f8ff2d6879
enable metal gpu acceleration
...
ggml-metal.metal must be in the same directory as the ollama binary
otherwise llama.cpp will not be able to find it and load it.
1. go generate llama/llama_metal.go
2. go build .
3. ./ollama serve
2023-07-05 18:11:12 -07:00
52ac534889
remove debug print statements
2023-07-05 16:29:12 -04:00
328c06eb14
separate routes
2023-07-05 15:37:49 -04:00
79ebc5c5a1
wip
2023-07-04 00:47:00 -04:00
a3ce37e7da
move prompt templates out of python bindings
2023-07-03 17:14:20 -04:00
7c257434fc
add llama.cpp go bindings
2023-07-03 16:32:48 -04:00
3432181e3e
wip go engine
...
Co-authored-by: Patrick Devine <pdevine@sonic.net>
2023-07-03 15:22:44 -04:00
b361fa72ec
reorganize directories
2023-06-25 13:08:03 -04:00
d3709f85b5
build server into desktop app
2023-06-25 00:30:02 -04:00
369108e1ad
Add dependencies to `requirements.txt`
2023-06-24 21:13:26 -04:00
5142ba2dad
Update README.md
2023-06-23 22:51:54 -04:00
c5bafaff54
package server with client
2023-06-23 18:38:22 -04:00
f0eee3faa0
build server executable
2023-06-23 17:23:30 -04:00
ebec1c61db
load and unload model endpoints
2023-06-23 14:47:57 -04:00
0758cb2d4b
llama server wrapper
2023-06-23 13:10:13 -04:00
8fa91332fa
initial commit
2023-06-22 18:31:40 -04:00
Michael Yang
Jeffrey Morgan
Bruce MacDonald
Michael Chiang