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

fix e2e rejection-sampler error (#5341)

Browse Source 
### What this PR does / why we need it?
Fixed the error in the CI process for
vllm-ascend/tests/e2e/nightly/ops/triton/test_rejection_sampler.py
Error: test_rejection_sampler_block_verify_triton_kernel: duplicate
parametrization of 'vocab_size'.

- vLLM version: release/v0.13.0
- vLLM main:
bc0a5a0c08

Signed-off-by: chenaoxuan <cax1165@163.com>
This commit is contained in:
Aoxuan Chen
2025-12-25 11:39:38 +08:00
committed by GitHub
parent 2ae0bad96d
commit 8caad0510d
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tests/e2e/nightly/ops/triton/test_rejection_sampler.py
View File

@@ -61,7 +61,7 @@ IS_GREEDY = torch.zeros(NUM_TOKENS, dtype=torch.bool, device=DEVICE)
@pytest.mark.parametrize("bonus_token_ids", [BONUS_TOKEN_IDS])
@pytest.mark.parametrize("uniform_probs", [UNIFORM_PROBS])
@pytest.mark.parametrize("is_greedy", [IS_GREEDY])
@pytest.mark.parametrize("vocab_size", [BATCH_SIZE])
@pytest.mark.parametrize("batch_size", [BATCH_SIZE])
@pytest.mark.parametrize("max_spec_len", [MAX_SPEC_LEN])
@pytest.mark.parametrize("vocab_size", [VOCAB_SIZE])
@torch.inference_mode()