Edit docker build command

Bump vllm version of sagemaker config
lightonai · Sep 9, 2024 · 7ff4582 · 7ff4582
1 parent a28db04
commit 7ff4582
Show file tree

Hide file tree

Showing 4 changed files with 4 additions and 4 deletions.
diff --git a/docker/build.sh b/docker/build.sh
@@ -1 +1 @@
-DOCKER_BUILDKIT=1 docker build . --target vllm-openai --tag vllm --build-arg max_jobs=8 --build-arg nvcc_threads=8
+DOCKER_BUILDKIT=1 docker build . --target vllm-openai --tag vllm --build-arg max_jobs=8 --build-arg nvcc_threads=8 --build-arg VLLM_MAX_SIZE_MB=300
diff --git a/docker/deploy.sh b/docker/deploy.sh
@@ -31,7 +31,7 @@ then
 fi
 
 # Build the Docker image with specified build arguments.
-if DOCKER_BUILDKIT=1 docker build . --target vllm-openai --tag "${REPOSITORY_NAME}" --build-arg max_jobs=8 --build-arg nvcc_threads=8
+if DOCKER_BUILDKIT=1 docker build . --target vllm-openai --tag "${REPOSITORY_NAME}" --build-arg max_jobs=8 --build-arg nvcc_threads=8 --build-arg VLLM_MAX_SIZE_MB=300
 then 
     docker tag "${REPOSITORY_NAME}" "$CONTAINER_URI:${VERSION_NUMBER}"
     docker push "$CONTAINER_URI:${VERSION_NUMBER}"

diff --git a/sagemaker/configs/llama3_1-70b.json b/sagemaker/configs/llama3_1-70b.json
@@ -1,6 +1,6 @@
 {
     "model": "meta-llama/Meta-Llama-3.1-70B-Instruct",
-    "image": "vllm:0.5.3.post1-2",
+    "image": "vllm:0.6.0-1",
     "sagemaker_instance_type": "ml.p4d.24xlarge",
     "env_vars": {
         "TENSOR_PARALLEL_SIZE": "8",

diff --git a/sagemaker/configs/llama3_1-8b.json b/sagemaker/configs/llama3_1-8b.json
@@ -1,6 +1,6 @@
 {
     "model": "meta-llama/Meta-Llama-3.1-8B-Instruct",
-    "image": "vllm:0.5.3.post1-2",
+    "image": "vllm:0.6.0-1",
     "sagemaker_instance_type": "ml.g5.2xlarge",
     "env_vars": {
         "TENSOR_PARALLEL_SIZE": "1",
Original file line number	Diff line number	Diff line change
		@@ -1 +1 @@
		DOCKER_BUILDKIT=1 docker build . --target vllm-openai --tag vllm --build-arg max_jobs=8 --build-arg nvcc_threads=8
		DOCKER_BUILDKIT=1 docker build . --target vllm-openai --tag vllm --build-arg max_jobs=8 --build-arg nvcc_threads=8 --build-arg VLLM_MAX_SIZE_MB=300