update workflows with cuda build

.github/workflows/build-commit.yml

@@ -75,6 +75,17 @@ jobs:
     # See: https://docs.github.com/en/free-pro-team@latest/actions/learn-github-actions/managing-complex-workflows#using-a-build-matrix
     runs-on: ubuntu-latest
 
+    strategy:
+          matrix:
+            include:
+            - build: 'avx2'
+              defines: ''
+            - build: 'avx'
+              defines: '-DLLAMA_AVX2=OFF'
+            - build: 'avx512'
+              defines: '-DLLAMA_AVX512=ON'
+
+
     steps:
     - uses: actions/checkout@v3
       with:

.github/workflows/docker-image.yml

@@ -13,6 +13,12 @@ jobs:
 
     runs-on: ubuntu-latest
 
+    strategy:
+      matrix:
+        config:
+          - {tag: "", dockerfile: "./Dockerfile.default"}
+          - {tags: "-cuda", dockerfile: "./Dockerfile.cuda"}
+
     steps:
     
     - name: Checkout
@@ -49,10 +55,19 @@ jobs:
 
     - name: Build and push release
       uses: docker/build-push-action@v4
-      if: startsWith(github.ref, 'refs/tags/')
+      if: startsWith(github.ref, 'refs/tags/') && matrix.config.tag == ""
       with:
         push: true
         tags: ghcr.io/ravenscroftj/turbopilot:${{ github.ref_name }}, ghcr.io/ravenscroftj/turbopilot:latest
         context: ${{github.workspace}}
-        platforms: linux/amd64,linux/arm64
-      
+        platforms: linux/amd64,linux/arm64  
+
+
+    - name: Build and push release (CUDA)
+      uses: docker/build-push-action@v4
+      if: startsWith(github.ref, 'refs/tags/') && matrix.config.tag != ""
+      with:
+        push: true
+        tags: ghcr.io/ravenscroftj/turbopilot:${{ github.ref_name }}
+        context: ${{github.workspace}}
+        platforms: linux/amd64,linux/arm64  

Dockerfile.cuda

@@ -0,0 +1,29 @@
+FROM nvidia/cuda:11.7.1-devel-ubuntu22.04 AS build
+
+RUN apt-get update && apt-get install -y build-essential cmake libboost-dev
+
+ADD ./ggml /build/ggml
+
+RUN mkdir /build/ggml/build
+
+WORKDIR  /build/ggml/build
+
+RUN cmake -DGGML_CUBLAS=ON -DCMAKE_CUDA_COMPILER=/usr/local/cuda/bin/nvcc ..
+RUN make codegen-serve
+
+FROM nvidia/cuda:11.7.1-cudnn8-runtime-ubuntu22.04 AS runtime
+
+
+WORKDIR /app
+
+COPY --from=build /build/ggml/build/bin/codegen-serve /app/codegen-serve
+
+ENV THREADS=4
+
+ENV MODEL="/models/codegen-2B-multi-ggml-4bit-quant.bin"
+
+COPY ./run.sh /app/
+
+EXPOSE 18080
+
+CMD /app/run.sh