xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-06-09 21:42:13 +08:00
Code Issues Packages Projects Releases Wiki Activity

[CI/Build] Enable eager mode for Samplers-Test group

Browse Source 
Signed-off-by: qli88 <qiang.li2@amd.com>
This commit is contained in:
qli88 2025-11-21 07:22:43 +00:00
parent 0075bfffd4
commit eea8b6634c
2 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
tests/samplers/test_logprobs.py
View File

@ -25,7 +25,7 @@ def test_ranks(
flat_logprobs, flat_logprobs,
example_prompts, example_prompts,
): ):
with vllm_runner(model, dtype=dtype, max_logprobs=MAX_LOGPROBS) as vllm_model: with vllm_runner(model, dtype=dtype, max_logprobs=MAX_LOGPROBS, enforce_eager=True) as vllm_model:
tokenizer = vllm_model.llm.get_tokenizer() tokenizer = vllm_model.llm.get_tokenizer()
example_prompt_tokens = [tokenizer.encode(prompt) for prompt in example_prompts] example_prompt_tokens = [tokenizer.encode(prompt) for prompt in example_prompts]
sampling_params = SamplingParams( sampling_params = SamplingParams(

2
tests/samplers/test_no_bad_words.py
View File

@ -94,7 +94,7 @@ class TestTwoTokenBadWord:
)[0] )[0]
def test_two_token_bad_word(self, vllm_runner): def test_two_token_bad_word(self, vllm_runner):
with vllm_runner(self.MODEL, dtype="half") as llm: with vllm_runner(self.MODEL, dtype="half", enforce_eager=True) as llm:
output_token_ids = self._generate(llm) output_token_ids = self._generate(llm)
assert output_token_ids[:2] == [ assert output_token_ids[:2] == [
self.target_token_id1, self.target_token_id1,