EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[CI] Skip test_suffix_correctness (#4820)

Browse Source 
### What this PR does / why we need it?
Currently, suffix decoding has known correctness issue see
https://github.com/vllm-project/vllm-ascend/actions/runs/20033509824/job/57457565620?pr=4781"

Signed-off-by: wangli <wangli858794774@gmail.com>
This commit is contained in:
Li Wang
2025-12-09 11:48:13 +08:00
committed by GitHub
parent 2b819bb35b
commit 1c70f5c922
1 changed files with 4 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/e2e/singlecard/spec_decode_v1/test_v1_spec_decode.py
View File

@@ -151,6 +151,8 @@ def test_eagle_correctness(
assert matches > int(0.66 * len(ref_outputs))
@pytest.mark.skip(
"Fix me, suffix decoding now exists some known accuracy issue, skip it")
def test_suffix_correctness(
test_prompts: list[list[dict[str, Any]]],
sampling_config: SamplingParams,
@@ -187,6 +189,8 @@ def test_suffix_correctness(
assert matches > int(0.66 * len(ref_outputs))
@pytest.mark.skip(
"Fix me, suffix decoding now exists some functional issue, skip it")
def test_suffix_acceptance(
test_prompts: list[list[dict[str, Any]]],
sampling_config: SamplingParams,