obscura/vllm
2
0
Fork 0
mirror of https://github.com/vllm-project/vllm.git synced 2026-08-06 05:48:08 +00:00
Code Issues Packages Projects Releases Wiki Activity

[CI] Bump huggingface-hub from v1.10.2 to v1.22.0 (#47551)

Browse Source 
Signed-off-by: Harry Mellor <[email protected]>
This commit is contained in:
Harry Mellor
2026-07-04 07:45:45 -07:00
committed by GitHub
parent 1d354c694e
commit f1445f6dbd
25 changed files with 49 additions and 45 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.buildkite/test-amd.yaml
+2 -2
View File
@@ -2248,7 +2248,7 @@ steps:
- pytest -v -s v1/worker
- pytest -v -s -m 'not cpu_test' v1/kv_connector/unit
- pytest -v -s -m 'not cpu_test' v1/metrics
- pip install -U git+https://github.com/robertgshaw2-redhat/lm-evaluation-harness.git@streaming-api
- pip install -U git+https://github.com/vllm-project/lm-evaluation-harness.git@streaming-api
# - export HSA_NO_SCRATCH_RECLAIM=1
- pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine
@@ -3318,7 +3318,7 @@ steps:
- pytest -v -s v1/worker
- pytest -v -s -m 'not cpu_test' v1/kv_connector/unit
- pytest -v -s -m 'not cpu_test' v1/metrics
- pip install -U git+https://github.com/robertgshaw2-redhat/lm-evaluation-harness.git@streaming-api
- pip install -U git+https://github.com/vllm-project/lm-evaluation-harness.git@streaming-api
- pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine
- label: V1 Sample + Logits # TBD
.buildkite/test_areas/misc.yaml
+1 -1
View File
@@ -103,7 +103,7 @@ steps:
- pytest -v -s -m 'not cpu_test' v1/kv_connector/unit
- pytest -v -s -m 'not cpu_test' v1/metrics
# Integration test for streaming correctness (requires special branch).
- pip install -U git+https://github.com/robertgshaw2-redhat/lm-evaluation-harness.git@streaming-api
- pip install -U git+https://github.com/vllm-project/lm-evaluation-harness.git@streaming-api
- pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine
mirror:
amd:
requirements/test/cpu.txt
+4 -4
View File
@@ -112,9 +112,10 @@ charset-normalizer==3.4.0
# via requests
chz==0.3.0
# via gpt-oss
click==8.1.7
click==8.4.2
# via
# black
# huggingface-hub
# jiwer
# nltk
# ray
@@ -309,7 +310,7 @@ h2==4.3.0
# via httpx
harfile==0.5.0
# via schemathesis
hf-xet==1.4.3
hf-xet==1.5.1
# via huggingface-hub
hiredis==3.0.0
# via tensorizer
@@ -335,7 +336,7 @@ httpx==0.27.2
# schemathesis
httpx-sse==0.4.3
# via mcp
huggingface-hub==1.10.2
huggingface-hub==1.22.0
# via
# accelerate
# datasets
@@ -1182,7 +1183,6 @@ typer==0.26.8
# fastapi-cli
# fastapi-cloud-cli
# fastsafetensors
# huggingface-hub
# perceptron
# transformers
typing-extensions==4.15.0
requirements/test/cuda.txt
+4 -4
View File
@@ -117,9 +117,10 @@ charset-normalizer==3.4.0
# via requests
chz==0.3.0
# via gpt-oss
click==8.1.7
click==8.4.2
# via
# black
# huggingface-hub
# jiwer
# nltk
# ray
@@ -330,7 +331,7 @@ h2==4.3.0
# via httpx
harfile==0.5.0
# via schemathesis
hf-xet==1.4.3
hf-xet==1.5.1
# via huggingface-hub
hiredis==3.0.0
# via tensorizer
@@ -356,7 +357,7 @@ httpx==0.27.2
# schemathesis
httpx-sse==0.4.3
# via mcp
huggingface-hub==1.10.2
huggingface-hub==1.22.0
# via
# accelerate
# datasets
@@ -1285,7 +1286,6 @@ typer==0.26.8
# fastapi-cli
# fastapi-cloud-cli
# fastsafetensors
# huggingface-hub
# perceptron
# transformers
typing-extensions==4.15.0
requirements/test/rocm.txt
+4 -4
View File
@@ -116,9 +116,10 @@ choreographer==1.2.1
# via kaleido
chz==0.4.0
# via gpt-oss
click==8.3.1
click==8.4.2
# via
# black
# huggingface-hub
# jiwer
# nltk
# ray
@@ -323,7 +324,7 @@ h2==4.3.0
# via httpx
harfile==0.5.0
# via schemathesis
hf-xet==1.4.3
hf-xet==1.5.1
# via huggingface-hub
hiredis==3.3.1
# via tensorizer
@@ -349,7 +350,7 @@ httpx==0.27.2
# schemathesis
httpx-sse==0.4.3
# via mcp
huggingface-hub==1.10.2
huggingface-hub==1.22.0
# via
# accelerate
# datasets
@@ -1244,7 +1245,6 @@ typer==0.24.1
# fastapi-cli
# fastapi-cloud-cli
# fastsafetensors
# huggingface-hub
# perceptron
# transformers
typing-extensions==4.15.0
requirements/test/xpu.txt
+4 -4
View File
@@ -83,8 +83,9 @@ charset-normalizer==3.4.6
# via requests
chz==0.4.0
# via gpt-oss
click==8.3.1
click==8.4.2
# via
# huggingface-hub
# jiwer
# nltk
# rich-toolkit
@@ -206,7 +207,7 @@ h11==0.16.0
# uvicorn
harfile==0.4.0
# via schemathesis
hf-xet==1.4.3
hf-xet==1.5.1
# via huggingface-hub
html2text==2025.4.15
# via gpt-oss
@@ -227,7 +228,7 @@ httpx==0.28.1
# schemathesis
httpx-sse==0.4.3
# via mcp
huggingface-hub==1.10.2
huggingface-hub==1.22.0
# via
# accelerate
# datasets
@@ -959,7 +960,6 @@ typer==0.24.1
# via
# fastapi-cli
# fastapi-cloud-cli
# huggingface-hub
# transformers
typing-extensions==4.15.0
# via
tests/benchmarks/test_bfcl_dataset.py
+1 -1
View File
@@ -21,7 +21,7 @@ def _patch_hf_api(side_effect):
@pytest.fixture(scope="session")
def hf_tokenizer() -> PreTrainedTokenizerBase:
return AutoTokenizer.from_pretrained("gpt2")
return AutoTokenizer.from_pretrained("openai-community/gpt2")
_FAKE_ROWS = {
tests/benchmarks/test_custom_dataset_seed.py
+1 -1
View File
@@ -12,7 +12,7 @@ from vllm.benchmarks.datasets import get_samples
@pytest.fixture(scope="session")
def hf_tokenizer() -> PreTrainedTokenizerBase:
return AutoTokenizer.from_pretrained("gpt2")
return AutoTokenizer.from_pretrained("openai-community/gpt2")
def _write_jsonl(path: Path, n_rows: int) -> None:
tests/benchmarks/test_random_dataset.py
+1 -1
View File
@@ -17,7 +17,7 @@ from vllm.benchmarks.datasets import (
@pytest.fixture(scope="session")
def hf_tokenizer() -> PreTrainedTokenizerBase:
# Use a small, commonly available tokenizer
return AutoTokenizer.from_pretrained("gpt2")
return AutoTokenizer.from_pretrained("openai-community/gpt2")
class Params(NamedTuple):
tests/benchmarks/test_random_multimodal_dataset_video.py
+1 -1
View File
@@ -16,7 +16,7 @@ from vllm.benchmarks.datasets import RandomMultiModalDataset, SampleRequest
@pytest.fixture(scope="session")
def hf_tokenizer() -> PreTrainedTokenizerBase:
"""Use a small, commonly available tokenizer."""
return AutoTokenizer.from_pretrained("gpt2")
return AutoTokenizer.from_pretrained("openai-community/gpt2")
@pytest.fixture
tests/benchmarks/test_txt_slices_dataset.py
+2 -2
View File
@@ -13,7 +13,7 @@ from vllm.benchmarks.datasets.create_txt_slices_dataset import create_txt_slices
@pytest.fixture(scope="session")
def hf_tokenizer() -> PreTrainedTokenizerBase:
# Use a small, commonly available tokenizer
return AutoTokenizer.from_pretrained("gpt2")
return AutoTokenizer.from_pretrained("openai-community/gpt2")
text_content = """
@@ -39,7 +39,7 @@ def test_create_txt_slices_jsonl(
create_txt_slices_jsonl(
input_path=str(txt_path),
output_path=str(jsonl_path),
tokenizer_name="gpt2",
tokenizer_name="openai-community/gpt2",
num_prompts=10,
input_len=10,
output_len=10,
tests/compile/test_aot_compile.py
+2 -2
View File
@@ -502,7 +502,7 @@ def test_gpt2_cache_hit(monkeypatch: pytest.MonkeyPatch):
m.setenv("VLLM_USE_AOT_COMPILE", "1")
# First compilation - initialize model and generate
llm_model = LLM(
model="gpt2",
model="openai-community/gpt2",
compilation_config=CompilationConfig(
mode=CompilationMode.VLLM_COMPILE,
),
@@ -519,7 +519,7 @@ def test_gpt2_cache_hit(monkeypatch: pytest.MonkeyPatch):
# Second compilation - should hit cache
m.setenv("VLLM_FORCE_AOT_LOAD", "1")
llm_model = LLM(
model="gpt2",
model="openai-community/gpt2",
compilation_config=CompilationConfig(
mode=CompilationMode.VLLM_COMPILE,
),
tests/compile/test_dynamic_shapes_compilation.py
+1 -1
View File
@@ -24,7 +24,7 @@ from vllm.utils.torch_utils import is_torch_equal_or_newer
def get_test_models():
"""Get list of models to test based on PyTorch version"""
models = [
"gpt2",
"openai-community/gpt2",
"Qwen/Qwen2-7B-Instruct",
"meta-llama/Llama-3.1-8B",
]
tests/distributed/test_pipeline_parallel.py
+1 -1
View File
@@ -114,7 +114,7 @@ TEXT_GENERATION_MODELS = {
"tiiuae/falcon-7b": PPTestSettings.fast(),
"google/gemma-1.1-2b-it": PPTestSettings.fast(),
"google/gemma-2-9b": PPTestSettings.fast(),
"gpt2": PPTestSettings.fast(),
"openai-community/gpt2": PPTestSettings.fast(),
"EleutherAI/gpt-j-6b": PPTestSettings.fast(),
"EleutherAI/pythia-1.4b": PPTestSettings.fast(),
"ibm/PowerLM-3b": PPTestSettings.fast(),
tests/entrypoints/openai/completion/test_prompt_validation.py
+2 -2
View File
@@ -18,7 +18,7 @@ from vllm.renderers.embed_utils import safe_load_prompt_embeds
@pytest.mark.asyncio
async def test_empty_prompt():
model_name = "gpt2"
model_name = "openai-community/gpt2"
server_args = ["--enforce-eager"]
with RemoteOpenAIServer(model_name, server_args) as remote_server:
client = remote_server.get_async_client()
@@ -38,7 +38,7 @@ async def test_empty_prompt():
@pytest.mark.asyncio
async def test_out_of_vocab_token_ids():
model_name = "gpt2"
model_name = "openai-community/gpt2"
server_args = ["--enforce-eager"]
with RemoteOpenAIServer(model_name, server_args) as remote_server:
client = remote_server.get_async_client()
tests/lora/test_chatglm3_tp.py
+1
View File
@@ -115,6 +115,7 @@ def test_chatglm3_lora_tp4_fully_sharded_loras(chatglm3_lora_files):
enable_lora=True,
max_loras=2,
max_lora_rank=64,
max_num_seqs=16,
tensor_parallel_size=4,
trust_remote_code=True,
fully_sharded_loras=True,
tests/models/registry.py
+1 -1
View File
@@ -296,7 +296,7 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
"GlmMoeDsaForCausalLM": _HfExamplesInfo(
"zai-org/GLM-5", min_transformers_version="5.0.1", is_available_online=False
),
"GPT2LMHeadModel": _HfExamplesInfo("openai-community/gpt2", {"alias": "gpt2"}),
"GPT2LMHeadModel": _HfExamplesInfo("openai-community/gpt2"),
"GPTBigCodeForCausalLM": _HfExamplesInfo(
"bigcode/starcoder",
extras={
tests/renderers/test_chat_utils_prompt_embeds.py
+1 -1
View File
@@ -40,7 +40,7 @@ from vllm.renderers.hf import (
# Qwen2TokenizerFast (SentencePiece BPE variant)
# BertTokenizerFast (WordPiece)
TOKENIZER_IDS: Final[list[str]] = [
"gpt2",
"openai-community/gpt2",
"Qwen/Qwen2.5-1.5B-Instruct",
"bert-base-uncased",
]
tests/tokenizers_/test_basic.py
+4 -2
View File
@@ -23,7 +23,7 @@ def _assert_tokenizer_like(tokenizer: object):
def test_tokenizer_like_protocol():
tokenizer = get_tokenizer("gpt2", use_fast=True)
tokenizer = get_tokenizer("openai-community/gpt2", use_fast=True)
assert isinstance(tokenizer, PreTrainedTokenizerFast)
_assert_tokenizer_like(tokenizer)
@@ -43,7 +43,9 @@ def test_tokenizer_like_protocol():
_assert_tokenizer_like(tokenizer)
@pytest.mark.parametrize("tokenizer_name", ["facebook/opt-125m", "gpt2"])
@pytest.mark.parametrize(
"tokenizer_name", ["facebook/opt-125m", "openai-community/gpt2"]
)
def test_tokenizer_revision(tokenizer_name: str):
# Assume that "main" branch always exists
tokenizer = get_tokenizer(tokenizer_name, revision="main")
tests/tokenizers_/test_detokenize.py
+1 -1
View File
@@ -33,7 +33,7 @@ TRUTH = [
TOKENIZERS = [
"facebook/opt-125m",
"gpt2",
"openai-community/gpt2",
"bigcode/tiny_starcoder_py",
"EleutherAI/gpt-j-6b",
"EleutherAI/pythia-70m",
tests/tokenizers_/test_hf.py
+2 -2
View File
@@ -14,7 +14,7 @@ from vllm.tokenizers.hf import (
)
@pytest.mark.parametrize("model_id", ["gpt2", "zai-org/chatglm3-6b"])
@pytest.mark.parametrize("model_id", ["openai-community/gpt2", "zai-org/chatglm3-6b"])
def test_cached_tokenizer(model_id: str):
reference_tokenizer = AutoTokenizer.from_pretrained(
model_id, trust_remote_code=True
@@ -47,7 +47,7 @@ def _check_consistency(target: TokenizerLike, expected: TokenizerLike):
assert target.encode("prompt") == expected.encode("prompt")
@pytest.mark.parametrize("model_id", ["gpt2"])
@pytest.mark.parametrize("model_id", ["openai-community/gpt2"])
def test_thread_pool_tokenizer_pickle(model_id: str):
"""Regression test for issue #45433: the thread-pool tokenizer wrapper
reconstructs through maybe_make_thread_pool on unpickling, which used to
tests/tool_parsers/conftest.py
+1 -1
View File
@@ -9,4 +9,4 @@ from vllm.tokenizers import TokenizerLike
@pytest.fixture(scope="module")
def default_tokenizer() -> TokenizerLike:
return AutoTokenizer.from_pretrained("gpt2")
return AutoTokenizer.from_pretrained("openai-community/gpt2")
tests/tool_parsers/test_gigachat3_tool_parser.py
+1 -1
View File
@@ -19,7 +19,7 @@ from vllm.tool_parsers import ToolParser, ToolParserManager
def default_tokenizer() -> TokenizerLike:
"""Override module-scoped default_tokenizer because gigachat tests
mutate the tokenizer via ``add_tokens``."""
return AutoTokenizer.from_pretrained("gpt2")
return AutoTokenizer.from_pretrained("openai-community/gpt2")
MSG_SEP_TOKEN = "<|message_sep|>\n\n"
tests/v1/structured_output/test_backend_guidance.py
+1 -1
View File
@@ -17,7 +17,7 @@ from vllm.v1.structured_output import StructuredOutputManager
from vllm.v1.structured_output.backend_guidance import GuidanceBackend
from vllm.v1.structured_output.backend_types import StructuredOutputOptions
TOKENIZER = "gpt2"
TOKENIZER = "openai-community/gpt2"
@pytest.fixture(scope="module")
vllm/transformers_utils/repo_utils.py
+5 -4
View File
@@ -229,10 +229,11 @@ def get_model_path(model: str | Path, revision: str | None = None):
if os.path.exists(model):
return model
assert huggingface_hub.constants.HF_HUB_OFFLINE
common_kwargs = {
"local_files_only": huggingface_hub.constants.HF_HUB_OFFLINE,
"revision": revision,
}
common_kwargs = dict(
local_files_only=huggingface_hub.constants.HF_HUB_OFFLINE,
ignore_patterns="*",
revision=revision,
)
if envs.VLLM_USE_MODELSCOPE:
from modelscope.hub.snapshot_download import snapshot_download