update docker builds

.github/workflows/docker-image.yml

@@ -17,7 +17,8 @@ jobs:
       matrix:
         config:
           - {tag: "", dockerfile: "./Dockerfile.default", platforms: "linux/amd64,linux/arm64"}
-          - {tag: "-cuda", dockerfile: "./Dockerfile.cuda", platforms: "linux/amd64"}
+          - {tag: "-cuda11", dockerfile: "./Dockerfile.cuda11", platforms: "linux/amd64"}
+          - {tag: "-cuda12", dockerfile: "./Dockerfile.cuda12", platforms: "linux/amd64"}
 
     steps:
     
@@ -54,7 +55,7 @@ jobs:
         platforms: ${{matrix.config.platforms}}
 
 
-    - name: Build and push release
+    - name: Build and push release (Main Latest Build)
       uses: docker/build-push-action@v4
       if: startsWith(github.ref, 'refs/tags/') && matrix.config.tag == ''
       with:
@@ -65,12 +66,12 @@ jobs:
         platforms: ${{matrix.config.platforms}} 
 
 
-    - name: Build and push release (CUDA)
+    - name: Build and push release (Accelerated Builds)
       uses: docker/build-push-action@v4
       if: startsWith(github.ref, 'refs/tags/') && matrix.config.tag != ''
       with:
         file: ${{matrix.config.dockerfile}}
         push: true
-        tags: ghcr.io/ravenscroftj/turbopilot:${{ github.ref_name }}
+        tags: ghcr.io/ravenscroftj/turbopilot:${{ github.ref_name }}${{matrix.config.tag}}
         context: ${{github.workspace}}
         platforms: linux/amd64,linux/arm64  

Dockerfile.cuda11

@@ -0,0 +1,39 @@
+FROM nvidia/cuda:11.7.1-devel-ubuntu22.04 AS build
+
+ENV DEBIAN_FRONTEND=noninteractive
+
+# inlude kitware apt repo to allow us to grab latest cmake
+RUN apt-get update && apt-get install ca-certificates gpg wget
+RUN wget -O - https://apt.kitware.com/keys/kitware-archive-latest.asc 2>/dev/null | gpg --dearmor - | tee /usr/share/keyrings/kitware-archive-keyring.gpg >/dev/null
+RUN echo 'deb [signed-by=/usr/share/keyrings/kitware-archive-keyring.gpg] https://apt.kitware.com/ubuntu/ jammy main' | tee /etc/apt/sources.list.d/kitware.list >/dev/null
+
+RUN apt-get update && apt-get install -y build-essential cmake libboost-dev
+
+
+ADD ./ /turbopilot
+
+RUN mkdir /turbopilot/build
+
+WORKDIR  /turbopilot/build
+
+RUN cmake -DGGML_CUBLAS=ON -DCMAKE_CUDA_COMPILER=/usr/local/cuda/bin/nvcc ..
+RUN make turbopilot
+
+FROM nvidia/cuda:11.7.1-cudnn8-runtime-ubuntu22.04 AS runtime
+
+
+WORKDIR /app
+
+COPY --from=build /turbopilot/build/bin/turbopilot /app/turbopilot
+
+ENV THREADS=4
+
+ENV MODEL="/models/codegen-2B-multi-ggml-4bit-quant.bin"
+
+ENV BATCHSIZE=64
+
+COPY ./run.sh /app/
+
+EXPOSE 18080
+
+CMD /app/run.sh

Dockerfile.cuda12

@@ -0,0 +1,37 @@
+FROM nvidia/cuda:12.2.0-devel-ubuntu20.04 AS build
+
+ENV DEBIAN_FRONTEND=noninteractive
+
+# inlude kitware apt repo to allow us to grab latest cmake
+RUN apt-get update && apt-get install ca-certificates gpg wget
+RUN wget -O - https://apt.kitware.com/keys/kitware-archive-latest.asc 2>/dev/null | gpg --dearmor - | tee /usr/share/keyrings/kitware-archive-keyring.gpg >/dev/null
+RUN echo 'deb [signed-by=/usr/share/keyrings/kitware-archive-keyring.gpg] https://apt.kitware.com/ubuntu/ focal main' | tee /etc/apt/sources.list.d/kitware.list >/dev/null
+
+RUN apt-get update && apt-get install -y build-essential cmake libboost-dev
+
+ADD ./ /turbopilot
+
+RUN mkdir /turbopilot/build
+
+WORKDIR  /turbopilot/build
+
+RUN cmake -DGGML_CUBLAS=ON -DCMAKE_CUDA_COMPILER=/usr/local/cuda/bin/nvcc ..
+RUN make turbopilot
+
+FROM nvidia/cuda:12.2.0-runtime-ubuntu20.04 AS runtime
+
+WORKDIR /app
+
+COPY --from=build /turbopilot/build/bin/turbopilot /app/turbopilot
+
+ENV THREADS=4
+
+ENV MODEL="/models/codegen-2B-multi-ggml-4bit-quant.bin"
+
+ENV BATCHSIZE=64
+
+COPY ./run.sh /app/
+
+EXPOSE 18080
+
+CMD /app/run.sh