog.Debug("resetting model to expire immediately to make room", "model", runnerToExpire.model, "refCount", runnerToExpire.refCount)
+				if runnerToExpire.expireTimer != nil {
+					runnerToExpire.expireTimer.Stop()
+					runnerToExpire.expireTimer = nil
+				}
+				runnerToExpire.sessionDuration = 0
+				if runnerToExpire.refCount <= 0 {
+					s.expiredCh <- runnerToExpire
+				}
+				runnerToExpire.refMu.Unlock()
+				// Wait for the unload to happen
+				// Note: at this point we're queueing up all incoming requests, even if they were for
+				// a different model that's loaded and not scheduled to be removed.
+				slog.Debug("waiting for pending requests to complete and unload to occur", "model", runnerToExpire.model)
+				select {
+				case <-ctx.Done():
+					slog.Debug("shutting down scheduler pending loop")
+					return
+				case <-s.unloadedCh:
+					slog.Debug("unload completed", "model", runnerToExpire.model)
+					continue
+				}
+			}
+		case <-s.unloadedCh:
+			// An unload request when there are no pending request can be ignored
+			slog.Debug("ignoring unload event with no pending requests")
+		}
+	}
+}
+
+func (s *Scheduler) processCompleted(ctx context.Context) {
+	// Process completed requests, expired timers, and unloading models
+	for {
+		select {
+		case <-ctx.Done():
+			slog.Debug("shutting down scheduler completed loop")
+			return
+		case finished := <-s.finishedReqCh:
+			s.loadedMu.Lock()
+			runner := s.loaded[finished.model.ModelPath]
+			s.loadedMu.Unlock()
+			if runner == nil {
+				slog.Error("finished requeset signal received after model unloaded", "model", finished.model.ModelPath)
+				continue
+			}
+			runner.refMu.Lock()
+			runner.refCount--
+			if runner.refCount <= 0 {
+				if runner.sessionDuration <= 0 {
+					slog.Debug("runner with zero duration has gone idle, expiring to unload", "model", runner.model)
+					if runner.expireTimer != nil {
+						runner.expireTimer.Stop()
+						runner.expireTimer = nil
+					}
+					s.expiredCh <- runner
+				} else if runner.expireTimer == nil {
+					slog.Debug("runner with non-zero duration has gone idle, adding timer", "model", runner.model, "duration", runner.sessionDuration)
+					runner.expireTimer = time.AfterFunc(runner.sessionDuration, func() {
+						slog.Debug("timer expired, expiring to unload", "model", runner.model)
+						runner.refMu.Lock()
+						defer runner.refMu.Unlock()
+						if runner.expireTimer != nil {
+							runner.expireTimer.Stop()
+							runner.expireTimer = nil
+						}
+						s.expiredCh <- runner
+					})
+				} else {
+					slog.Debug("runner with non-zero duration has gone idle, resetting timer", "model", runner.model, "duration", runner.sessionDuration)
+					runner.expireTimer.Reset(runner.sessionDuration)
+				}
+			}
+			slog.Debug("after processing request finished event", "model", runner.model, "refCount", runner.refCount)
+			runner.refMu.Unlock()
+		case runner := <-s.expiredCh:
+			slog.Debug("runner expired event received", "model", runner.model)
+			runner.refMu.Lock()
+			if runner.refCount > 0 {
+				// Shouldn't happen, but safeguard to ensure no leaked runners
+				slog.Debug("expired event with positive ref count, retrying", "model", runner.model, "refCount", runner.refCount)
+				go func(runner *runnerRef) {
+					// We can't unload yet, but want to as soon as the current request completes
+					// So queue up another expired event
+					time.Sleep(10 * time.Millisecond)
+					s.expiredCh <- runner
+				}(runner)
+				runner.refMu.Unlock()
+				continue
+			}
+
+			slog.Debug("got lock to unload", "model", runner.model)
+			runner.unload()
+			s.loadedMu.Lock()
+			delete(s.loaded, runner.model)
+			s.loadedMu.Unlock()
+			slog.Debug("runner released", "model", runner.model)
+			runner.refMu.Unlock()
+			slog.Debug("sending an unloaded event", "model", runner.model)
+			s.unloadedCh <- struct{}{}
+		}
+	}
+}
+
+// Complete the pending request and send the runner back to the requester
+// Wires up a finished event after the request context is completed
+// Updates session duration, and resets expiration timer
+func (pending *LlmRequest) useLoadedRunner(runner *runnerRef, finished chan *LlmRequest) {
+	runner.refMu.Lock()
+	defer runner.refMu.Unlock()
+	runner.refCount++
+	if runner.expireTimer != nil {
+		runner.expireTimer.Stop()
+		runner.expireTimer = nil
+	}
+	runner.sessionDuration = pending.sessionDuration
+	pending.successCh <- runner
+	go func() {
+		<-pending.ctx.Done()
+		slog.Debug("context for request finished")
+		finished <- pending
+	}()
+}
+
+func (s *Scheduler) load(req *LlmRequest, ggml *llm.GGML, gpus gpu.GpuInfoList) {
+	llama, err := s.newServerFn(gpus, req.model.ModelPath, ggml, req.model.AdapterPaths, req.model.ProjectorPaths, req.opts)
+	if err != nil {
+		// some older models are not compatible with newer versions of llama.cpp
+		// show a generalized compatibility error until there is a better way to
+		// check for model compatibility
+		if errors.Is(llm.ErrUnsupportedFormat, err) || strings.Contains(err.Error(), "failed to load model") {
+			err = fmt.Errorf("%v: this model may be incompatible with your version of Ollama. If you previously pulled this model, try updating it by running `ollama pull %s`", err, req.model.ShortName)
+		}
+		slog.Info("NewLlamaServer failed", "model", req.model.ModelPath, "error", err)
+		req.errCh <- err
+		return
+	}
+	runner := &runnerRef{}
+	runner.model = req.model.ModelPath
+	runner.adapters = req.model.AdapterPaths
+	runner.projectors = req.model.ProjectorPaths
+	runner.llama = llama
+	runner.Options = &req.opts
+	runner.sessionDuration = req.sessionDuration
+	runner.gpus = gpus
+	runner.estimatedVRAM = llama.EstimatedVRAM()
+	runner.loading = true
+	runner.refCount = 1
+	runner.refMu.Lock()
+	s.loadedMu.Lock()
+	s.loaded[req.model.ModelPath] = runner
+	slog.Info("loaded runners", "count", len(s.loaded))
+	s.loadedMu.Unlock()
+
+	go func() {
+		defer runner.refMu.Unlock()
+		if err = llama.WaitUntilRunning(req.ctx); err != nil {
+			slog.Error("error loading llama server", "error", err)
+			runner.refCount--
+			req.errCh <- err
+			slog.Debug("triggering expiration for failed load", "model", runner.model)
+			s.expiredCh <- runner
+			return
+		}
+		slog.Debug("finished setting up runner", "model", req.model.ModelPath)
+		runner.loading = false
+		go func() {
+			<-req.ctx.Done()
+			slog.Debug("context for request finished")
+			s.finishedReqCh <- req
+		}()
+		req.successCh <- runner
+	}()
+}
+
+func (s *Scheduler) updateFreeSpace(allGpus gpu.GpuInfoList) {
+	type predKey struct {
+		Library string
+		ID      string
+	}
+	predMap := map[predKey]uint64{} // Sum up the total predicted usage per GPU for all runners
+	s.loadedMu.Lock()
+	for _, r := range s.loaded {
+		r.refMu.Lock()
+		gpuIDs := make([]string, 0, len(r.gpus))
+		if r.llama != nil {
+
+			// TODO this should be broken down by GPU instead of assuming uniform spread
+			estimatedVRAMPerGPU := r.llama.EstimatedVRAM() / uint64(len(r.gpus))
+			for _, gpu := range r.gpus {
+				gpuIDs = append(gpuIDs, gpu.ID)
+			}
+			for _, gpu := range allGpus {
+				if slices.Contains(gpuIDs, gpu.ID) {
+					predMap[predKey{gpu.Library, gpu.ID}] += estimatedVRAMPerGPU
+				}
+			}
+		} else {
+			slog.Warn("unexpected nil runner reference, memory prediction may be incorrect")
+		}
+		r.refMu.Unlock()
+	}
+	s.loadedMu.Unlock()
+
+	// Now that we've summed up all the GPU usage predictions across all the loaded runners, update the gpu list
+	for i := range allGpus {
+		if p, ok := predMap[predKey{allGpus[i].Library, allGpus[i].ID}]; ok {
+			slog.Debug("gpu reported", "gpu", allGpus[i].ID, "library", allGpus[i].Library, "available", format.HumanBytes2(allGpus[i].FreeMemory))
+			if p > allGpus[i].TotalMemory {
+				// Shouldn't happen
+				slog.Warn("predicted usage exceeds VRAM", "gpu", allGpus[i].ID, "totalMemory", allGpus[i].TotalMemory, "predicted", p)
+				allGpus[i].FreeMemory = 0
+			} else if (allGpus[i].TotalMemory - p) < allGpus[i].FreeMemory { // predicted free is smaller than reported free, use it
+				// TODO maybe we should just always trust our numbers, since cuda's free memory reporting is laggy
+				// and we might unload models we didn't actually need to.  The risk is if some other GPU intensive app is loaded
+				// after we start our first runner, then we'll never acount for that, so picking the smallest free value seems prudent.
+				allGpus[i].FreeMemory = allGpus[i].TotalMemory - p
+			}
+			slog.Info("updated VRAM", "gpu", allGpus[i].ID, "library", allGpus[i].Library, "total", format.HumanBytes2(allGpus[i].TotalMemory), "available", format.HumanBytes2(allGpus[i].FreeMemory))
+		}
+	}
+}
+
+type runnerRef struct {
+	refMu sync.Mutex
+	// refCond   sync.Cond // Signaled on transition from 1 -> 0 refCount
+	refCount uint // prevent unloading if > 0
+	// unloading bool      // set to true when we are trying to unload the runner
+
+	llama         llm.LlamaServer
+	loading       bool            // True only during initial load, then false forever
+	gpus          gpu.GpuInfoList // Recorded at time of provisioning
+	estimatedVRAM uint64
+
+	sessionDuration time.Duration
+	expireTimer     *time.Timer
+
+	model      string
+	adapters   []string
+	projectors []string
+	*api.Options
+}
+
+// The refMu must already be held when calling unload
+func (runner *runnerRef) unload() {
+	if runner.expireTimer != nil {
+		runner.expireTimer.Stop()
+		runner.expireTimer = nil
+	}
+	if runner.llama != nil {
+		runner.llama.Close()
+	}
+	runner.llama = nil
+	runner.adapters = nil
+	runner.projectors = nil
+	runner.Options = nil
+	runner.gpus = nil
+}
+
+func (runner *runnerRef) needsReload(ctx context.Context, req *LlmRequest) bool {
+	slog.Debug("evaluating already loaded", "model", req.model.ModelPath)
+	runner.refMu.Lock()
+	defer runner.refMu.Unlock()
+
+	timeout := 10 * time.Second
+	if runner.loading {
+		timeout = 2 * time.Minute // Initial load can take a long time for big models on slow systems...
+	}
+
+	// Don't reload runner if num_gpu=-1 was provided
+	optsExisting := runner.Options.Runner
+	optsNew := req.opts.Runner
+	if optsNew.NumGPU < 0 {
+		optsExisting.NumGPU = -1
+		optsNew.NumGPU = -1
+	}
+
+	ctx, cancel := context.WithTimeout(ctx, timeout)
+	defer cancel()
+	if !reflect.DeepEqual(runner.adapters, req.model.AdapterPaths) || // have the adapters changed?
+		!reflect.DeepEqual(runner.projectors, req.model.ProjectorPaths) || // have the projectors changed?
+		!reflect.DeepEqual(optsExisting, optsNew) || // have the runner options changed?
+		runner.llama.Ping(ctx) != nil {
+		return true
+	}
+
+	return false
+}
+
+type ByDuration []*runnerRef
+
+func (a ByDuration) Len() int      { return len(a) }
+func (a ByDuration) Swap(i, j int) { a[i], a[j] = a[j], a[i] }
+func (a ByDuration) Less(i, j int) bool {
+	// uint64 to turn negative time (never unload) to largest
+	return uint64(a[i].sessionDuration) < uint64(a[j].sessionDuration)
+}
+
+// TODO - future consideration to pick runners based on size
+// type BySize []*runnerRef
+// func (a BySize) Len() int           { return len(a) }
+// func (a BySize) Swap(i, j int)      { a[i], a[j] = a[j], a[i] }
+// func (a BySize) Less(i, j int) bool { return a[i].estimatedVRAM < a[j].estimatedVRAM }
+
+// pickBestFitGPUs will try to find the optimal placement of the model in the available GPUs where the model fully fits
+// If the model can not be fit fully within the available GPU(s) nil is returned
+func pickBestFitGPUs(req *LlmRequest, ggml *llm.GGML, gpus gpu.GpuInfoList) gpu.GpuInfoList {
+	var estimatedVRAM uint64
+	for _, gl := range gpus.ByLibrary() {
+		var ok bool
+		sgl := append(make(gpu.GpuInfoList, 0, len(gl)), gl...)
+
+		// TODO - potentially sort by performance capability, existing models loaded, etc.
+		// Note: at present, this will favor more VRAM over faster GPU speed in mixed setups
+		sort.Sort(sort.Reverse(gpu.ByFreeMemory(sgl)))
+
+		// First attempt to fit the model into a single GPU
+		for _, g := range sgl {
+			if ok, estimatedVRAM = llm.PredictServerFit([]gpu.GpuInfo{g}, ggml, req.model.AdapterPaths, req.model.ProjectorPaths, req.opts); ok {
+				slog.Debug("new model will fit in available VRAM in single GPU, loading", "model", req.model.ModelPath, "gpu", g.ID, "available", g.FreeMemory, "required", format.HumanBytes2(estimatedVRAM))
+				return []gpu.GpuInfo{g}
+			}
+		}
+
+		// TODO future refinements
+		// - if multiple Libraries, see if any single GPU in any Library will fit
+		// - try subsets of GPUs instead of just falling back to 1 or all in a family
+
+		// Now try all the GPUs
+		if ok, estimatedVRAM = llm.PredictServerFit(gl, ggml, req.model.AdapterPaths, req.model.ProjectorPaths, req.opts); ok {
+			slog.Debug("new model will fit in available VRAM, loading", "model", req.model.ModelPath, "library", gl[0].Library, "required", format.HumanBytes2(estimatedVRAM))
+			return gl
+		}
+	}
+	return nil
+}
+
+// findRunnerToUnload finds a runner to unload to make room for a new model
+func (s *Scheduler) findRunnerToUnload(req *LlmRequest) *runnerRef {
+	s.loadedMu.Lock()
+	runnerList := make([]*runnerRef, 0, len(s.loaded))
+	for _, r := range s.loaded {
+		runnerList = append(runnerList, r)
+	}
+	s.loadedMu.Unlock()
+
+	// In the future we can enhance the algorithm to be smarter about picking the optimal runner to unload
+	// e.g., if we have multiple options, will one make room for the request?
+	sort.Sort(ByDuration(runnerList))
+
+	// First try to find a runner that's already idle
+	for _, runner := range runnerList {
+		runner.refMu.Lock()
+		rc := runner.refCount
+		runner.refMu.Unlock()
+		if rc == 0 {
+			slog.Debug("found an idle runner to unload")
+			return runner
+		}
+	}
+	// None appear idle, just wait for the one with the shortest duration
+	slog.Debug("no idle runners, picking the shortest duration", "count", len(runnerList))
+	return runnerList[0]
+}
+
+func (s *Scheduler) unloadAllRunners() {
+	s.loadedMu.Lock()
+	defer s.loadedMu.Unlock()
+	for model, runner := range s.loaded {
+		if runner.llama != nil {
+			slog.Debug("shutting down runner", "model", model)
+			runner.llama.Close()
+		}
+	}
+}

server/sched_test.go

@@ -0,0 +1,629 @@
+package server
+
+import (
+	"bytes"
+	"context"
+	"encoding/binary"
+	"fmt"
+	"log/slog"
+	"os"
+	"testing"
+	"time"
+
+	"github.com/ollama/ollama/api"
+	"github.com/ollama/ollama/app/lifecycle"
+	"github.com/ollama/ollama/format"
+	"github.com/ollama/ollama/gpu"
+	"github.com/ollama/ollama/llm"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func init() {
+	os.Setenv("OLLAMA_DEBUG", "1")
+	lifecycle.InitLogging()
+}
+
+func TestInitScheduler(t *testing.T) {
+	ctx, done := context.WithCancel(context.Background())
+	defer done()
+	initialMax := loadedMax
+	initialParallel := numParallel
+	s := InitScheduler(ctx)
+	require.Equal(t, initialMax, loadedMax)
+	s.loadedMu.Lock()
+	require.NotNil(t, s.loaded)
+	s.loadedMu.Unlock()
+
+	os.Setenv("OLLAMA_MAX_LOADED_MODELS", "blue")
+	s = InitScheduler(ctx)
+	require.Equal(t, initialMax, loadedMax)
+	s.loadedMu.Lock()
+	require.NotNil(t, s.loaded)
+	s.loadedMu.Unlock()
+
+	os.Setenv("OLLAMA_MAX_LOADED_MODELS", "0")
+	s = InitScheduler(ctx)
+	require.Equal(t, 0, loadedMax)
+	s.loadedMu.Lock()
+	require.NotNil(t, s.loaded)
+	s.loadedMu.Unlock()
+
+	os.Setenv("OLLAMA_NUM_PARALLEL", "blue")
+	_ = InitScheduler(ctx)
+	require.Equal(t, initialParallel, numParallel)
+	os.Setenv("OLLAMA_NUM_PARALLEL", "10")
+	_ = InitScheduler(ctx)
+	require.Equal(t, 10, numParallel)
+}
+
+func TestLoad(t *testing.T) {
+	ctx, done := context.WithTimeout(context.Background(), 20*time.Millisecond)
+	defer done()
+	s := InitScheduler(ctx)
+	var ggml *llm.GGML // value not used in tests
+	req := &LlmRequest{
+		ctx:             ctx,
+		model:           &Model{ModelPath: "foo"},
+		opts:            api.DefaultOptions(),
+		successCh:       make(chan *runnerRef, 1),
+		errCh:           make(chan error, 1),
+		sessionDuration: 2,
+	}
+	// Fail to load model first
+	s.newServerFn = func(gpus gpu.GpuInfoList, model string, ggml *llm.GGML, adapters []string, projectors []string, opts api.Options) (llm.LlamaServer, error) {
+		return nil, fmt.Errorf("something failed to load model blah")
+	}
+	gpus := gpu.GpuInfoList{}
+	s.load(req, ggml, gpus)
+	require.Len(t, req.successCh, 0)
+	require.Len(t, req.errCh, 1)
+	s.loadedMu.Lock()
+	require.Len(t, s.loaded, 0)
+	s.loadedMu.Unlock()
+	err := <-req.errCh
+	require.Contains(t, err.Error(), "this model may be incompatible")
+
+	server := &mockLlm{estimatedVRAM: 10}
+	s.newServerFn = func(gpus gpu.GpuInfoList, model string, ggml *llm.GGML, adapters []string, projectors []string, opts api.Options) (llm.LlamaServer, error) {
+		return server, nil
+	}
+	s.load(req, ggml, gpus)
+	select {
+	case err := <-req.errCh:
+		require.NoError(t, err)
+	case resp := <-req.successCh:
+		require.Equal(t, uint64(10), resp.estimatedVRAM)
+		require.Equal(t, uint(1), resp.refCount)
+		s.loadedMu.Lock()
+		require.Len(t, s.loaded, 1)
+		s.loadedMu.Unlock()
+	}
+
+	req.model.ModelPath = "dummy_model_path"
+	server.waitResp = fmt.Errorf("wait failure")
+	s.load(req, ggml, gpus)
+	select {
+	case err := <-req.errCh:
+		require.Contains(t, err.Error(), "wait failure")
+	case resp := <-req.successCh:
+		t.Errorf("unexpected success %v", resp)
+	}
+	s.loadedMu.Lock()
+	runner := s.loaded["dummy_model_path"]
+	s.loadedMu.Unlock()
+	require.NotNil(t, runner)
+	require.Equal(t, uint(0), runner.refCount)
+	time.Sleep(1 * time.Millisecond)
+	require.Len(t, s.expiredCh, 1)
+}
+
+type bundle struct {
+	ctx     context.Context //nolint:containedctx
+	ctxDone func()
+	srv     *mockLlm
+	req     *LlmRequest
+	ggml    *llm.GGML
+}
+
+func (scenario *bundle) newServer(gpus gpu.GpuInfoList, model string, ggml *llm.GGML, adapters []string, projectors []string, opts api.Options) (llm.LlamaServer, error) {
+	return scenario.srv, nil
+}
+
+func newScenario(t *testing.T, ctx context.Context, modelName string, estimatedVRAM uint64) *bundle {
+	scenario := &bundle{}
+	scenario.ctx, scenario.ctxDone = context.WithCancel(ctx)
+	t.Helper()
+
+	f, err := os.CreateTemp(t.TempDir(), modelName)
+	assert.Nil(t, err)
+	defer f.Close()
+
+	gguf := llm.NewGGUFV3(binary.LittleEndian)
+	err = gguf.Encode(f, llm.KV{
+		"general.architecture":          "llama",
+		"general.name":                  "name",
+		"llama.context_length":          uint32(32),
+		"llama.embedding_length":        uint32(4096),
+		"llama.block_count":             uint32(1),
+		"llama.attention.head_count":    uint32(32),
+		"llama.attention.head_count_kv": uint32(32),
+		"tokenizer.ggml.tokens":         []string{" "},
+		"tokenizer.ggml.scores":         []float32{0},
+		"tokenizer.ggml.token_type":     []int32{0},
+	}, []llm.Tensor{
+		{Name: "blk.0.attn.weight", Kind: uint32(0), Offset: uint64(0), Shape: []uint64{1, 1, 1, 1}, WriterTo: &bytes.Reader{}},
+	})
+	assert.Nil(t, err)
+
+	fname := f.Name()
+	model := &Model{Name: modelName, ModelPath: fname}
+	scenario.ggml, err = llm.LoadModel(model.ModelPath)
+	require.NoError(t, err)
+
+	scenario.req = &LlmRequest{
+		ctx:             scenario.ctx,
+		model:           model,
+		opts:            api.DefaultOptions(),
+		sessionDuration: 5 * time.Millisecond,
+		successCh:       make(chan *runnerRef, 1),
+		errCh:           make(chan error, 1),
+	}
+	scenario.srv = &mockLlm{estimatedVRAM: estimatedVRAM}
+	return scenario
+}
+
+func TestRequests(t *testing.T) {
+	ctx, done := context.WithTimeout(context.Background(), 500*time.Millisecond)
+	defer done()
+
+	// Same model, same request
+	scenario1a := newScenario(t, ctx, "ollama-model-1", 10)
+	scenario1a.req.sessionDuration = 0
+	scenario1b := newScenario(t, ctx, "ollama-model-1", 11)
+	scenario1b.req.model = scenario1a.req.model
+	scenario1b.ggml = scenario1a.ggml
+	scenario1b.req.sessionDuration = 0
+
+	// simple reload of same model
+	scenario2a := newScenario(t, ctx, "ollama-model-1", 20)
+	scenario2a.req.model = scenario1a.req.model
+	scenario2a.ggml = scenario1a.ggml
+
+	// Multiple loaded models
+	scenario3a := newScenario(t, ctx, "ollama-model-3a", 1*format.GigaByte)
+	scenario3b := newScenario(t, ctx, "ollama-model-3b", 24*format.GigaByte)
+	scenario3c := newScenario(t, ctx, "ollama-model-4a", 30)
+	scenario3c.req.opts.NumGPU = 0                           // CPU load, will be allowed
+	scenario3d := newScenario(t, ctx, "ollama-model-3c", 30) // Needs prior unloaded
+
+	s := InitScheduler(ctx)
+	s.getGpuFn = func() gpu.GpuInfoList {
+		g := gpu.GpuInfo{Library: "metal"}
+		g.TotalMemory = 24 * format.GigaByte
+		g.FreeMemory = 12 * format.GigaByte
+		return []gpu.GpuInfo{g}
+	}
+	s.newServerFn = scenario1a.newServer
+	slog.Info("scenario1a")
+	s.pendingReqCh <- scenario1a.req
+	require.Len(t, s.pendingReqCh, 1)
+	s.Run(ctx)
+	select {
+	case resp := <-scenario1a.req.successCh:
+		require.Equal(t, resp.llama, scenario1a.srv)
+		require.Len(t, s.pendingReqCh, 0)
+		require.Len(t, scenario1a.req.errCh, 0)
+	case <-ctx.Done():
+		t.Errorf("timeout")
+	}
+
+	// Same runner as first request due to not needing a reload
+	s.newServerFn = scenario1b.newServer
+	slog.Info("scenario1b")
+	s.pendingReqCh <- scenario1b.req
+	select {
+	case resp := <-scenario1b.req.successCh:
+		require.Equal(t, resp.llama, scenario1a.srv)
+		require.Len(t, s.pendingReqCh, 0)
+		require.Len(t, scenario1b.req.errCh, 0)
+	case <-ctx.Done():
+		t.Errorf("timeout")
+	}
+
+	// Trigger a reload
+	s.newServerFn = scenario2a.newServer
+	scenario2a.req.model.AdapterPaths = []string{"new"}
+	slog.Info("scenario2a")
+	s.pendingReqCh <- scenario2a.req
+	// finish first two requests, so model can reload
+	time.Sleep(1 * time.Millisecond)
+	scenario1a.ctxDone()
+	scenario1b.ctxDone()
+	select {
+	case resp := <-scenario2a.req.successCh:
+		require.Equal(t, resp.llama, scenario2a.srv)
+		require.Len(t, s.pendingReqCh, 0)
+		require.Len(t, scenario2a.req.errCh, 0)
+	case <-ctx.Done():
+		t.Errorf("timeout")
+	}
+
+	loadedMax = 1
+	s.newServerFn = scenario3a.newServer
+	slog.Info("scenario3a")
+	s.pendingReqCh <- scenario3a.req
+	// finish prior request, so new model can load
+	time.Sleep(1 * time.Millisecond)
+	scenario2a.ctxDone()
+	select {
+	case resp := <-scenario3a.req.successCh:
+		require.Equal(t, resp.llama, scenario3a.srv)
+		require.Len(t, s.pendingReqCh, 0)
+		require.Len(t, scenario3a.req.errCh, 0)
+	case <-ctx.Done():
+		t.Errorf("timeout")
+	}
+	s.loadedMu.Lock()
+	require.Len(t, s.loaded, 1)
+	s.loadedMu.Unlock()
+
+	loadedMax = 0
+	s.newServerFn = scenario3b.newServer
+	slog.Info("scenario3b")
+	s.pendingReqCh <- scenario3b.req
+	select {
+	case resp := <-scenario3b.req.successCh:
+		require.Equal(t, resp.llama, scenario3b.srv)
+		require.Len(t, s.pendingReqCh, 0)
+		require.Len(t, scenario3b.req.errCh, 0)
+	case <-ctx.Done():
+		t.Errorf("timeout")
+	}
+	s.loadedMu.Lock()
+	require.Len(t, s.loaded, 2)
+	s.loadedMu.Unlock()
+
+	// This is a CPU load with NumGPU = 0 so it should load
+	s.newServerFn = scenario3c.newServer
+	slog.Info("scenario3c")
+	s.pendingReqCh <- scenario3c.req
+	select {
+	case resp := <-scenario3c.req.successCh:
+		require.Equal(t, resp.llama, scenario3c.srv)
+		require.Len(t, s.pendingReqCh, 0)
+		require.Len(t, scenario3c.req.errCh, 0)
+	case <-ctx.Done():
+		t.Errorf("timeout")
+	}
+	s.loadedMu.Lock()
+	require.Len(t, s.loaded, 3)
+	s.loadedMu.Unlock()
+
+	// Try to load a model that wont fit
+	s.newServerFn = scenario3d.newServer
+	slog.Info("scenario3d")
+	s.loadedMu.Lock()
+	require.Len(t, s.loaded, 3)
+	s.loadedMu.Unlock()
+	scenario3a.ctxDone() // Won't help since this one isn't big enough to make room
+	time.Sleep(2 * time.Millisecond)
+	s.pendingReqCh <- scenario3d.req
+	// finish prior request, so new model can load
+	time.Sleep(6 * time.Millisecond)
+	s.loadedMu.Lock()
+	require.Len(t, s.loaded, 2)
+	s.loadedMu.Unlock()
+	scenario3b.ctxDone()
+	select {
+	case resp := <-scenario3d.req.successCh:
+		require.Equal(t, resp.llama, scenario3d.srv)
+		require.Len(t, s.pendingReqCh, 0)
+		require.Len(t, scenario3d.req.errCh, 0)
+	case <-ctx.Done():
+		t.Errorf("timeout")
+	}
+	s.loadedMu.Lock()
+	require.Len(t, s.loaded, 2)
+	s.loadedMu.Unlock()
+}
+
+func TestGetRunner(t *testing.T) {
+	ctx, done := context.WithTimeout(context.Background(), 100*time.Millisecond)
+	defer done()
+
+	// Same model, same request
+	scenario1a := newScenario(t, ctx, "ollama-model-1a", 10)
+	scenario1a.req.sessionDuration = 0
+	scenario1b := newScenario(t, ctx, "ollama-model-1b", 10)
+	scenario1b.req.sessionDuration = 0
+	scenario1c := newScenario(t, ctx, "ollama-model-1c", 10)
+	scenario1c.req.sessionDuration = 0
+	maxQueuedRequests = 1
+	s := InitScheduler(ctx)
+	s.getGpuFn = func() gpu.GpuInfoList {
+		g := gpu.GpuInfo{Library: "metal"}
+		g.TotalMemory = 24 * format.GigaByte
+		g.FreeMemory = 12 * format.GigaByte
+		return []gpu.GpuInfo{g}
+	}
+	s.newServerFn = scenario1a.newServer
+	slog.Info("scenario1a")
+	successCh1a, errCh1a := s.GetRunner(scenario1a.ctx, scenario1a.req.model, scenario1a.req.opts, scenario1a.req.sessionDuration)
+	require.Len(t, s.pendingReqCh, 1)
+	slog.Info("scenario1b")
+	successCh1b, errCh1b := s.GetRunner(scenario1b.ctx, scenario1b.req.model, scenario1b.req.opts, scenario1b.req.sessionDuration)
+	require.Len(t, s.pendingReqCh, 1)
+	require.Len(t, successCh1b, 0)
+	require.Len(t, errCh1b, 1)
+	err := <-errCh1b
+	require.Contains(t, err.Error(), "server busy")
+	s.Run(ctx)
+	select {
+	case resp := <-successCh1a:
+		require.Equal(t, resp.llama, scenario1a.srv)
+		require.Len(t, s.pendingReqCh, 0)
+		require.Len(t, errCh1a, 0)
+	case <-ctx.Done():
+		t.Errorf("timeout")
+	}
+	scenario1a.ctxDone()
+	s.loadedMu.Lock()
+	require.Len(t, s.loaded, 1)
+	s.loadedMu.Unlock()
+
+	scenario1c.req.model.ModelPath = "bad path"
+	slog.Info("scenario1c")
+	successCh1c, errCh1c := s.GetRunner(scenario1c.ctx, scenario1c.req.model, scenario1c.req.opts, scenario1c.req.sessionDuration)
+	require.Len(t, s.pendingReqCh, 0)
+	require.Len(t, successCh1c, 0)
+	require.Len(t, errCh1c, 0)
+
+	time.Sleep(5 * time.Millisecond)
+	s.loadedMu.Lock()
+	require.Len(t, s.loaded, 0)
+	s.loadedMu.Unlock()
+	require.Len(t, errCh1c, 1)
+	err = <-errCh1c
+	require.Contains(t, err.Error(), "bad path")
+	scenario1b.ctxDone()
+}
+
+// TODO - add one scenario that triggers the bogus finished event with positive ref count
+func TestPrematureExpired(t *testing.T) {
+	ctx, done := context.WithTimeout(context.Background(), 500*time.Millisecond)
+	defer done()
+
+	// Same model, same request
+	scenario1a := newScenario(t, ctx, "ollama-model-1a", 10)
+	s := InitScheduler(ctx)
+	s.getGpuFn = func() gpu.GpuInfoList {
+		g := gpu.GpuInfo{Library: "metal"}
+		g.TotalMemory = 24 * format.GigaByte
+		g.FreeMemory = 12 * format.GigaByte
+		return []gpu.GpuInfo{g}
+	}
+	s.newServerFn = scenario1a.newServer
+	successCh1a, errCh1a := s.GetRunner(scenario1a.ctx, scenario1a.req.model, scenario1a.req.opts, scenario1a.req.sessionDuration)
+	require.Len(t, s.pendingReqCh, 1)
+	s.Run(ctx)
+	select {
+	case resp := <-successCh1a:
+		require.Equal(t, resp.llama, scenario1a.srv)
+		require.Len(t, s.pendingReqCh, 0)
+		require.Len(t, errCh1a, 0)
+		s.loadedMu.Lock()
+		require.Len(t, s.loaded, 1)
+		s.loadedMu.Unlock()
+		slog.Info("sending premature expired event now")
+		s.expiredCh <- resp // Shouldn't happen in real life, but make sure its safe
+	case <-ctx.Done():
+		t.Errorf("timeout")
+	}
+	time.Sleep(scenario1a.req.sessionDuration)
+	scenario1a.ctxDone()
+	time.Sleep(20 * time.Millisecond)
+	require.LessOrEqual(t, len(s.finishedReqCh), 1)
+	time.Sleep(10 * time.Millisecond)
+	require.Len(t, s.finishedReqCh, 0)
+	s.loadedMu.Lock()
+	require.Len(t, s.loaded, 0)
+	s.loadedMu.Unlock()
+
+	// also shouldn't happen in real life
+	s.finishedReqCh <- scenario1a.req
+	time.Sleep(5 * time.Millisecond)
+}
+
+func TestUseLoadedRunner(t *testing.T) {
+	ctx, done := context.WithTimeout(context.Background(), 100*time.Millisecond)
+	req := &LlmRequest{
+		ctx:             ctx,
+		opts:            api.DefaultOptions(),
+		successCh:       make(chan *runnerRef, 1),
+		sessionDuration: 2,
+	}
+	finished := make(chan *LlmRequest)
+	llm1 := &mockLlm{}
+	r1 := &runnerRef{llama: llm1, sessionDuration: 1}
+	req.useLoadedRunner(r1, finished)
+	require.Equal(t, uint(1), r1.refCount)
+	require.Equal(t, time.Duration(2), r1.sessionDuration)
+	select {
+	case success := <-req.successCh:
+		require.Equal(t, r1, success)
+	case <-ctx.Done():
+		t.Errorf("timeout")
+	}
+	done()
+	fin := <-finished
+	require.Equal(t, req, fin)
+}
+
+func TestUpdateFreeSpace(t *testing.T) {
+	ctx, done := context.WithTimeout(context.Background(), 100*time.Millisecond)
+	defer done()
+	gpus := gpu.GpuInfoList{
+		{
+			Library: "a",
+			ID:      "1",
+		},
+		{
+			Library: "a",
+			ID:      "2",
+		},
+	}
+	gpus[0].TotalMemory = 1000
+	gpus[0].FreeMemory = 900
+	gpus[1].TotalMemory = 2000
+	gpus[1].FreeMemory = 1900
+	llm1 := &mockLlm{estimatedVRAM: 100}
+	llm2 := &mockLlm{estimatedVRAM: 200}
+	r1 := &runnerRef{llama: llm1, gpus: gpus}
+	r2 := &runnerRef{llama: llm2, gpus: gpus}
+
+	s := InitScheduler(ctx)
+	s.loadedMu.Lock()
+	s.loaded["a"] = r1
+	s.loaded["b"] = r2
+	s.loadedMu.Unlock()
+
+	s.updateFreeSpace(gpus)
+	require.Equal(t, uint64(850), gpus[0].FreeMemory)
+	require.Equal(t, uint64(1850), gpus[1].FreeMemory)
+}
+
+func TestFindRunnerToUnload(t *testing.T) {
+	ctx, done := context.WithTimeout(context.Background(), 100*time.Millisecond)
+	defer done()
+	req := &LlmRequest{
+		ctx:  ctx,
+		opts: api.DefaultOptions(),
+	}
+	r1 := &runnerRef{refCount: 1, sessionDuration: 1}
+	r2 := &runnerRef{sessionDuration: 2}
+
+	s := InitScheduler(ctx)
+	s.loadedMu.Lock()
+	s.loaded["a"] = r1
+	s.loaded["b"] = r2
+	s.loadedMu.Unlock()
+
+	resp := s.findRunnerToUnload(req)
+	require.Equal(t, r2, resp)
+	r2.refCount = 1
+	resp = s.findRunnerToUnload(req)
+	require.Equal(t, r1, resp)
+
+}
+
+func TestNeedsReload(t *testing.T) {
+	ctx, done := context.WithTimeout(context.Background(), 100*time.Millisecond)
+	defer done()
+
+	llm := &mockLlm{}
+	do := api.DefaultOptions()
+	runner := &runnerRef{
+		adapters:   []string{"adapter1"},
+		projectors: []string{"projector1"},
+		Options:    &do,
+		llama:      llm,
+	}
+	req := &LlmRequest{
+		model: &Model{
+			AdapterPaths:   []string{"adapter2"},
+			ProjectorPaths: []string{"projector2"},
+		},
+		opts: api.DefaultOptions(),
+	}
+	resp := runner.needsReload(ctx, req)
+	require.True(t, resp)
+	req.model.AdapterPaths = runner.adapters
+	resp = runner.needsReload(ctx, req)
+	require.True(t, resp)
+	req.model.ProjectorPaths = runner.projectors
+	runner.loading = true
+	req.opts.NumBatch = 1234
+	resp = runner.needsReload(ctx, req)
+	require.True(t, resp)
+	req.opts.NumBatch = runner.Options.NumBatch
+	llm.pingResp = fmt.Errorf("foo")
+	resp = runner.needsReload(ctx, req)
+	require.True(t, resp)
+	llm.pingResp = nil
+	resp = runner.needsReload(ctx, req)
+	require.False(t, resp)
+	req.opts.NumGPU = 99
+	resp = runner.needsReload(ctx, req)
+	require.True(t, resp)
+	req.opts.NumGPU = -1
+	resp = runner.needsReload(ctx, req)
+	require.False(t, resp)
+}
+
+func TestUnloadAllRunners(t *testing.T) {
+	ctx, done := context.WithTimeout(context.Background(), 100*time.Millisecond)
+	defer done()
+
+	llm1 := &mockLlm{}
+	llm2 := &mockLlm{}
+	s := InitScheduler(ctx)
+	s.unloadAllRunners()
+
+	r1 := &runnerRef{llama: llm1}
+	r2 := &runnerRef{llama: llm2}
+
+	s.loadedMu.Lock()
+	s.loaded["a"] = r1
+	s.loaded["b"] = r2
+	s.loadedMu.Unlock()
+	s.unloadAllRunners()
+
+	require.True(t, llm1.closeCalled)
+	require.True(t, llm2.closeCalled)
+}
+
+func TestUnload(t *testing.T) {
+	llm1 := &mockLlm{}
+	r1 := &runnerRef{llama: llm1}
+	r2 := &runnerRef{adapters: []string{"A"}}
+	r1.unload()
+	require.True(t, llm1.closeCalled)
+	r2.unload()
+	require.Nil(t, r2.adapters)
+}
+
+type mockLlm struct {
+	pingResp          error
+	waitResp          error
+	completionResp    error
+	embeddingResp     []float64
+	embeddingRespErr  error
+	tokenizeResp      []int
+	tokenizeRespErr   error
+	detokenizeResp    string
+	detonekizeRespErr error
+	closeResp         error
+	closeCalled       bool
+	estimatedVRAM     uint64
+}
+
+func (s *mockLlm) Ping(ctx context.Context) error             { return s.pingResp }
+func (s *mockLlm) WaitUntilRunning(ctx context.Context) error { return s.waitResp }
+func (s *mockLlm) Completion(ctx context.Context, req llm.CompletionRequest, fn func(llm.CompletionResponse)) error {
+	return s.completionResp
+}
+func (s *mockLlm) Embedding(ctx context.Context, prompt string) ([]float64, error) {
+	return s.embeddingResp, s.embeddingRespErr
+}
+func (s *mockLlm) Tokenize(ctx context.Context, content string) ([]int, error) {
+	return s.tokenizeResp, s.tokenizeRespErr
+}
+func (s *mockLlm) Detokenize(ctx context.Context, tokens []int) (string, error) {
+	return s.detokenizeResp, s.detonekizeRespErr
+}
+func (s *mockLlm) Close() error {
+	s.closeCalled = true
+	return s.closeResp
+}
+func (s *mockLlm) EstimatedVRAM() uint64 { return s.estimatedVRAM }

types/errtypes/errtypes.go

@@ -0,0 +1,18 @@
+// Package errtypes contains custom error types
+package errtypes
+
+import (
+	"fmt"
+	"strings"
+)
+
+const UnknownOllamaKeyErrMsg = "unknown ollama key"
+
+// TODO: This should have a structured response from the API
+type UnknownOllamaKey struct {
+	Key string
+}
+
+func (e *UnknownOllamaKey) Error() string {
+	return fmt.Sprintf("unauthorized: %s %q", UnknownOllamaKeyErrMsg, strings.TrimSpace(e.Key))
+}

types/model/digest.go

@@ -1,83 +0,0 @@
-package model
-
-import (
-	"fmt"
-	"log/slog"
-	"strings"
-	"unicode"
-)
-
-// Digest represents a digest of a model Manifest. It is a comparable value
-// type and is immutable.
-//
-// The zero Digest is not a valid digest.
-type Digest struct {
-	s string
-}
-
-// Type returns the digest type of the digest.
-//
-// Example:
-//
-//	ParseDigest("sha256-1234").Type() // returns "sha256"
-func (d Digest) Type() string {
-	typ, _, _ := strings.Cut(d.s, "-")
-	return typ
-}
-
-// String returns the digest in the form of "<digest-type>-<digest>", or the
-// empty string if the digest is invalid.
-func (d Digest) String() string { return d.s }
-
-// IsValid returns true if the digest is valid (not zero).
-//
-// A valid digest may be created only by ParseDigest, or
-// ParseName(name).Digest().
-func (d Digest) IsValid() bool { return d.s != "" }
-
-// LogValue implements slog.Value.
-func (d Digest) LogValue() slog.Value {
-	return slog.StringValue(d.String())
-}
-
-var (
-	_ slog.LogValuer = Digest{}
-)
-
-// ParseDigest parses a string in the form of "<digest-type>-<digest>" into a
-// Digest.
-func ParseDigest(s string) Digest {
-	typ, digest, ok := strings.Cut(s, "-")
-	if !ok {
-		typ, digest, ok = strings.Cut(s, ":")
-	}
-	if ok && isValidDigestType(typ) && isValidHex(digest) {
-		return Digest{s: fmt.Sprintf("%s-%s", typ, digest)}
-	}
-	return Digest{}
-}
-
-func isValidDigestType(s string) bool {
-	if len(s) == 0 {
-		return false
-	}
-	for _, r := range s {
-		if !unicode.IsLower(r) && !unicode.IsDigit(r) {
-			return false
-		}
-	}
-	return true
-}
-
-func isValidHex(s string) bool {
-	if len(s) == 0 {
-		return false
-	}
-	for i := range s {
-		c := s[i]
-		if c < '0' || c > '9' && c < 'a' || c > 'f' {
-			return false
-		}
-	}
-	return true
-}

types/model/digest_test.go

@@ -1,46 +0,0 @@
-package model
-
-import "testing"
-
-var testDigests = map[string]Digest{
-	"":                 {},
-	"sha256-1234":      {s: "sha256-1234"},
-	"sha256-5678":      {s: "sha256-5678"},
-	"blake2-9abc":      {s: "blake2-9abc"},
-	"-1234":            {},
-	"sha256-":          {},
-	"sha256-1234-5678": {},
-	"sha256-P":         {}, //         invalid  hex
-	"sha256-1234P":     {},
-	"---":              {},
-}
-
-func TestDigestParse(t *testing.T) {
-	// Test cases.
-	for s, want := range testDigests {
-		got := ParseDigest(s)
-		t.Logf("ParseDigest(%q) = %#v", s, got)
-		if got != want {
-			t.Errorf("ParseDigest(%q) = %q; want %q", s, got, want)
-		}
-	}
-}
-
-func TestDigestString(t *testing.T) {
-	// Test cases.
-	for s, d := range testDigests {
-		want := s
-		if !d.IsValid() {
-			want = ""
-		}
-		got := d.String()
-		if got != want {
-			t.Errorf("ParseDigest(%q).String() = %q; want %q", s, got, want)
-		}
-
-		got = ParseDigest(s).String()
-		if got != want {
-			t.Errorf("roundtrip ParseDigest(%q).String() = %q; want %q", s, got, want)
-		}
-	}
-}

types/model/file.go

@@ -0,0 +1,303 @@
+package model
+
+import (
+	"bufio"
+	"bytes"
+	"errors"
+	"fmt"
+	"io"
+	"strconv"
+	"strings"
+)
+
+type File struct {
+	Commands []Command
+}
+
+func (f File) String() string {
+	var sb strings.Builder
+	for _, cmd := range f.Commands {
+		fmt.Fprintln(&sb, cmd.String())
+	}
+
+	return sb.String()
+}
+
+type Command struct {
+	Name string
+	Args string
+}
+
+func (c Command) String() string {
+	var sb strings.Builder
+	switch c.Name {
+	case "model":
+		fmt.Fprintf(&sb, "FROM %s", c.Args)
+	case "license", "template", "system", "adapter":
+		fmt.Fprintf(&sb, "%s %s", strings.ToUpper(c.Name), quote(c.Args))
+	case "message":
+		role, message, _ := strings.Cut(c.Args, ": ")
+		fmt.Fprintf(&sb, "MESSAGE %s %s", role, quote(message))
+	default:
+		fmt.Fprintf(&sb, "PARAMETER %s %s", c.Name, quote(c.Args))
+	}
+
+	return sb.String()
+}
+
+type state int
+
+const (
+	stateNil state = iota
+	stateName
+	stateValue
+	stateParameter
+	stateMessage
+	stateComment
+)
+
+var (
+	errMissingFrom        = errors.New("no FROM line")
+	errInvalidMessageRole = errors.New("message role must be one of \"system\", \"user\", or \"assistant\"")
+	errInvalidCommand     = errors.New("command must be one of \"from\", \"license\", \"template\", \"system\", \"adapter\", \"parameter\", or \"message\"")
+)
+
+func ParseFile(r io.Reader) (*File, error) {
+	var cmd Command
+	var curr state
+	var b bytes.Buffer
+	var role string
+
+	var f File
+
+	br := bufio.NewReader(r)
+	for {
+		r, _, err := br.ReadRune()
+		if errors.Is(err, io.EOF) {
+			break
+		} else if err != nil {
+			return nil, err
+		}
+
+		next, r, err := parseRuneForState(r, curr)
+		if errors.Is(err, io.ErrUnexpectedEOF) {
+			return nil, fmt.Errorf("%w: %s", err, b.String())
+		} else if err != nil {
+			return nil, err
+		}
+
+		// process the state transition, some transitions need to be intercepted and redirected
+		if next != curr {
+			switch curr {
+			case stateName:
+				if !isValidCommand(b.String()) {
+					return nil, errInvalidCommand
+				}
+
+				// next state sometimes depends on the current buffer value
+				switch s := strings.ToLower(b.String()); s {
+				case "from":
+					cmd.Name = "model"
+				case "parameter":
+					// transition to stateParameter which sets command name
+					next = stateParameter
+				case "message":
+					// transition to stateMessage which validates the message role
+					next = stateMessage
+					fallthrough
+				default:
+					cmd.Name = s
+				}
+			case stateParameter:
+				cmd.Name = b.String()
+			case stateMessage:
+				if !isValidMessageRole(b.String()) {
+					return nil, errInvalidMessageRole
+				}
+
+				role = b.String()
+			case stateComment, stateNil:
+				// pass
+			case stateValue:
+				s, ok := unquote(b.String())
+				if !ok || isSpace(r) {
+					if _, err := b.WriteRune(r); err != nil {
+						return nil, err
+					}
+
+					continue
+				}
+
+				if role != "" {
+					s = role + ": " + s
+					role = ""
+				}
+
+				cmd.Args = s
+				f.Commands = append(f.Commands, cmd)
+			}
+
+			b.Reset()
+			curr = next
+		}
+
+		if strconv.IsPrint(r) {
+			if _, err := b.WriteRune(r); err != nil {
+				return nil, err
+			}
+		}
+	}
+
+	// flush the buffer
+	switch curr {
+	case stateComment, stateNil:
+		// pass; nothing to flush
+	case stateValue:
+		s, ok := unquote(b.String())
+		if !ok {
+			return nil, io.ErrUnexpectedEOF
+		}
+
+		if role != "" {
+			s = role + ": " + s
+		}
+
+		cmd.Args = s
+		f.Commands = append(f.Commands, cmd)
+	default:
+		return nil, io.ErrUnexpectedEOF
+	}
+
+	for _, cmd := range f.Commands {
+		if cmd.Name == "model" {
+			return &f, nil
+		}
+	}
+
+	return nil, errMissingFrom
+}
+
+func parseRuneForState(r rune, cs state) (state, rune, error) {
+	switch cs {
+	case stateNil:
+		switch {
+		case r == '#':
+			return stateComment, 0, nil
+		case isSpace(r), isNewline(r):
+			return stateNil, 0, nil
+		default:
+			return stateName, r, nil
+		}
+	case stateName:
+		switch {
+		case isAlpha(r):
+			return stateName, r, nil
+		case isSpace(r):
+			return stateValue, 0, nil
+		default:
+			return stateNil, 0, errInvalidCommand
+		}
+	case stateValue:
+		switch {
+		case isNewline(r):
+			return stateNil, r, nil
+		case isSpace(r):
+			return stateNil, r, nil
+		default:
+			return stateValue, r, nil
+		}
+	case stateParameter:
+		switch {
+		case isAlpha(r), isNumber(r), r == '_':
+			return stateParameter, r, nil
+		case isSpace(r):
+			return stateValue, 0, nil
+		default:
+			return stateNil, 0, io.ErrUnexpectedEOF
+		}
+	case stateMessage:
+		switch {
+		case isAlpha(r):
+			return stateMessage, r, nil
+		case isSpace(r):
+			return stateValue, 0, nil
+		default:
+			return stateNil, 0, io.ErrUnexpectedEOF
+		}
+	case stateComment:
+		switch {
+		case isNewline(r):
+			return stateNil, 0, nil
+		default:
+			return stateComment, 0, nil
+		}
+	default:
+		return stateNil, 0, errors.New("")
+	}
+}
+
+func quote(s string) string {
+	if strings.Contains(s, "\n") || strings.HasPrefix(s, " ") || strings.HasSuffix(s, " ") {
+		if strings.Contains(s, "\"") {
+			return `"""` + s + `"""`
+		}
+
+		return `"` + s + `"`
+	}
+
+	return s
+}
+
+func unquote(s string) (string, bool) {
+	if len(s) == 0 {
+		return "", false
+	}
+
+	// TODO: single quotes
+	if len(s) >= 3 && s[:3] == `"""` {
+		if len(s) >= 6 && s[len(s)-3:] == `"""` {
+			return s[3 : len(s)-3], true
+		}
+
+		return "", false
+	}
+
+	if len(s) >= 1 && s[0] == '"' {
+		if len(s) >= 2 && s[len(s)-1] == '"' {
+			return s[1 : len(s)-1], true
+		}
+
+		return "", false
+	}
+
+	return s, true
+}
+
+func isAlpha(r rune) bool {
+	return r >= 'a' && r <= 'z' || r >= 'A' && r <= 'Z'
+}
+
+func isNumber(r rune) bool {
+	return r >= '0' && r <= '9'
+}
+
+func isSpace(r rune) bool {
+	return r == ' ' || r == '\t'
+}
+
+func isNewline(r rune) bool {
+	return r == '\r' || r == '\n'
+}
+
+func isValidMessageRole(role string) bool {
+	return role == "system" || role == "user" || role == "assistant"
+}
+
+func isValidCommand(cmd string) bool {
+	switch strings.ToLower(cmd) {
+	case "from", "license", "template", "system", "adapter", "parameter", "message":
+		return true
+	default:
+		return false
+	}
+}

types/model/file_test.go

@@ -0,0 +1,507 @@
+package model
+
+import (
+	"bytes"
+	"fmt"
+	"io"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestParseFileFile(t *testing.T) {
+	input := `
+FROM model1
+ADAPTER adapter1
+LICENSE MIT
+PARAMETER param1 value1
+PARAMETER param2 value2
+TEMPLATE template1
+`
+
+	reader := strings.NewReader(input)
+
+	modelfile, err := ParseFile(reader)
+	assert.NoError(t, err)
+
+	expectedCommands := []Command{
+		{Name: "model", Args: "model1"},
+		{Name: "adapter", Args: "adapter1"},
+		{Name: "license", Args: "MIT"},
+		{Name: "param1", Args: "value1"},
+		{Name: "param2", Args: "value2"},
+		{Name: "template", Args: "template1"},
+	}
+
+	assert.Equal(t, expectedCommands, modelfile.Commands)
+}
+
+func TestParseFileFrom(t *testing.T) {
+	var cases = []struct {
+		input    string
+		expected []Command
+		err      error
+	}{
+		{
+			"FROM foo",
+			[]Command{{Name: "model", Args: "foo"}},
+			nil,
+		},
+		{
+			"FROM /path/to/model",
+			[]Command{{Name: "model", Args: "/path/to/model"}},
+			nil,
+		},
+		{
+			"FROM /path/to/model/fp16.bin",
+			[]Command{{Name: "model", Args: "/path/to/model/fp16.bin"}},
+			nil,
+		},
+		{
+			"FROM llama3:latest",
+			[]Command{{Name: "model", Args: "llama3:latest"}},
+			nil,
+		},
+		{
+			"FROM llama3:7b-instruct-q4_K_M",
+			[]Command{{Name: "model", Args: "llama3:7b-instruct-q4_K_M"}},
+			nil,
+		},
+		{
+			"", nil, errMissingFrom,
+		},
+		{
+			"PARAMETER param1 value1",
+			nil,
+			errMissingFrom,
+		},
+		{
+			"PARAMETER param1 value1\nFROM foo",
+			[]Command{{Name: "param1", Args: "value1"}, {Name: "model", Args: "foo"}},
+			nil,
+		},
+	}
+
+	for _, c := range cases {
+		t.Run("", func(t *testing.T) {
+			modelfile, err := ParseFile(strings.NewReader(c.input))
+			assert.ErrorIs(t, err, c.err)
+			if modelfile != nil {
+				assert.Equal(t, c.expected, modelfile.Commands)
+			}
+		})
+	}
+}
+
+func TestParseFileParametersMissingValue(t *testing.T) {
+	input := `
+FROM foo
+PARAMETER param1
+`
+
+	reader := strings.NewReader(input)
+
+	_, err := ParseFile(reader)
+	assert.ErrorIs(t, err, io.ErrUnexpectedEOF)
+}
+
+func TestParseFileBadCommand(t *testing.T) {
+	input := `
+FROM foo
+BADCOMMAND param1 value1
+`
+	_, err := ParseFile(strings.NewReader(input))
+	assert.ErrorIs(t, err, errInvalidCommand)
+
+}
+
+func TestParseFileMessages(t *testing.T) {
+	var cases = []struct {
+		input    string
+		expected []Command
+		err      error
+	}{
+		{
+			`
+FROM foo
+MESSAGE system You are a file parser. Always parse things.
+`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "message", Args: "system: You are a file parser. Always parse things."},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+MESSAGE system You are a file parser. Always parse things.`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "message", Args: "system: You are a file parser. Always parse things."},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+MESSAGE system You are a file parser. Always parse things.
+MESSAGE user Hey there!
+MESSAGE assistant Hello, I want to parse all the things!
+`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "message", Args: "system: You are a file parser. Always parse things."},
+				{Name: "message", Args: "user: Hey there!"},
+				{Name: "message", Args: "assistant: Hello, I want to parse all the things!"},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+MESSAGE system """
+You are a multiline file parser. Always parse things.
+"""
+			`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "message", Args: "system: \nYou are a multiline file parser. Always parse things.\n"},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+MESSAGE badguy I'm a bad guy!
+`,
+			nil,
+			errInvalidMessageRole,
+		},
+		{
+			`
+FROM foo
+MESSAGE system
+`,
+			nil,
+			io.ErrUnexpectedEOF,
+		},
+		{
+			`
+FROM foo
+MESSAGE system`,
+			nil,
+			io.ErrUnexpectedEOF,
+		},
+	}
+
+	for _, c := range cases {
+		t.Run("", func(t *testing.T) {
+			modelfile, err := ParseFile(strings.NewReader(c.input))
+			assert.ErrorIs(t, err, c.err)
+			if modelfile != nil {
+				assert.Equal(t, c.expected, modelfile.Commands)
+			}
+		})
+	}
+}
+
+func TestParseFileQuoted(t *testing.T) {
+	var cases = []struct {
+		multiline string
+		expected  []Command
+		err       error
+	}{
+		{
+			`
+FROM foo
+SYSTEM """
+This is a
+multiline system.
+"""
+			`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "system", Args: "\nThis is a\nmultiline system.\n"},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+SYSTEM """
+This is a
+multiline system."""
+			`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "system", Args: "\nThis is a\nmultiline system."},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+SYSTEM """This is a
+multiline system."""
+			`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "system", Args: "This is a\nmultiline system."},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+SYSTEM """This is a multiline system."""
+			`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "system", Args: "This is a multiline system."},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+SYSTEM """This is a multiline system.""
+			`,
+			nil,
+			io.ErrUnexpectedEOF,
+		},
+		{
+			`
+FROM foo
+SYSTEM "
+			`,
+			nil,
+			io.ErrUnexpectedEOF,
+		},
+		{
+			`
+FROM foo
+SYSTEM """
+This is a multiline system with "quotes".
+"""
+`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "system", Args: "\nThis is a multiline system with \"quotes\".\n"},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+SYSTEM """"""
+`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "system", Args: ""},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+SYSTEM ""
+`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "system", Args: ""},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+SYSTEM "'"
+`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "system", Args: "'"},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+SYSTEM """''"'""'""'"'''''""'""'"""
+`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "system", Args: `''"'""'""'"'''''""'""'`},
+			},
+			nil,
+		},
+		{
+			`
+FROM foo
+TEMPLATE """
+{{ .Prompt }}
+"""`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+				{Name: "template", Args: "\n{{ .Prompt }}\n"},
+			},
+			nil,
+		},
+	}
+
+	for _, c := range cases {
+		t.Run("", func(t *testing.T) {
+			modelfile, err := ParseFile(strings.NewReader(c.multiline))
+			assert.ErrorIs(t, err, c.err)
+			if modelfile != nil {
+				assert.Equal(t, c.expected, modelfile.Commands)
+			}
+		})
+	}
+}
+
+func TestParseFileParameters(t *testing.T) {
+	var cases = map[string]struct {
+		name, value string
+	}{
+		"numa true":                    {"numa", "true"},
+		"num_ctx 1":                    {"num_ctx", "1"},
+		"num_batch 1":                  {"num_batch", "1"},
+		"num_gqa 1":                    {"num_gqa", "1"},
+		"num_gpu 1":                    {"num_gpu", "1"},
+		"main_gpu 1":                   {"main_gpu", "1"},
+		"low_vram true":                {"low_vram", "true"},
+		"f16_kv true":                  {"f16_kv", "true"},
+		"logits_all true":              {"logits_all", "true"},
+		"vocab_only true":              {"vocab_only", "true"},
+		"use_mmap true":                {"use_mmap", "true"},
+		"use_mlock true":               {"use_mlock", "true"},
+		"num_thread 1":                 {"num_thread", "1"},
+		"num_keep 1":                   {"num_keep", "1"},
+		"seed 1":                       {"seed", "1"},
+		"num_predict 1":                {"num_predict", "1"},
+		"top_k 1":                      {"top_k", "1"},
+		"top_p 1.0":                    {"top_p", "1.0"},
+		"tfs_z 1.0":                    {"tfs_z", "1.0"},
+		"typical_p 1.0":                {"typical_p", "1.0"},
+		"repeat_last_n 1":              {"repeat_last_n", "1"},
+		"temperature 1.0":              {"temperature", "1.0"},
+		"repeat_penalty 1.0":           {"repeat_penalty", "1.0"},
+		"presence_penalty 1.0":         {"presence_penalty", "1.0"},
+		"frequency_penalty 1.0":        {"frequency_penalty", "1.0"},
+		"mirostat 1":                   {"mirostat", "1"},
+		"mirostat_tau 1.0":             {"mirostat_tau", "1.0"},
+		"mirostat_eta 1.0":             {"mirostat_eta", "1.0"},
+		"penalize_newline true":        {"penalize_newline", "true"},
+		"stop ### User:":               {"stop", "### User:"},
+		"stop ### User: ":              {"stop", "### User: "},
+		"stop \"### User:\"":           {"stop", "### User:"},
+		"stop \"### User: \"":          {"stop", "### User: "},
+		"stop \"\"\"### User:\"\"\"":   {"stop", "### User:"},
+		"stop \"\"\"### User:\n\"\"\"": {"stop", "### User:\n"},
+		"stop <|endoftext|>":           {"stop", "<|endoftext|>"},
+		"stop <|eot_id|>":              {"stop", "<|eot_id|>"},
+		"stop </s>":                    {"stop", "</s>"},
+	}
+
+	for k, v := range cases {
+		t.Run(k, func(t *testing.T) {
+			var b bytes.Buffer
+			fmt.Fprintln(&b, "FROM foo")
+			fmt.Fprintln(&b, "PARAMETER", k)
+			modelfile, err := ParseFile(&b)
+			assert.NoError(t, err)
+
+			assert.Equal(t, []Command{
+				{Name: "model", Args: "foo"},
+				{Name: v.name, Args: v.value},
+			}, modelfile.Commands)
+		})
+	}
+}
+
+func TestParseFileComments(t *testing.T) {
+	var cases = []struct {
+		input    string
+		expected []Command
+	}{
+		{
+			`
+# comment
+FROM foo
+	`,
+			[]Command{
+				{Name: "model", Args: "foo"},
+			},
+		},
+	}
+
+	for _, c := range cases {
+		t.Run("", func(t *testing.T) {
+			modelfile, err := ParseFile(strings.NewReader(c.input))
+			assert.NoError(t, err)
+			assert.Equal(t, c.expected, modelfile.Commands)
+		})
+	}
+}
+
+func TestParseFileFormatParseFile(t *testing.T) {
+	var cases = []string{
+		`
+FROM foo
+ADAPTER adapter1
+LICENSE MIT
+PARAMETER param1 value1
+PARAMETER param2 value2
+TEMPLATE template1
+MESSAGE system You are a file parser. Always parse things.
+MESSAGE user Hey there!
+MESSAGE assistant Hello, I want to parse all the things!
+`,
+		`
+FROM foo
+ADAPTER adapter1
+LICENSE MIT
+PARAMETER param1 value1
+PARAMETER param2 value2
+TEMPLATE template1
+MESSAGE system """
+You are a store greeter. Always responsed with "Hello!".
+"""
+MESSAGE user Hey there!
+MESSAGE assistant Hello, I want to parse all the things!
+`,
+		`
+FROM foo
+ADAPTER adapter1
+LICENSE """
+Very long and boring legal text.
+Blah blah blah.
+"Oh look, a quote!"
+"""
+
+PARAMETER param1 value1
+PARAMETER param2 value2
+TEMPLATE template1
+MESSAGE system """
+You are a store greeter. Always responsed with "Hello!".
+"""
+MESSAGE user Hey there!
+MESSAGE assistant Hello, I want to parse all the things!
+`,
+	}
+
+	for _, c := range cases {
+		t.Run("", func(t *testing.T) {
+			modelfile, err := ParseFile(strings.NewReader(c))
+			assert.NoError(t, err)
+
+			modelfile2, err := ParseFile(strings.NewReader(modelfile.String()))
+			assert.NoError(t, err)
+
+			assert.Equal(t, modelfile, modelfile2)
+		})
+	}
+
+}

types/model/testdata/fuzz/FuzzName/d37463aa416f6bab

@@ -1,2 +1,2 @@
 go test fuzz v1
-string("/0")
+string("00@")