5f8241c25c
### What this PR does / why we need it? Change as little existing code as possible to add v1 pooling task's support, notice that i move down the `vllm.v1.worker.gpu_input_batch` to vllm-ascend, Considering the frequent changes in upstream interfaces, in order to decouple, so i move it here ### How was this patch tested? CI passed with new added/existing test, and I have a simple test was first conducted locally which is adapted from https://www.modelscope.cn/models/Qwen/Qwen3-Embedding-0.6B, just like bellow： ```python import os import torch from vllm import LLM os.environ["VLLM_USE_MODELSCOPE"]="True" def get_detailed_instruct(task_description: str, query: str) -> str: return f'Instruct: {task_description}\nQuery:{query}' # Each query must come with a one-sentence instruction that describes the task task = 'Given a web search query, retrieve relevant passages that answer the query' queries = [ get_detailed_instruct(task, 'What is the capital of China?'), get_detailed_instruct(task, 'Explain gravity') ] # No need to add instruction for retrieval documents documents = [ "The capital of China is Beijing.", "Gravity is a force that attracts two bodies towards each other. It gives weight to physical objects and is responsible for the movement of planets around the sun." ] input_texts = queries + documents model = LLM(model="Qwen/Qwen3-Embedding-0.6B", task="embed") outputs = model.embed(input_texts) embeddings = torch.tensor([o.outputs.embedding for o in outputs]) scores = (embeddings[:2] @ embeddings[2:].T) print(scores.tolist()) # [[0.7620252966880798, 0.14078938961029053], [0.1358368694782257, 0.6013815999031067]] ``` --------- Signed-off-by: wangli <wangli858794774@gmail.com> Signed-off-by: wangli <858794774@qq.com> Co-authored-by: wangli <858794774@qq.com>
73 lines
2.3 KiB
Python
73 lines
2.3 KiB
Python
#
|
|
# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
|
|
# Copyright 2023 The vLLM team.
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
# This file is a part of the vllm-ascend project.
|
|
# Adapted from vllm/tests/basic_correctness/test_basic_correctness.py
|
|
#
|
|
from collections.abc import Sequence
|
|
from typing import Optional
|
|
|
|
import pytest
|
|
from modelscope import snapshot_download # type: ignore[import-untyped]
|
|
|
|
from tests.conftest import HfRunner
|
|
from tests.utils import check_embeddings_close, matryoshka_fy
|
|
from vllm_ascend.utils import vllm_version_is
|
|
|
|
|
|
def run_embedding_correctness_test(
|
|
hf_model: "HfRunner",
|
|
inputs: list[str],
|
|
vllm_outputs: Sequence[list[float]],
|
|
dimensions: Optional[int] = None,
|
|
):
|
|
hf_outputs = hf_model.encode(inputs)
|
|
if dimensions:
|
|
hf_outputs = matryoshka_fy(hf_outputs, dimensions)
|
|
|
|
check_embeddings_close(
|
|
embeddings_0_lst=hf_outputs,
|
|
embeddings_1_lst=vllm_outputs,
|
|
name_0="hf",
|
|
name_1="vllm",
|
|
tol=1e-2,
|
|
)
|
|
|
|
|
|
# dummy to avoid pytest collect nothing and exit code 5
|
|
def test_dummy():
|
|
assert True
|
|
|
|
|
|
@pytest.mark.skipif(vllm_version_is("0.9.1"),
|
|
reason="vLLM 0.9.1 does not support embed task for v1")
|
|
def test_embed_models_correctness(hf_runner, vllm_runner):
|
|
queries = ['What is the capital of China?', 'Explain gravity']
|
|
|
|
model_name = snapshot_download("Qwen/Qwen3-Embedding-0.6B")
|
|
with vllm_runner(
|
|
model_name,
|
|
task="embed",
|
|
enforce_eager=True,
|
|
) as vllm_model:
|
|
vllm_outputs = vllm_model.encode(queries)
|
|
|
|
with hf_runner(
|
|
model_name,
|
|
dtype="float32",
|
|
is_sentence_transformer=True,
|
|
) as hf_model:
|
|
run_embedding_correctness_test(hf_model, queries, vllm_outputs)
|