ci: use clang for windows cpu builds

clang outputs are faster. we were previously building with clang via gcc
wrapper in cgo but this was missed during the build updates so there was
a drop in performance

.github/workflows/release.yaml

@@ -160,6 +160,10 @@ jobs:
           echo "$hipPath\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
           echo "CC=$hipPath\bin\clang.exe" | Out-File -FilePath $env:GITHUB_ENV -Append
           echo "CXX=$hipPath\bin\clang++.exe" | Out-File -FilePath $env:GITHUB_ENV -Append
+      - if: matrix.preset == 'CPU'
+        run: |
+          echo "CC=clang.exe" | Out-File -FilePath $env:GITHUB_ENV -Append
+          echo "CXX=clang++.exe" | Out-File -FilePath $env:GITHUB_ENV -Append
       - if: ${{ !cancelled() && steps.cache-install.outputs.cache-hit != 'true' }}
         uses: actions/cache/save@v4
         with:

README.md

@@ -384,6 +384,7 @@ See the [API documentation](./docs/api.md) for all endpoints.
 - [OpenDeepResearcher-via-searxng](https://github.com/benhaotang/OpenDeepResearcher-via-searxng) (A Deep Research equivent endpoint with Ollama support for running locally)
 - [AntSK](https://github.com/AIDotNet/AntSK) (Out-of-the-box & Adaptable RAG Chatbot)
 - [MaxKB](https://github.com/1Panel-dev/MaxKB/) (Ready-to-use & flexible RAG Chatbot)
+- [yla](https://github.com/danielekp/yla) (Web interface to freely interact with your customized models)
 
 ### Cloud
 

server/sched.go

@@ -179,7 +179,7 @@ func (s *Scheduler) processPending(ctx context.Context) {
 						if allReliable {
 							// HACK
 							os.Setenv("OLLAMA_MAX_LOADED_MODELS", strconv.Itoa(defaultModelsPerGPU*len(gpus)))
-							slog.Debug("updating default concurrency", "OLLAMA_MAX_LOADED_MODELS", envconfig.MaxRunners, "gpu_count", len(gpus))
+							slog.Debug("updating default concurrency", "OLLAMA_MAX_LOADED_MODELS", envconfig.MaxRunners(), "gpu_count", len(gpus))
 						} else {
 							// HACK
 							os.Setenv("OLLAMA_MAX_LOADED_MODELS", strconv.Itoa(len(gpus)))