[v1] Add FlashAttention selection and implement normal / padding-free / dynamic batching (#10469)

This commit is contained in:
jiaqiw09
2026-05-21 17:14:19 +08:00
committed by GitHub
parent 7e20db5735
commit bdcb92d035
23 changed files with 507 additions and 105 deletions

View File

@@ -35,10 +35,3 @@ def test_sync_sampler():
"role": "assistant",
"content": [{"type": "text", "value": "This is a test."}],
}
if __name__ == "__main__":
"""
python tests_v1/sampler/test_cli_sampler.py
"""
test_sync_sampler()