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

[Misc] Replace os environ to monkeypatch in test suite (#14516)

Browse Source 
Signed-off-by: sibi <85477603+t-sibiraj@users.noreply.github.com>
Signed-off-by: Aaron Pham <contact@aarnphm.xyz>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
Co-authored-by: Aaron Pham <contact@aarnphm.xyz>
This commit is contained in:
Sibi
2025-03-17 11:35:57 +08:00
committed by GitHub
parent 1e799b7ec1
commit a73e183e36
43 changed files with 1900 additions and 1658 deletions
Download Patch File Download Diff File Expand all files Collapse all files

38
tests/distributed/test_pp_cudagraph.py
View File

@@ -1,11 +1,15 @@
# SPDX-License-Identifier: Apache-2.0
from __future__ import annotations
import os
from typing import TYPE_CHECKING
import pytest
from ..utils import compare_two_settings, fork_new_process_for_each_test
if TYPE_CHECKING:
from typing_extensions import LiteralString
@pytest.mark.parametrize("PP_SIZE, MODEL_NAME", [
(2, "JackFram/llama-160m"),
@@ -15,18 +19,24 @@ from ..utils import compare_two_settings, fork_new_process_for_each_test
"FLASHINFER",
])
@fork_new_process_for_each_test
def test_pp_cudagraph(PP_SIZE, MODEL_NAME, ATTN_BACKEND):
cudagraph_args = [
# use half precision for speed and memory savings in CI environment
"--dtype",
"float16",
"--pipeline-parallel-size",
str(PP_SIZE),
"--distributed-executor-backend",
"mp",
]
os.environ["VLLM_ATTENTION_BACKEND"] = ATTN_BACKEND
def test_pp_cudagraph(
monkeypatch: pytest.MonkeyPatch,
PP_SIZE: int,
MODEL_NAME: str,
ATTN_BACKEND: LiteralString,
):
with monkeypatch.context() as m:
cudagraph_args = [
# use half precision for speed and memory savings in CI environment
"--dtype",
"float16",
"--pipeline-parallel-size",
str(PP_SIZE),
"--distributed-executor-backend",
"mp",
]
m.setenv("VLLM_ATTENTION_BACKEND", ATTN_BACKEND)
eager_args = cudagraph_args + ["--enforce-eager"]
eager_args = cudagraph_args + ["--enforce-eager"]
compare_two_settings(MODEL_NAME, eager_args, cudagraph_args)
compare_two_settings(MODEL_NAME, eager_args, cudagraph_args)