[CI] Bump huggingface-hub from v1.10.2 to v1.22.0 (#47551)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
@@ -2248,7 +2248,7 @@ steps:
|
|||||||
- pytest -v -s v1/worker
|
- pytest -v -s v1/worker
|
||||||
- pytest -v -s -m 'not cpu_test' v1/kv_connector/unit
|
- pytest -v -s -m 'not cpu_test' v1/kv_connector/unit
|
||||||
- pytest -v -s -m 'not cpu_test' v1/metrics
|
- 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
|
# - export HSA_NO_SCRATCH_RECLAIM=1
|
||||||
- pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine
|
- 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 v1/worker
|
||||||
- pytest -v -s -m 'not cpu_test' v1/kv_connector/unit
|
- pytest -v -s -m 'not cpu_test' v1/kv_connector/unit
|
||||||
- pytest -v -s -m 'not cpu_test' v1/metrics
|
- 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
|
- pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine
|
||||||
|
|
||||||
- label: V1 Sample + Logits # TBD
|
- label: V1 Sample + Logits # TBD
|
||||||
|
|||||||
@@ -103,7 +103,7 @@ steps:
|
|||||||
- pytest -v -s -m 'not cpu_test' v1/kv_connector/unit
|
- pytest -v -s -m 'not cpu_test' v1/kv_connector/unit
|
||||||
- pytest -v -s -m 'not cpu_test' v1/metrics
|
- pytest -v -s -m 'not cpu_test' v1/metrics
|
||||||
# Integration test for streaming correctness (requires special branch).
|
# 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
|
- pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine
|
||||||
mirror:
|
mirror:
|
||||||
amd:
|
amd:
|
||||||
|
|||||||
@@ -112,9 +112,10 @@ charset-normalizer==3.4.0
|
|||||||
# via requests
|
# via requests
|
||||||
chz==0.3.0
|
chz==0.3.0
|
||||||
# via gpt-oss
|
# via gpt-oss
|
||||||
click==8.1.7
|
click==8.4.2
|
||||||
# via
|
# via
|
||||||
# black
|
# black
|
||||||
|
# huggingface-hub
|
||||||
# jiwer
|
# jiwer
|
||||||
# nltk
|
# nltk
|
||||||
# ray
|
# ray
|
||||||
@@ -309,7 +310,7 @@ h2==4.3.0
|
|||||||
# via httpx
|
# via httpx
|
||||||
harfile==0.5.0
|
harfile==0.5.0
|
||||||
# via schemathesis
|
# via schemathesis
|
||||||
hf-xet==1.4.3
|
hf-xet==1.5.1
|
||||||
# via huggingface-hub
|
# via huggingface-hub
|
||||||
hiredis==3.0.0
|
hiredis==3.0.0
|
||||||
# via tensorizer
|
# via tensorizer
|
||||||
@@ -335,7 +336,7 @@ httpx==0.27.2
|
|||||||
# schemathesis
|
# schemathesis
|
||||||
httpx-sse==0.4.3
|
httpx-sse==0.4.3
|
||||||
# via mcp
|
# via mcp
|
||||||
huggingface-hub==1.10.2
|
huggingface-hub==1.22.0
|
||||||
# via
|
# via
|
||||||
# accelerate
|
# accelerate
|
||||||
# datasets
|
# datasets
|
||||||
@@ -1182,7 +1183,6 @@ typer==0.26.8
|
|||||||
# fastapi-cli
|
# fastapi-cli
|
||||||
# fastapi-cloud-cli
|
# fastapi-cloud-cli
|
||||||
# fastsafetensors
|
# fastsafetensors
|
||||||
# huggingface-hub
|
|
||||||
# perceptron
|
# perceptron
|
||||||
# transformers
|
# transformers
|
||||||
typing-extensions==4.15.0
|
typing-extensions==4.15.0
|
||||||
|
|||||||
@@ -117,9 +117,10 @@ charset-normalizer==3.4.0
|
|||||||
# via requests
|
# via requests
|
||||||
chz==0.3.0
|
chz==0.3.0
|
||||||
# via gpt-oss
|
# via gpt-oss
|
||||||
click==8.1.7
|
click==8.4.2
|
||||||
# via
|
# via
|
||||||
# black
|
# black
|
||||||
|
# huggingface-hub
|
||||||
# jiwer
|
# jiwer
|
||||||
# nltk
|
# nltk
|
||||||
# ray
|
# ray
|
||||||
@@ -330,7 +331,7 @@ h2==4.3.0
|
|||||||
# via httpx
|
# via httpx
|
||||||
harfile==0.5.0
|
harfile==0.5.0
|
||||||
# via schemathesis
|
# via schemathesis
|
||||||
hf-xet==1.4.3
|
hf-xet==1.5.1
|
||||||
# via huggingface-hub
|
# via huggingface-hub
|
||||||
hiredis==3.0.0
|
hiredis==3.0.0
|
||||||
# via tensorizer
|
# via tensorizer
|
||||||
@@ -356,7 +357,7 @@ httpx==0.27.2
|
|||||||
# schemathesis
|
# schemathesis
|
||||||
httpx-sse==0.4.3
|
httpx-sse==0.4.3
|
||||||
# via mcp
|
# via mcp
|
||||||
huggingface-hub==1.10.2
|
huggingface-hub==1.22.0
|
||||||
# via
|
# via
|
||||||
# accelerate
|
# accelerate
|
||||||
# datasets
|
# datasets
|
||||||
@@ -1285,7 +1286,6 @@ typer==0.26.8
|
|||||||
# fastapi-cli
|
# fastapi-cli
|
||||||
# fastapi-cloud-cli
|
# fastapi-cloud-cli
|
||||||
# fastsafetensors
|
# fastsafetensors
|
||||||
# huggingface-hub
|
|
||||||
# perceptron
|
# perceptron
|
||||||
# transformers
|
# transformers
|
||||||
typing-extensions==4.15.0
|
typing-extensions==4.15.0
|
||||||
|
|||||||
@@ -116,9 +116,10 @@ choreographer==1.2.1
|
|||||||
# via kaleido
|
# via kaleido
|
||||||
chz==0.4.0
|
chz==0.4.0
|
||||||
# via gpt-oss
|
# via gpt-oss
|
||||||
click==8.3.1
|
click==8.4.2
|
||||||
# via
|
# via
|
||||||
# black
|
# black
|
||||||
|
# huggingface-hub
|
||||||
# jiwer
|
# jiwer
|
||||||
# nltk
|
# nltk
|
||||||
# ray
|
# ray
|
||||||
@@ -323,7 +324,7 @@ h2==4.3.0
|
|||||||
# via httpx
|
# via httpx
|
||||||
harfile==0.5.0
|
harfile==0.5.0
|
||||||
# via schemathesis
|
# via schemathesis
|
||||||
hf-xet==1.4.3
|
hf-xet==1.5.1
|
||||||
# via huggingface-hub
|
# via huggingface-hub
|
||||||
hiredis==3.3.1
|
hiredis==3.3.1
|
||||||
# via tensorizer
|
# via tensorizer
|
||||||
@@ -349,7 +350,7 @@ httpx==0.27.2
|
|||||||
# schemathesis
|
# schemathesis
|
||||||
httpx-sse==0.4.3
|
httpx-sse==0.4.3
|
||||||
# via mcp
|
# via mcp
|
||||||
huggingface-hub==1.10.2
|
huggingface-hub==1.22.0
|
||||||
# via
|
# via
|
||||||
# accelerate
|
# accelerate
|
||||||
# datasets
|
# datasets
|
||||||
@@ -1244,7 +1245,6 @@ typer==0.24.1
|
|||||||
# fastapi-cli
|
# fastapi-cli
|
||||||
# fastapi-cloud-cli
|
# fastapi-cloud-cli
|
||||||
# fastsafetensors
|
# fastsafetensors
|
||||||
# huggingface-hub
|
|
||||||
# perceptron
|
# perceptron
|
||||||
# transformers
|
# transformers
|
||||||
typing-extensions==4.15.0
|
typing-extensions==4.15.0
|
||||||
|
|||||||
@@ -83,8 +83,9 @@ charset-normalizer==3.4.6
|
|||||||
# via requests
|
# via requests
|
||||||
chz==0.4.0
|
chz==0.4.0
|
||||||
# via gpt-oss
|
# via gpt-oss
|
||||||
click==8.3.1
|
click==8.4.2
|
||||||
# via
|
# via
|
||||||
|
# huggingface-hub
|
||||||
# jiwer
|
# jiwer
|
||||||
# nltk
|
# nltk
|
||||||
# rich-toolkit
|
# rich-toolkit
|
||||||
@@ -206,7 +207,7 @@ h11==0.16.0
|
|||||||
# uvicorn
|
# uvicorn
|
||||||
harfile==0.4.0
|
harfile==0.4.0
|
||||||
# via schemathesis
|
# via schemathesis
|
||||||
hf-xet==1.4.3
|
hf-xet==1.5.1
|
||||||
# via huggingface-hub
|
# via huggingface-hub
|
||||||
html2text==2025.4.15
|
html2text==2025.4.15
|
||||||
# via gpt-oss
|
# via gpt-oss
|
||||||
@@ -227,7 +228,7 @@ httpx==0.28.1
|
|||||||
# schemathesis
|
# schemathesis
|
||||||
httpx-sse==0.4.3
|
httpx-sse==0.4.3
|
||||||
# via mcp
|
# via mcp
|
||||||
huggingface-hub==1.10.2
|
huggingface-hub==1.22.0
|
||||||
# via
|
# via
|
||||||
# accelerate
|
# accelerate
|
||||||
# datasets
|
# datasets
|
||||||
@@ -959,7 +960,6 @@ typer==0.24.1
|
|||||||
# via
|
# via
|
||||||
# fastapi-cli
|
# fastapi-cli
|
||||||
# fastapi-cloud-cli
|
# fastapi-cloud-cli
|
||||||
# huggingface-hub
|
|
||||||
# transformers
|
# transformers
|
||||||
typing-extensions==4.15.0
|
typing-extensions==4.15.0
|
||||||
# via
|
# via
|
||||||
|
|||||||
@@ -21,7 +21,7 @@ def _patch_hf_api(side_effect):
|
|||||||
|
|
||||||
@pytest.fixture(scope="session")
|
@pytest.fixture(scope="session")
|
||||||
def hf_tokenizer() -> PreTrainedTokenizerBase:
|
def hf_tokenizer() -> PreTrainedTokenizerBase:
|
||||||
return AutoTokenizer.from_pretrained("gpt2")
|
return AutoTokenizer.from_pretrained("openai-community/gpt2")
|
||||||
|
|
||||||
|
|
||||||
_FAKE_ROWS = {
|
_FAKE_ROWS = {
|
||||||
|
|||||||
@@ -12,7 +12,7 @@ from vllm.benchmarks.datasets import get_samples
|
|||||||
|
|
||||||
@pytest.fixture(scope="session")
|
@pytest.fixture(scope="session")
|
||||||
def hf_tokenizer() -> PreTrainedTokenizerBase:
|
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:
|
def _write_jsonl(path: Path, n_rows: int) -> None:
|
||||||
|
|||||||
@@ -17,7 +17,7 @@ from vllm.benchmarks.datasets import (
|
|||||||
@pytest.fixture(scope="session")
|
@pytest.fixture(scope="session")
|
||||||
def hf_tokenizer() -> PreTrainedTokenizerBase:
|
def hf_tokenizer() -> PreTrainedTokenizerBase:
|
||||||
# Use a small, commonly available tokenizer
|
# Use a small, commonly available tokenizer
|
||||||
return AutoTokenizer.from_pretrained("gpt2")
|
return AutoTokenizer.from_pretrained("openai-community/gpt2")
|
||||||
|
|
||||||
|
|
||||||
class Params(NamedTuple):
|
class Params(NamedTuple):
|
||||||
|
|||||||
@@ -16,7 +16,7 @@ from vllm.benchmarks.datasets import RandomMultiModalDataset, SampleRequest
|
|||||||
@pytest.fixture(scope="session")
|
@pytest.fixture(scope="session")
|
||||||
def hf_tokenizer() -> PreTrainedTokenizerBase:
|
def hf_tokenizer() -> PreTrainedTokenizerBase:
|
||||||
"""Use a small, commonly available tokenizer."""
|
"""Use a small, commonly available tokenizer."""
|
||||||
return AutoTokenizer.from_pretrained("gpt2")
|
return AutoTokenizer.from_pretrained("openai-community/gpt2")
|
||||||
|
|
||||||
|
|
||||||
@pytest.fixture
|
@pytest.fixture
|
||||||
|
|||||||
@@ -13,7 +13,7 @@ from vllm.benchmarks.datasets.create_txt_slices_dataset import create_txt_slices
|
|||||||
@pytest.fixture(scope="session")
|
@pytest.fixture(scope="session")
|
||||||
def hf_tokenizer() -> PreTrainedTokenizerBase:
|
def hf_tokenizer() -> PreTrainedTokenizerBase:
|
||||||
# Use a small, commonly available tokenizer
|
# Use a small, commonly available tokenizer
|
||||||
return AutoTokenizer.from_pretrained("gpt2")
|
return AutoTokenizer.from_pretrained("openai-community/gpt2")
|
||||||
|
|
||||||
|
|
||||||
text_content = """
|
text_content = """
|
||||||
@@ -39,7 +39,7 @@ def test_create_txt_slices_jsonl(
|
|||||||
create_txt_slices_jsonl(
|
create_txt_slices_jsonl(
|
||||||
input_path=str(txt_path),
|
input_path=str(txt_path),
|
||||||
output_path=str(jsonl_path),
|
output_path=str(jsonl_path),
|
||||||
tokenizer_name="gpt2",
|
tokenizer_name="openai-community/gpt2",
|
||||||
num_prompts=10,
|
num_prompts=10,
|
||||||
input_len=10,
|
input_len=10,
|
||||||
output_len=10,
|
output_len=10,
|
||||||
|
|||||||
@@ -502,7 +502,7 @@ def test_gpt2_cache_hit(monkeypatch: pytest.MonkeyPatch):
|
|||||||
m.setenv("VLLM_USE_AOT_COMPILE", "1")
|
m.setenv("VLLM_USE_AOT_COMPILE", "1")
|
||||||
# First compilation - initialize model and generate
|
# First compilation - initialize model and generate
|
||||||
llm_model = LLM(
|
llm_model = LLM(
|
||||||
model="gpt2",
|
model="openai-community/gpt2",
|
||||||
compilation_config=CompilationConfig(
|
compilation_config=CompilationConfig(
|
||||||
mode=CompilationMode.VLLM_COMPILE,
|
mode=CompilationMode.VLLM_COMPILE,
|
||||||
),
|
),
|
||||||
@@ -519,7 +519,7 @@ def test_gpt2_cache_hit(monkeypatch: pytest.MonkeyPatch):
|
|||||||
# Second compilation - should hit cache
|
# Second compilation - should hit cache
|
||||||
m.setenv("VLLM_FORCE_AOT_LOAD", "1")
|
m.setenv("VLLM_FORCE_AOT_LOAD", "1")
|
||||||
llm_model = LLM(
|
llm_model = LLM(
|
||||||
model="gpt2",
|
model="openai-community/gpt2",
|
||||||
compilation_config=CompilationConfig(
|
compilation_config=CompilationConfig(
|
||||||
mode=CompilationMode.VLLM_COMPILE,
|
mode=CompilationMode.VLLM_COMPILE,
|
||||||
),
|
),
|
||||||
|
|||||||
@@ -24,7 +24,7 @@ from vllm.utils.torch_utils import is_torch_equal_or_newer
|
|||||||
def get_test_models():
|
def get_test_models():
|
||||||
"""Get list of models to test based on PyTorch version"""
|
"""Get list of models to test based on PyTorch version"""
|
||||||
models = [
|
models = [
|
||||||
"gpt2",
|
"openai-community/gpt2",
|
||||||
"Qwen/Qwen2-7B-Instruct",
|
"Qwen/Qwen2-7B-Instruct",
|
||||||
"meta-llama/Llama-3.1-8B",
|
"meta-llama/Llama-3.1-8B",
|
||||||
]
|
]
|
||||||
|
|||||||
@@ -114,7 +114,7 @@ TEXT_GENERATION_MODELS = {
|
|||||||
"tiiuae/falcon-7b": PPTestSettings.fast(),
|
"tiiuae/falcon-7b": PPTestSettings.fast(),
|
||||||
"google/gemma-1.1-2b-it": PPTestSettings.fast(),
|
"google/gemma-1.1-2b-it": PPTestSettings.fast(),
|
||||||
"google/gemma-2-9b": PPTestSettings.fast(),
|
"google/gemma-2-9b": PPTestSettings.fast(),
|
||||||
"gpt2": PPTestSettings.fast(),
|
"openai-community/gpt2": PPTestSettings.fast(),
|
||||||
"EleutherAI/gpt-j-6b": PPTestSettings.fast(),
|
"EleutherAI/gpt-j-6b": PPTestSettings.fast(),
|
||||||
"EleutherAI/pythia-1.4b": PPTestSettings.fast(),
|
"EleutherAI/pythia-1.4b": PPTestSettings.fast(),
|
||||||
"ibm/PowerLM-3b": PPTestSettings.fast(),
|
"ibm/PowerLM-3b": PPTestSettings.fast(),
|
||||||
|
|||||||
@@ -18,7 +18,7 @@ from vllm.renderers.embed_utils import safe_load_prompt_embeds
|
|||||||
|
|
||||||
@pytest.mark.asyncio
|
@pytest.mark.asyncio
|
||||||
async def test_empty_prompt():
|
async def test_empty_prompt():
|
||||||
model_name = "gpt2"
|
model_name = "openai-community/gpt2"
|
||||||
server_args = ["--enforce-eager"]
|
server_args = ["--enforce-eager"]
|
||||||
with RemoteOpenAIServer(model_name, server_args) as remote_server:
|
with RemoteOpenAIServer(model_name, server_args) as remote_server:
|
||||||
client = remote_server.get_async_client()
|
client = remote_server.get_async_client()
|
||||||
@@ -38,7 +38,7 @@ async def test_empty_prompt():
|
|||||||
|
|
||||||
@pytest.mark.asyncio
|
@pytest.mark.asyncio
|
||||||
async def test_out_of_vocab_token_ids():
|
async def test_out_of_vocab_token_ids():
|
||||||
model_name = "gpt2"
|
model_name = "openai-community/gpt2"
|
||||||
server_args = ["--enforce-eager"]
|
server_args = ["--enforce-eager"]
|
||||||
with RemoteOpenAIServer(model_name, server_args) as remote_server:
|
with RemoteOpenAIServer(model_name, server_args) as remote_server:
|
||||||
client = remote_server.get_async_client()
|
client = remote_server.get_async_client()
|
||||||
|
|||||||
@@ -115,6 +115,7 @@ def test_chatglm3_lora_tp4_fully_sharded_loras(chatglm3_lora_files):
|
|||||||
enable_lora=True,
|
enable_lora=True,
|
||||||
max_loras=2,
|
max_loras=2,
|
||||||
max_lora_rank=64,
|
max_lora_rank=64,
|
||||||
|
max_num_seqs=16,
|
||||||
tensor_parallel_size=4,
|
tensor_parallel_size=4,
|
||||||
trust_remote_code=True,
|
trust_remote_code=True,
|
||||||
fully_sharded_loras=True,
|
fully_sharded_loras=True,
|
||||||
|
|||||||
@@ -296,7 +296,7 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
|
|||||||
"GlmMoeDsaForCausalLM": _HfExamplesInfo(
|
"GlmMoeDsaForCausalLM": _HfExamplesInfo(
|
||||||
"zai-org/GLM-5", min_transformers_version="5.0.1", is_available_online=False
|
"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(
|
"GPTBigCodeForCausalLM": _HfExamplesInfo(
|
||||||
"bigcode/starcoder",
|
"bigcode/starcoder",
|
||||||
extras={
|
extras={
|
||||||
|
|||||||
@@ -40,7 +40,7 @@ from vllm.renderers.hf import (
|
|||||||
# Qwen2TokenizerFast (SentencePiece BPE variant)
|
# Qwen2TokenizerFast (SentencePiece BPE variant)
|
||||||
# BertTokenizerFast (WordPiece)
|
# BertTokenizerFast (WordPiece)
|
||||||
TOKENIZER_IDS: Final[list[str]] = [
|
TOKENIZER_IDS: Final[list[str]] = [
|
||||||
"gpt2",
|
"openai-community/gpt2",
|
||||||
"Qwen/Qwen2.5-1.5B-Instruct",
|
"Qwen/Qwen2.5-1.5B-Instruct",
|
||||||
"bert-base-uncased",
|
"bert-base-uncased",
|
||||||
]
|
]
|
||||||
|
|||||||
@@ -23,7 +23,7 @@ def _assert_tokenizer_like(tokenizer: object):
|
|||||||
|
|
||||||
|
|
||||||
def test_tokenizer_like_protocol():
|
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 isinstance(tokenizer, PreTrainedTokenizerFast)
|
||||||
_assert_tokenizer_like(tokenizer)
|
_assert_tokenizer_like(tokenizer)
|
||||||
|
|
||||||
@@ -43,7 +43,9 @@ def test_tokenizer_like_protocol():
|
|||||||
_assert_tokenizer_like(tokenizer)
|
_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):
|
def test_tokenizer_revision(tokenizer_name: str):
|
||||||
# Assume that "main" branch always exists
|
# Assume that "main" branch always exists
|
||||||
tokenizer = get_tokenizer(tokenizer_name, revision="main")
|
tokenizer = get_tokenizer(tokenizer_name, revision="main")
|
||||||
|
|||||||
@@ -33,7 +33,7 @@ TRUTH = [
|
|||||||
|
|
||||||
TOKENIZERS = [
|
TOKENIZERS = [
|
||||||
"facebook/opt-125m",
|
"facebook/opt-125m",
|
||||||
"gpt2",
|
"openai-community/gpt2",
|
||||||
"bigcode/tiny_starcoder_py",
|
"bigcode/tiny_starcoder_py",
|
||||||
"EleutherAI/gpt-j-6b",
|
"EleutherAI/gpt-j-6b",
|
||||||
"EleutherAI/pythia-70m",
|
"EleutherAI/pythia-70m",
|
||||||
|
|||||||
@@ -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):
|
def test_cached_tokenizer(model_id: str):
|
||||||
reference_tokenizer = AutoTokenizer.from_pretrained(
|
reference_tokenizer = AutoTokenizer.from_pretrained(
|
||||||
model_id, trust_remote_code=True
|
model_id, trust_remote_code=True
|
||||||
@@ -47,7 +47,7 @@ def _check_consistency(target: TokenizerLike, expected: TokenizerLike):
|
|||||||
assert target.encode("prompt") == expected.encode("prompt")
|
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):
|
def test_thread_pool_tokenizer_pickle(model_id: str):
|
||||||
"""Regression test for issue #45433: the thread-pool tokenizer wrapper
|
"""Regression test for issue #45433: the thread-pool tokenizer wrapper
|
||||||
reconstructs through maybe_make_thread_pool on unpickling, which used to
|
reconstructs through maybe_make_thread_pool on unpickling, which used to
|
||||||
|
|||||||
@@ -9,4 +9,4 @@ from vllm.tokenizers import TokenizerLike
|
|||||||
|
|
||||||
@pytest.fixture(scope="module")
|
@pytest.fixture(scope="module")
|
||||||
def default_tokenizer() -> TokenizerLike:
|
def default_tokenizer() -> TokenizerLike:
|
||||||
return AutoTokenizer.from_pretrained("gpt2")
|
return AutoTokenizer.from_pretrained("openai-community/gpt2")
|
||||||
|
|||||||
@@ -19,7 +19,7 @@ from vllm.tool_parsers import ToolParser, ToolParserManager
|
|||||||
def default_tokenizer() -> TokenizerLike:
|
def default_tokenizer() -> TokenizerLike:
|
||||||
"""Override module-scoped default_tokenizer because gigachat tests
|
"""Override module-scoped default_tokenizer because gigachat tests
|
||||||
mutate the tokenizer via ``add_tokens``."""
|
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"
|
MSG_SEP_TOKEN = "<|message_sep|>\n\n"
|
||||||
|
|||||||
@@ -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_guidance import GuidanceBackend
|
||||||
from vllm.v1.structured_output.backend_types import StructuredOutputOptions
|
from vllm.v1.structured_output.backend_types import StructuredOutputOptions
|
||||||
|
|
||||||
TOKENIZER = "gpt2"
|
TOKENIZER = "openai-community/gpt2"
|
||||||
|
|
||||||
|
|
||||||
@pytest.fixture(scope="module")
|
@pytest.fixture(scope="module")
|
||||||
|
|||||||
@@ -229,10 +229,11 @@ def get_model_path(model: str | Path, revision: str | None = None):
|
|||||||
if os.path.exists(model):
|
if os.path.exists(model):
|
||||||
return model
|
return model
|
||||||
assert huggingface_hub.constants.HF_HUB_OFFLINE
|
assert huggingface_hub.constants.HF_HUB_OFFLINE
|
||||||
common_kwargs = {
|
common_kwargs = dict(
|
||||||
"local_files_only": huggingface_hub.constants.HF_HUB_OFFLINE,
|
local_files_only=huggingface_hub.constants.HF_HUB_OFFLINE,
|
||||||
"revision": revision,
|
ignore_patterns="*",
|
||||||
}
|
revision=revision,
|
||||||
|
)
|
||||||
|
|
||||||
if envs.VLLM_USE_MODELSCOPE:
|
if envs.VLLM_USE_MODELSCOPE:
|
||||||
from modelscope.hub.snapshot_download import snapshot_download
|
from modelscope.hub.snapshot_download import snapshot_download
|
||||||
|
|||||||
Reference in New Issue
Block a user