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

[Misc] unify variable for LLM instance (#20996)

Browse Source 
Signed-off-by: Andy Xie <andy.xning@gmail.com>
This commit is contained in:
Ning Xie
2025-07-21 19:18:33 +08:00
committed by GitHub
parent e6b90a2805
commit d97841078b
53 changed files with 237 additions and 236 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
tests/v1/engine/test_llm_engine.py
View File

@@ -112,9 +112,9 @@ def test_compatibility_with_skip_tokenizer_init(
example_prompts,
structured_outputs=True,
)
model: LLM = vllm_model_skip_tokenizer_init.model
llm: LLM = vllm_model_skip_tokenizer_init.llm
with pytest.raises(ValueError):
_ = model.generate(example_prompts, sampling_params_list)
_ = llm.generate(example_prompts, sampling_params_list)
def test_parallel_sampling(vllm_model, example_prompts) -> None:
@@ -125,8 +125,8 @@ def test_parallel_sampling(vllm_model, example_prompts) -> None:
example_prompt: test fixture providing prompts for testing.
"""
sampling_params_list, n_list = _get_test_sampling_params(example_prompts)
model: LLM = vllm_model.model
outputs = model.generate(example_prompts, sampling_params_list)
llm: LLM = vllm_model.llm
outputs = llm.generate(example_prompts, sampling_params_list)
# Validate each request response
for out, n in zip(outputs, n_list):
@@ -166,10 +166,10 @@ def test_engine_metrics(vllm_runner, monkeypatch, example_prompts):
speculative_config=speculative_config,
disable_log_stats=False,
) as vllm_model:
model: LLM = vllm_model.model
llm: LLM = vllm_model.llm
sampling_params = SamplingParams(temperature=0.0,
max_tokens=max_tokens)
outputs = model.generate(example_prompts, sampling_params)
outputs = llm.generate(example_prompts, sampling_params)
n_prompts = len(example_prompts)
assert len(outputs) == n_prompts
@@ -180,7 +180,7 @@ def test_engine_metrics(vllm_runner, monkeypatch, example_prompts):
total_tokens += len(out.outputs[0].token_ids)
assert total_tokens == max_tokens * n_prompts
metrics = model.get_metrics()
metrics = llm.get_metrics()
def find_metric(name) -> list[Metric]:
found = []