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

[Core] Support logprobs with spec decode + async scheduling (#29223)

Browse Source 
Signed-off-by: Nick Hill <nhill@redhat.com>
This commit is contained in:
Nick Hill
2025-11-25 12:55:24 -08:00
committed by GitHub
parent e7d776273d
commit 4e57c6587f
4 changed files with 35 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
tests/v1/e2e/test_async_scheduling.py
View File

@@ -87,6 +87,11 @@ def test_with_spec_decoding(monkeypatch: pytest.MonkeyPatch):
# Set small draft model len to force doesn't-fit-in-drafter case.
spec_config_short = spec_config | {"max_model_len": 50}
test_sampling_params = [
dict(),
dict(logprobs=2),
]
# test_preemption, executor, async_scheduling,
# spec_config, test_prefill_chunking
test_configs = [
@@ -103,7 +108,7 @@ def test_with_spec_decoding(monkeypatch: pytest.MonkeyPatch):
(True, "uni", True, spec_config_short, True),
]
run_tests(monkeypatch, MTP_MODEL, test_configs, [{}])
run_tests(monkeypatch, MTP_MODEL, test_configs, test_sampling_params)
@dynamo_config.patch(cache_size_limit=16)