biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[ci] Use env var to control whether to use S3 bucket in CI (#13634)

Browse Source
This commit is contained in:
Kevin H. Luu
2025-02-22 19:19:45 -08:00
committed by GitHub
parent 322d2a27d6
commit 2c5e637b57
30 changed files with 222 additions and 231 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
tests/engine/test_executor.py
View File

@@ -6,17 +6,12 @@ from typing import Any, Callable, Dict, List, Optional, Tuple, Union
import pytest
from vllm.config import LoadFormat
from vllm.engine.arg_utils import AsyncEngineArgs, EngineArgs
from vllm.engine.async_llm_engine import AsyncLLMEngine
from vllm.engine.llm_engine import LLMEngine
from vllm.executor.uniproc_executor import UniProcExecutor
from vllm.sampling_params import SamplingParams
from ..conftest import MODEL_WEIGHTS_S3_BUCKET
RUNAI_STREAMER_LOAD_FORMAT = LoadFormat.RUNAI_STREAMER
class Mock:
...
@@ -38,12 +33,10 @@ class CustomUniExecutor(UniProcExecutor):
CustomUniExecutorAsync = CustomUniExecutor
@pytest.mark.parametrize("model",
[f"{MODEL_WEIGHTS_S3_BUCKET}/distilbert/distilgpt2"])
@pytest.mark.parametrize("model", ["distilbert/distilgpt2"])
def test_custom_executor_type_checking(model):
with pytest.raises(ValueError):
engine_args = EngineArgs(model=model,
load_format=RUNAI_STREAMER_LOAD_FORMAT,
distributed_executor_backend=Mock)
LLMEngine.from_engine_args(engine_args)
with pytest.raises(ValueError):
@@ -52,8 +45,7 @@ def test_custom_executor_type_checking(model):
AsyncLLMEngine.from_engine_args(engine_args)
@pytest.mark.parametrize("model",
[f"{MODEL_WEIGHTS_S3_BUCKET}/distilbert/distilgpt2"])
@pytest.mark.parametrize("model", ["distilbert/distilgpt2"])
def test_custom_executor(model, tmp_path):
cwd = os.path.abspath(".")
os.chdir(tmp_path)
@@ -62,7 +54,6 @@ def test_custom_executor(model, tmp_path):
engine_args = EngineArgs(
model=model,
load_format=RUNAI_STREAMER_LOAD_FORMAT,
distributed_executor_backend=CustomUniExecutor,
enforce_eager=True, # reduce test time
)
@@ -77,8 +68,7 @@ def test_custom_executor(model, tmp_path):
os.chdir(cwd)
@pytest.mark.parametrize("model",
[f"{MODEL_WEIGHTS_S3_BUCKET}/distilbert/distilgpt2"])
@pytest.mark.parametrize("model", ["distilbert/distilgpt2"])
def test_custom_executor_async(model, tmp_path):
cwd = os.path.abspath(".")
os.chdir(tmp_path)
@@ -87,7 +77,6 @@ def test_custom_executor_async(model, tmp_path):
engine_args = AsyncEngineArgs(
model=model,
load_format=RUNAI_STREAMER_LOAD_FORMAT,
distributed_executor_backend=CustomUniExecutorAsync,
enforce_eager=True, # reduce test time
)
@@ -106,8 +95,7 @@ def test_custom_executor_async(model, tmp_path):
os.chdir(cwd)
@pytest.mark.parametrize("model",
[f"{MODEL_WEIGHTS_S3_BUCKET}/distilbert/distilgpt2"])
@pytest.mark.parametrize("model", ["distilbert/distilgpt2"])
def test_respect_ray(model):
# even for TP=1 and PP=1,
# if users specify ray, we should use ray.
@@ -116,7 +104,6 @@ def test_respect_ray(model):
engine_args = EngineArgs(
model=model,
distributed_executor_backend="ray",
load_format=RUNAI_STREAMER_LOAD_FORMAT,
enforce_eager=True, # reduce test time
)
engine = LLMEngine.from_engine_args(engine_args)