diff --git a/.github/workflows/_e2e_test.yaml b/.github/workflows/_e2e_test.yaml index 904e4eb3..01ce509f 100644 --- a/.github/workflows/_e2e_test.yaml +++ b/.github/workflows/_e2e_test.yaml @@ -86,6 +86,7 @@ jobs: pip install uc-manager uv pip install -r requirements-dev.txt uv pip install -v -e . + uv pip install git+https://github.com/modelscope/modelscope.git@dbbcbf631fe6d10cc6446df2ad2fef24039fe7fe - name: Run vllm-project/vllm-ascend test env: @@ -177,6 +178,7 @@ jobs: pip install uc-manager uv pip install -r requirements-dev.txt uv pip install -v -e . + uv pip install git+https://github.com/modelscope/modelscope.git@dbbcbf631fe6d10cc6446df2ad2fef24039fe7fe - name: Run e2e test env: VLLM_WORKER_MULTIPROC_METHOD: spawn @@ -265,6 +267,7 @@ jobs: pip install uc-manager uv pip install -r requirements-dev.txt uv pip install -v -e . + uv pip install git+https://github.com/modelscope/modelscope.git@dbbcbf631fe6d10cc6446df2ad2fef24039fe7fe - name: Run vllm-project/vllm-ascend test (light) env: VLLM_WORKER_MULTIPROC_METHOD: spawn @@ -353,6 +356,7 @@ jobs: pip install uc-manager uv pip install -r requirements-dev.txt uv pip install -v -e . + uv pip install git+https://github.com/modelscope/modelscope.git@dbbcbf631fe6d10cc6446df2ad2fef24039fe7fe - name: Run vllm-project/vllm-ascend test (full) env: VLLM_WORKER_MULTIPROC_METHOD: spawn @@ -448,6 +452,7 @@ jobs: pip install uc-manager uv pip install -r requirements-dev.txt uv pip install -v -e . + uv pip install git+https://github.com/modelscope/modelscope.git@dbbcbf631fe6d10cc6446df2ad2fef24039fe7fe - name: Run vllm-project/vllm-ascend test for V1 Engine env: @@ -529,6 +534,7 @@ jobs: pip install uc-manager uv pip install -r requirements-dev.txt uv pip install -v -e . + uv pip install git+https://github.com/modelscope/modelscope.git@dbbcbf631fe6d10cc6446df2ad2fef24039fe7fe - name: Run vllm-project/vllm-ascend test env: @@ -590,6 +596,7 @@ jobs: pip install uc-manager uv pip install -r requirements-dev.txt uv pip install -v -e . + uv pip install git+https://github.com/modelscope/modelscope.git@dbbcbf631fe6d10cc6446df2ad2fef24039fe7fe - name: Run vllm-project/vllm-ascend test env: diff --git a/tests/e2e/conftest.py b/tests/e2e/conftest.py index edef3664..d124b9ce 100644 --- a/tests/e2e/conftest.py +++ b/tests/e2e/conftest.py @@ -34,6 +34,7 @@ import traceback from pathlib import Path from typing import Any, TypeVar +import huggingface_hub import numpy as np import openai import psutil @@ -1024,7 +1025,10 @@ class HfRunner: @pytest.fixture(scope="session") def ilama_lora_files(): - return snapshot_download(repo_id="vllm-ascend/ilama-text2sql-spider") + return snapshot_download( + repo_id="vllm-ascend/ilama-text2sql-spider", + local_files_only=huggingface_hub.constants.HF_HUB_OFFLINE, + ) @pytest.fixture(scope="session") diff --git a/tests/e2e/multicard/2-cards/test_external_launcher.py b/tests/e2e/multicard/2-cards/test_external_launcher.py index e74b3585..0ce41945 100644 --- a/tests/e2e/multicard/2-cards/test_external_launcher.py +++ b/tests/e2e/multicard/2-cards/test_external_launcher.py @@ -26,6 +26,7 @@ import sys from pathlib import Path from unittest.mock import patch +import huggingface_hub import pytest import torch_npu from modelscope import snapshot_download # type: ignore @@ -165,7 +166,10 @@ def test_qwen3_external_launcher_with_sleepmode(): def test_qwen3_external_launcher_with_sleepmode_level2(): script = Path(__file__).parent.parent.parent.parent.parent / "examples" / "offline_external_launcher.py" env = os.environ.copy() - model_path = snapshot_download("Qwen/Qwen3-8B") + model_path = snapshot_download( + "Qwen/Qwen3-8B", + local_files_only=huggingface_hub.constants.HF_HUB_OFFLINE, + ) # TODO: Add moe model test cmd = [ sys.executable, diff --git a/tests/e2e/singlecard/pooling/test_classification.py b/tests/e2e/singlecard/pooling/test_classification.py index 8bdd3660..e41ca01d 100644 --- a/tests/e2e/singlecard/pooling/test_classification.py +++ b/tests/e2e/singlecard/pooling/test_classification.py @@ -1,13 +1,14 @@ import torch from modelscope import snapshot_download # type: ignore[import-untyped] from transformers import AutoModelForSequenceClassification +import huggingface_hub from tests.e2e.conftest import HfRunner, VllmRunner def test_qwen_pooling_classify_correctness() -> None: - model_name = snapshot_download("Howeee/Qwen2.5-1.5B-apeach") + model_name = snapshot_download("Howeee/Qwen2.5-1.5B-apeach", local_files_only=huggingface_hub.constants.HF_HUB_OFFLINE,) prompts = [ "Hello, my name is", diff --git a/tests/e2e/singlecard/pooling/test_embedding.py b/tests/e2e/singlecard/pooling/test_embedding.py index 54eae677..50dc9ee9 100644 --- a/tests/e2e/singlecard/pooling/test_embedding.py +++ b/tests/e2e/singlecard/pooling/test_embedding.py @@ -18,6 +18,7 @@ # import pytest from modelscope import snapshot_download # type: ignore[import-untyped] +import huggingface_hub from tests.e2e.conftest import HfRunner, VllmRunner from tests.e2e.utils import check_embeddings_close @@ -32,7 +33,7 @@ MODELS = [ def test_embed_models_correctness(model: str): queries = ['What is the capital of China?', 'Explain gravity'] - model_name = snapshot_download(model) + model_name = snapshot_download(model, local_files_only=huggingface_hub.constants.HF_HUB_OFFLINE,) with VllmRunner( model_name, runner="pooling", @@ -60,7 +61,7 @@ def test_embed_models_correctness(model: str): def test_bge_m3_correctness(): queries = ['What is the capital of China?', 'Explain gravity'] - model_name = snapshot_download("BAAI/bge-m3") + model_name = snapshot_download("BAAI/bge-m3", local_files_only=huggingface_hub.constants.HF_HUB_OFFLINE,) with VllmRunner( model_name, runner="pooling", diff --git a/tests/e2e/singlecard/pooling/test_scoring.py b/tests/e2e/singlecard/pooling/test_scoring.py index fb81b6f2..2913a361 100644 --- a/tests/e2e/singlecard/pooling/test_scoring.py +++ b/tests/e2e/singlecard/pooling/test_scoring.py @@ -3,6 +3,7 @@ import pytest import torch import torch.nn.functional as F +import huggingface_hub from modelscope import snapshot_download # type: ignore[import-untyped] from tests.e2e.conftest import HfRunner, VllmRunner @@ -31,7 +32,7 @@ DTYPE = "half" @pytest.fixture(scope="module", params=CROSS_ENCODER_MODELS) def model_name(request): - yield snapshot_download(request.param) + yield snapshot_download(request.param, local_files_only=huggingface_hub.constants.HF_HUB_OFFLINE,) def test_cross_encoder_score_1_to_1(model_name): text_pair = [TEXTS_1[0], TEXTS_2[0]] @@ -100,7 +101,7 @@ def test_cross_encoder_score_N_to_N(model_name): @pytest.fixture(scope="module", params=EMBEDDING_MODELS) def emb_model_name(request): - yield snapshot_download(request.param) + yield snapshot_download(request.param, local_files_only=huggingface_hub.constants.HF_HUB_OFFLINE,) def test_embedding_score_1_to_1(emb_model_name): diff --git a/tools/send_mm_request.py b/tools/send_mm_request.py index 8b3033ab..164fba61 100644 --- a/tools/send_mm_request.py +++ b/tools/send_mm_request.py @@ -1,10 +1,15 @@ import base64 import os +import huggingface_hub import requests from modelscope import snapshot_download # type: ignore -mm_dir = snapshot_download("vllm-ascend/mm_request", repo_type="dataset") +mm_dir = snapshot_download( + "vllm-ascend/mm_request", + repo_type="dataset", + local_files_only=huggingface_hub.constants.HF_HUB_OFFLINE, +) image_path = os.path.join(mm_dir, "test_mm2.jpg") with open(image_path, "rb") as image_file: image_data = base64.b64encode(image_file.read()).decode("utf-8")