diff --git a/.buildkite/test-amd.yaml b/.buildkite/test-amd.yaml index 0049f35403409..37c6bd4276722 100644 --- a/.buildkite/test-amd.yaml +++ b/.buildkite/test-amd.yaml @@ -1323,7 +1323,7 @@ steps: - vllm/ - tests/weight_loading commands: - - bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models.txt + - bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models-amd.txt - label: Weight Loading Multiple GPU Test - Large Models # optional mirror_hardwares: [amdexperimental] @@ -1331,13 +1331,12 @@ steps: # grade: Blocking working_dir: "/vllm-workspace/tests" num_gpus: 2 - gpu: a100 optional: true source_file_dependencies: - vllm/ - tests/weight_loading commands: - - bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models-large.txt + - bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models-large-amd.txt - label: NixlConnector PD accuracy tests (Distributed) # 30min mirror_hardwares: [amdexperimental] diff --git a/tests/weight_loading/models-amd.txt b/tests/weight_loading/models-amd.txt new file mode 100644 index 0000000000000..e31e904c08af4 --- /dev/null +++ b/tests/weight_loading/models-amd.txt @@ -0,0 +1,3 @@ +fp8, amd/Meta-Llama-3.1-8B-Instruct-FP8-KV, main +None, amd/Llama-3.2-1B-Instruct-FP8-KV, main +fp8, amd/Mixtral-8x7B-Instruct-v0.1-FP8-KV, main diff --git a/tests/weight_loading/models-large-amd.txt b/tests/weight_loading/models-large-amd.txt new file mode 100644 index 0000000000000..b6f5b4b16b37f --- /dev/null +++ b/tests/weight_loading/models-large-amd.txt @@ -0,0 +1,3 @@ +fp8, amd/Meta-Llama-3.1-70B-Instruct-FP8-KV, main +None, microsoft/phi-4, main +fp8, amd/Mixtral-8x22B-Instruct-v0.1-FP8-KV, main