biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Attention] Support distinguishing between short extends and decodes (#37303)

Browse Source 
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
This commit is contained in:
Lucas Wilkinson
2026-03-20 10:49:36 -07:00
committed by GitHub
parent 79eb9369c5
commit e1d85e5c24
9 changed files with 176 additions and 133 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/v1/e2e/test_hybrid_chunked_prefill.py
View File

@@ -43,7 +43,7 @@ MESSAGES = [
pytest.param("Qwen/Qwen3.5-4B", marks=[large_gpu_mark(min_gb=40)]),
pytest.param(
"nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8",
marks=[large_gpu_mark(min_gb=80)] + multi_gpu_marks(num_gpus=2),
marks=[large_gpu_mark(min_gb=80)] + multi_gpu_marks(num_gpus=4),
),
],
)
@@ -68,7 +68,7 @@ def test_mtp_speculative_mixed_batch_short_prefill(
max_num_batched_tokens=chunk_size,
max_model_len=512,
enforce_eager=True,
tensor_parallel_size=2,
tensor_parallel_size=4,
trust_remote_code=True,
enable_chunked_prefill=True,
enable_prefix_caching=enable_prefix_caching,