bfafe30953
### What this PR does / why we need it?
This PR updates the CI configuration and adjusts a set of end-to-end
(e2e) tests under tests/e2e/multicard, in order to refactor the test
suite and ensure compatibility with current codebase and CI workflows.
1. tests/e2e/multicard/test_prefix_caching.py: change model to Qwen3-8B
and rename the test case
2. tests/e2e/multicard/test_quantization.py: rename the test case
3. tests/e2e/multicard/test_qwen3_moe.py: remove duplicate test and
rename test cases
4. tests/e2e/multicard/test_qwen3_next.py: rename test cases and change
the W8A8 pruning model to the W8A8 model and remove the eager parameter
5. tests/e2e/multicard/test_shared_expert_dp.py: rename test case and
remove the eager parameter
6. tests/e2e/multicard/test_single_request_aclgraph.py: rename test case
and change Qwen3-30B to Qwen3-0.6B
7. tests/e2e/multicard/test_torchair_graph_mode.py: delete test cases
about torchair
- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c
Signed-off-by: hfadzxy <starmoon_zhang@163.com>
47 lines
1.6 KiB
Python
47 lines
1.6 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.
|
|
#
|
|
import pytest
|
|
|
|
from tests.e2e.conftest import VllmRunner
|
|
|
|
MODELS = [
|
|
"Qwen/Qwen3-0.6B",
|
|
"deepseek-ai/DeepSeek-V2-Lite-Chat",
|
|
]
|
|
|
|
TENSOR_PARALLELS = [1]
|
|
PIPELINE_PARALLELS = [2]
|
|
DIST_EXECUTOR_BACKEND = ["mp", "ray"]
|
|
|
|
prompts = [
|
|
"Hello, my name is",
|
|
"The future of AI is",
|
|
]
|
|
|
|
|
|
@pytest.mark.parametrize("model", MODELS)
|
|
@pytest.mark.parametrize("tp_size", TENSOR_PARALLELS)
|
|
@pytest.mark.parametrize("pp_size", PIPELINE_PARALLELS)
|
|
@pytest.mark.parametrize("distributed_executor_backend", DIST_EXECUTOR_BACKEND)
|
|
def test_models_pp2(model: str, tp_size: int, pp_size: int,
|
|
distributed_executor_backend: str) -> None:
|
|
with VllmRunner(model,
|
|
tensor_parallel_size=tp_size,
|
|
pipeline_parallel_size=pp_size,
|
|
distributed_executor_backend=distributed_executor_backend,
|
|
gpu_memory_utilization=0.7) as vllm_model:
|
|
vllm_model.generate_greedy(prompts, 64) |