diff --git a/.buildkite/test-amd.yaml b/.buildkite/test-amd.yaml index 3675065791..954aae904a 100644 --- a/.buildkite/test-amd.yaml +++ b/.buildkite/test-amd.yaml @@ -2838,12 +2838,13 @@ steps: - uv pip install --system 'gpt-oss[eval]==0.0.5' - pytest -s -v evals/gpt_oss/test_gpqa_correctness.py --config-list-file=configs/models-gfx950.txt -- label: LM Eval Qwen3-5 Models (B200-MI355) # TBD - timeout_in_minutes: 120 +- label: LM Eval Qwen3-5 Models (B200-MI355) %N # TBD + timeout_in_minutes: 180 mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355] agent_pool: mi355_2 num_gpus: 2 optional: true + parallelism: 4 working_dir: "/vllm-workspace/tests" source_file_dependencies: - vllm/model_executor/models/qwen3_5.py @@ -2858,7 +2859,7 @@ steps: - vllm/_aiter_ops.py - vllm/platforms/rocm.py commands: - - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=configs/models-qwen35-mi355.txt + - pytest -s -v evals/gsm8k/test_gsm8k_correctness.py --config-list-file=configs/models-qwen35-mi355.txt --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT - label: LM Eval Small Models (2xB200-2xMI355) # TBD timeout_in_minutes: 180 diff --git a/tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-MXFP4-AITER-TP2.yaml b/tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-MXFP4-AITER-TP2.yaml index 657251a660..2c0431747d 100644 --- a/tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-MXFP4-AITER-TP2.yaml +++ b/tests/evals/gsm8k/configs/Qwen3.5-35B-A3B-MXFP4-AITER-TP2.yaml @@ -3,6 +3,7 @@ accuracy_threshold: 0.89 tolerance: 0.03 num_questions: 1319 num_fewshot: 5 +startup_max_wait_seconds: 3600 server_args: >- --max-model-len 4096 --tensor-parallel-size 2