xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-11 13:05:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

[CI] Tweaks to GPT-OSS Eval (Blackwell) for stability (#26030)

Browse Source 
Signed-off-by: mgoin <mgoin64@gmail.com>
This commit is contained in:
Michael Goin 2025-10-01 15:02:17 -04:00 committed by GitHub
parent c36f0aa300
commit ee04c0cd04
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 3 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
.buildkite/test-pipeline.yaml
View File

@ -845,7 +845,7 @@ steps:
- vllm/v1/attention/backends/flashinfer.py - vllm/v1/attention/backends/flashinfer.py
commands: commands:
- uv pip install --system 'gpt-oss[eval]==0.0.5' - uv pip install --system 'gpt-oss[eval]==0.0.5'
- pytest -s -v tests/evals/gpt_oss/test_gpqa_correctness.py --model openai/gpt-oss-20b --metric 0.58 --server-args '--tensor-parallel-size 2' - pytest -s -v tests/evals/gpt_oss/test_gpqa_correctness.py --model openai/gpt-oss-20b --metric 0.58
- label: Blackwell Quantized MoE Test - label: Blackwell Quantized MoE Test
timeout_in_minutes: 60 timeout_in_minutes: 60

5
tests/evals/gpt_oss/test_gpqa_correctness.py
View File

@ -26,7 +26,8 @@ def run_gpqa_eval(model_name: str, base_url: str) -> float:
# Build the command to run the evaluation # Build the command to run the evaluation
cmd = [ cmd = [
sys.executable, "-m", "gpt_oss.evals", "--eval", "gpqa", "--model", sys.executable, "-m", "gpt_oss.evals", "--eval", "gpqa", "--model",
model_name, "--reasoning-effort", "low", "--base-url", base_url model_name, "--reasoning-effort", "low", "--base-url", base_url,
"--n-threads", "200"
] ]
try: try:
@ -72,8 +73,6 @@ def test_gpqa_correctness(request):
# Add standard server arguments # Add standard server arguments
server_args.extend([ server_args.extend([
"--max-model-len",
"32768",
"--trust-remote-code", "--trust-remote-code",
]) ])