jeans/ollama
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

first pass at linux gpu support (#454)

Browse Source 
* linux gpu support
* handle multiple gpus
* add cuda docker image (#488)
---------

Co-authored-by: Michael Yang <mxyng@pm.me>
This commit is contained in:
Bruce MacDonald
2023-09-12 11:04:35 -04:00
committed by GitHub
parent 45ac07cd02
commit f221637053
9 changed files with 158 additions and 22 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
llm/llama.cpp/generate_darwin_arm64.go
View File

@@ -1,12 +1,16 @@
package llm
//go:generate git submodule init
//go:generate git submodule update --force ggml gguf
//go:generate git -C ggml apply ../ggml_patch/0001-add-detokenize-endpoint.patch
//go:generate git -C ggml apply ../ggml_patch/0002-34B-model-support.patch
//go:generate git -C ggml apply ../ggml_patch/0003-metal-fix-synchronization-in-new-matrix-multiplicati.patch
//go:generate git -C ggml apply ../ggml_patch/0004-metal-add-missing-barriers-for-mul-mat-2699.patch
//go:generate cmake --fresh -S ggml -B ggml/build/gpu -DLLAMA_METAL=on -DLLAMA_ACCELERATE=on -DLLAMA_K_QUANTS=on -DCMAKE_SYSTEM_PROCESSOR=arm64 -DCMAKE_OSX_ARCHITECTURES=arm64 -DCMAKE_OSX_DEPLOYMENT_TARGET=11.0
//go:generate git submodule update --force ggml
//go:generate -command git-apply git -C ggml apply
//go:generate git-apply ../ggml_patch/0001-add-detokenize-endpoint.patch
//go:generate git-apply ../ggml_patch/0002-34B-model-support.patch
//go:generate git-apply ../ggml_patch/0003-metal-fix-synchronization-in-new-matrix-multiplicati.patch
//go:generate git-apply ../ggml_patch/0004-metal-add-missing-barriers-for-mul-mat-2699.patch
//go:generate cmake -S ggml -B ggml/build/gpu -DLLAMA_METAL=on -DLLAMA_ACCELERATE=on -DLLAMA_K_QUANTS=on -DCMAKE_SYSTEM_PROCESSOR=arm64 -DCMAKE_OSX_ARCHITECTURES=arm64 -DCMAKE_OSX_DEPLOYMENT_TARGET=11.0
//go:generate cmake --build ggml/build/gpu --target server --config Release
//go:generate git submodule update --force gguf
//go:generate cmake -S gguf -B gguf/build/gpu -DLLAMA_METAL=on -DLLAMA_ACCELERATE=on -DLLAMA_K_QUANTS=on -DCMAKE_SYSTEM_PROCESSOR=arm64 -DCMAKE_OSX_ARCHITECTURES=arm64 -DCMAKE_OSX_DEPLOYMENT_TARGET=11.0
//go:generate cmake --build gguf/build/gpu --target server --config Release