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

Convert formatting to use ruff instead of yapf + isort (#26247)

Browse Source 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2025-10-05 15:06:22 +01:00
committed by GitHub
parent 17edd8a807
commit d6953beb91
1508 changed files with 115244 additions and 94146 deletions
Download Patch File Download Diff File Expand all files Collapse all files

67
tests/models/language/pooling_mteb_test/test_qwen3_reranker.py
View File

@@ -18,46 +18,45 @@ qwen3_reranker_hf_overrides = {
}
RERANK_MODELS = [
LASTPoolingRerankModelInfo("Qwen/Qwen3-Reranker-0.6B",
architecture="Qwen3ForSequenceClassification",
mteb_score=0.25736,
hf_overrides=qwen3_reranker_hf_overrides,
enable_test=True),
LASTPoolingRerankModelInfo("Qwen/Qwen3-Reranker-4B",
architecture="Qwen3ForSequenceClassification",
hf_overrides=qwen3_reranker_hf_overrides,
enable_test=False)
LASTPoolingRerankModelInfo(
"Qwen/Qwen3-Reranker-0.6B",
architecture="Qwen3ForSequenceClassification",
mteb_score=0.25736,
hf_overrides=qwen3_reranker_hf_overrides,
enable_test=True,
),
LASTPoolingRerankModelInfo(
"Qwen/Qwen3-Reranker-4B",
architecture="Qwen3ForSequenceClassification",
hf_overrides=qwen3_reranker_hf_overrides,
enable_test=False,
),
]
class Qwen3RerankerHfRunner(HfRunner):
def __init__(self,
model_name: str,
dtype: str = "auto",
*args: Any,
**kwargs: Any) -> None:
def __init__(
self, model_name: str, dtype: str = "auto", *args: Any, **kwargs: Any
) -> None:
from transformers import AutoModelForCausalLM, AutoTokenizer
super().__init__(model_name, dtype, auto_cls=AutoModelForCausalLM)
self.tokenizer = AutoTokenizer.from_pretrained(model_name,
padding_side='left')
self.tokenizer = AutoTokenizer.from_pretrained(model_name, padding_side="left")
self.token_false_id = self.tokenizer.convert_tokens_to_ids("no")
self.token_true_id = self.tokenizer.convert_tokens_to_ids("yes")
def predict(self, prompts: list[list[str]], *args,
**kwargs) -> torch.Tensor:
def predict(self, prompts: list[list[str]], *args, **kwargs) -> torch.Tensor:
def process_inputs(pairs):
inputs = self.tokenizer(pairs,
padding=False,
truncation='longest_first',
return_attention_mask=False)
for i, ele in enumerate(inputs['input_ids']):
inputs['input_ids'][i] = ele
inputs = self.tokenizer.pad(inputs,
padding=True,
return_tensors="pt")
inputs = self.tokenizer(
pairs,
padding=False,
truncation="longest_first",
return_attention_mask=False,
)
for i, ele in enumerate(inputs["input_ids"]):
inputs["input_ids"][i] = ele
inputs = self.tokenizer.pad(inputs, padding=True, return_tensors="pt")
for key in inputs:
inputs[key] = inputs[key].to(self.model.device)
return inputs
@@ -82,20 +81,18 @@ class Qwen3RerankerHfRunner(HfRunner):
@pytest.mark.parametrize("model_info", RERANK_MODELS)
def test_rerank_models_mteb(vllm_runner, model_info: RerankModelInfo) -> None:
mteb_test_rerank_models(Qwen3RerankerHfRunner, vllm_runner, model_info)
@pytest.mark.parametrize("model_info", RERANK_MODELS)
@multi_gpu_test(num_gpus=2)
def test_rerank_models_mteb_tp(vllm_runner,
model_info: RerankModelInfo) -> None:
def test_rerank_models_mteb_tp(vllm_runner, model_info: RerankModelInfo) -> None:
assert model_info.architecture == "Qwen3ForSequenceClassification"
vllm_extra_kwargs: dict[str, Any] = {
"tensor_parallel_size": 2,
}
mteb_test_rerank_models(Qwen3RerankerHfRunner, vllm_runner, model_info,
vllm_extra_kwargs)
mteb_test_rerank_models(
Qwen3RerankerHfRunner, vllm_runner, model_info, vllm_extra_kwargs
)