ci: 1.20.0 release(fix cuda dependency) (#69)

kibae · Nov 4, 2024 · e1b8546 · e1b8546
1 parent a446878
commit e1b8546
Show file tree

Hide file tree

Showing 5 changed files with 21 additions and 21 deletions.
diff --git a/README.md b/README.md
@@ -158,11 +158,11 @@ sudo cmake --install build --prefix /usr/local/onnxruntime-server
 # Docker
 
 - Docker hub: [kibaes/onnxruntime-server](https://hub.docker.com/r/kibaes/onnxruntime-server)
-    - [`1.20.0-linux-cuda11`](https://github.com/kibae/onnxruntime-server/blob/main/deploy/build-docker/linux-cuda11.dockerfile) amd64(CUDA 11.x, cuDNN 8.x)
+    - [`1.20.0-linux-cuda12`](https://github.com/kibae/onnxruntime-server/blob/main/deploy/build-docker/linux-cuda12.dockerfile) amd64(CUDA 12.x, cuDNN 9.x)
     - [`1.20.0-linux-cpu`](https://github.com/kibae/onnxruntime-server/blob/main/deploy/build-docker/linux-cpu.dockerfile) amd64, arm64
 
 ```shell
-DOCKER_IMAGE=kibae/onnxruntime-server:1.20.0-linux-cuda11 # or kibae/onnxruntime-server:1.20.0-linux-cpu	
+DOCKER_IMAGE=kibae/onnxruntime-server:1.20.0-linux-cuda12 # or kibae/onnxruntime-server:1.20.0-linux-cpu	
 
 docker pull ${DOCKER_IMAGE}
 

diff --git a/deploy/build-docker/build.sh b/deploy/build-docker/build.sh
@@ -29,19 +29,19 @@ if [ "$1" != "--target=cpu" ]; then
   # |  |     |  |  |  | |  |  |  |  /  /_\  \ |______>   <   | '_ \  |__   _|
   # |  `----.|  `--'  | |  '--'  | /  _____  \      /  .  \  | (_) |    | |
   #  \______| \______/  |_______/ /__/     \__\    /__/ \__\  \___/     |_|
-  POSTFIX=linux-cuda11
-  IMAGE_NAME=${IMAGE_PREFIX}:${VERSION}-${POSTFIX}
-
-  docker buildx build --platform linux/amd64 -t ${IMAGE_NAME} -f deploy/build-docker/${POSTFIX}.dockerfile --load . || exit 1
-  ./deploy/build-docker/docker-image-test.sh ${IMAGE_NAME} 1 || exit 1
-  docker buildx build --platform linux/amd64 -t ${IMAGE_NAME} -f deploy/build-docker/${POSTFIX}.dockerfile --push . || exit 1
-
-
-#  POSTFIX=linux-cuda12
+#  POSTFIX=linux-cuda11
 #  IMAGE_NAME=${IMAGE_PREFIX}:${VERSION}-${POSTFIX}
 #
 #  docker buildx build --platform linux/amd64 -t ${IMAGE_NAME} -f deploy/build-docker/${POSTFIX}.dockerfile --load . || exit 1
 #  ./deploy/build-docker/docker-image-test.sh ${IMAGE_NAME} 1 || exit 1
 #  docker buildx build --platform linux/amd64 -t ${IMAGE_NAME} -f deploy/build-docker/${POSTFIX}.dockerfile --push . || exit 1
+
+
+  POSTFIX=linux-cuda12
+  IMAGE_NAME=${IMAGE_PREFIX}:${VERSION}-${POSTFIX}
+
+  docker buildx build --platform linux/amd64 -t ${IMAGE_NAME} -f deploy/build-docker/${POSTFIX}.dockerfile --load . || exit 1
+  ./deploy/build-docker/docker-image-test.sh ${IMAGE_NAME} 1 || exit 1
+  docker buildx build --platform linux/amd64 -t ${IMAGE_NAME} -f deploy/build-docker/${POSTFIX}.dockerfile --push . || exit 1
 fi
 
diff --git a/deploy/build-docker/docker-compose.yaml b/deploy/build-docker/docker-compose.yaml
@@ -5,7 +5,7 @@ services:
   onnxruntime_server_simple:
     # After the docker container is up, you can use the REST API (http://localhost:8080).
     # API documentation will be available at http://localhost:8080/api-docs.
-    image: kibaes/onnxruntime-server:1.20.0-linux-cuda11
+    image: kibaes/onnxruntime-server:1.20.0-linux-cuda12
     ports:
       - "8080:80" # for http backend
     volumes:
@@ -29,7 +29,7 @@ services:
   onnxruntime_server_advanced:
     # After the docker container is up, you can use the REST API (http://localhost, https://localhost).
     # API documentation will be available at http://localhost/api-docs.
-    image: kibaes/onnxruntime-server:1.20.0-linux-cuda11
+    image: kibaes/onnxruntime-server:1.20.0-linux-cuda12
     ports:
       - "80:80" # for http backend
       - "443:443" # for https backend

diff --git a/deploy/build-docker/linux-cuda12.dockerfile b/deploy/build-docker/linux-cuda12.dockerfile
@@ -1,4 +1,4 @@
-FROM nvidia/cuda:12.3.2-cudnn9-devel-ubuntu22.04 AS builder
+FROM nvidia/cuda:12.4.1-cudnn-devel-ubuntu22.04 AS builder
 
 RUN apt update && apt install -y curl wget git build-essential cmake pkg-config libboost-all-dev libssl-dev
 RUN mkdir -p /app/source
@@ -15,18 +15,18 @@ RUN case ${TARGETPLATFORM} in \
          "linux/amd64")  ./download-onnxruntime.sh linux x64-gpu ;; \
     esac
 
-RUN cmake -DCUDA_SDK_ROOT_DIR=/usr/local/cuda -DBoost_USE_STATIC_LIBS=ON -DOPENSSL_USE_STATIC_LIBS=ON -B build -S . -DCMAKE_BUILD_TYPE=Release
+RUN cmake -DCUDA_SDK_ROOT_DIR=/usr/local/cuda-12 -DBoost_USE_STATIC_LIBS=ON -DOPENSSL_USE_STATIC_LIBS=ON -B build -S . -DCMAKE_BUILD_TYPE=Release
 RUN cmake --build build --parallel 4 --target onnxruntime_server_standalone
 RUN cmake --install build --prefix /app/onnxruntime-server
 
 # target
-FROM nvidia/cuda:12.3.2-cudnn9-runtime-ubuntu22.04 AS target
+FROM nvidia/cuda:12.4.1-cudnn-runtime-ubuntu22.04 AS target
 COPY --from=builder /app/onnxruntime-server /app
 COPY --from=builder /usr/local/onnxruntime /usr/local/onnxruntime
 
 WORKDIR /app
 RUN mkdir -p models logs certs
 
 ENV ONNX_SERVER_CONFIG_PRIORITY=env
-ENV LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/lib64:/usr/local/onnxruntime/lib
+ENV LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda-12/lib64:/usr/local/onnxruntime/lib
 ENTRYPOINT ["/app/bin/onnxruntime_server", "--model-dir", "models", "--log-file", "logs/app.log", "--access-log-file", "logs/access.log", "--tcp-port", "6432", "--http-port", "80"]
diff --git a/docs/docker.md b/docs/docker.md
@@ -5,7 +5,7 @@
 
 # Supported tags and respective Dockerfile links
 
-- [`1.20.0-linux-cuda11`](https://github.com/kibae/onnxruntime-server/blob/main/deploy/build-docker/linux-cuda11.dockerfile) amd64(CUDA 11.x, cuDNN 8.x)
+- [`1.20.0-linux-cuda12`](https://github.com/kibae/onnxruntime-server/blob/main/deploy/build-docker/linux-cuda12.dockerfile) amd64(CUDA 12.x, cuDNN 9.x)
 - [`1.20.0-linux-cpu`](https://github.com/kibae/onnxruntime-server/blob/main/deploy/build-docker/linux-cpu.dockerfile) amd64, arm64
 
 # How to use this image
@@ -28,7 +28,7 @@
     - API documentation will be available at http://localhost/api-docs.
 
 ```shell
-DOCKER_IMAGE=kibae/onnxruntime-server:1.20.0-linux-cuda11 # or kibae/onnxruntime-server:1.20.0-linux-cpu	
+DOCKER_IMAGE=kibae/onnxruntime-server:1.20.0-linux-cuda12 # or kibae/onnxruntime-server:1.20.0-linux-cpu	
 
 docker pull ${DOCKER_IMAGE}
 
@@ -69,7 +69,7 @@ services:
   onnxruntime_server_simple:
     # After the docker container is up, you can use the REST API (http://localhost:8080).
     # API documentation will be available at http://localhost:8080/api-docs.
-    image: kibaes/onnxruntime-server:1.20.0-linux-cuda11
+    image: kibaes/onnxruntime-server:1.20.0-linux-cuda12
     ports:
       - "8080:80" # for http backend
     volumes:
@@ -101,7 +101,7 @@ services:
   onnxruntime_server_advanced:
     # After the docker container is up, you can use the REST API (http://localhost, https://localhost).
     # API documentation will be available at http://localhost/api-docs.
-    image: kibaes/onnxruntime-server:1.20.0-linux-cuda11
+    image: kibaes/onnxruntime-server:1.20.0-linux-cuda12
     ports:
       - "80:80" # for http backend
       - "443:443" # for https backend