Adding entrypoints in docker file

Signed-off-by: Yuan Zhou <[email protected]>
vllm-project · Jul 2, 2024 · 551c442 · 551c442
1 parent 30ee3ce
commit 551c442
Showing 1 changed file with 4 additions and 2 deletions.
diff --git a/.buildkite/run-cpu-test.sh b/.buildkite/run-cpu-test.sh
@@ -12,8 +12,10 @@ trap remove_docker_container EXIT
 remove_docker_container
 
 # Run the image
-docker run -itd -v ~/.cache/huggingface:/root/.cache/huggingface --cpuset-cpus=48-95 --cpuset-mems=1 --network host -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --name cpu-test cpu-test
-docker run -itd -v ~/.cache/huggingface:/root/.cache/huggingface --cpuset-cpus=48-95 --cpuset-mems=1 --network host -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --name cpu-test-avx2 cpu-test-avx2
+docker run -itd --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --cpuset-cpus=48-95 \
+  --cpuset-mems=1 --network host -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --name cpu-test cpu-test
+docker run -itd --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --cpuset-cpus=48-95 \
+  --cpuset-mems=1 --network host -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --name cpu-test-avx2 cpu-test-avx2
 
 # offline inference
 docker exec cpu-test bash -c "python3 examples/offline_inference.py"