EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

Fix flasky tests (#2212)

Browse Source
This commit is contained in:
Lianmin Zheng
2024-11-26 23:06:20 -08:00
committed by GitHub
parent 0b46b951ae
commit c754652fcd
2 changed files with 2 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
python/sglang/srt/managers/detokenizer_manager.py
View File

@@ -25,8 +25,6 @@ from sglang.srt.managers.io_struct import (
BatchEmbeddingOut, BatchEmbeddingOut,
BatchStrOut, BatchStrOut,
BatchTokenIDOut, BatchTokenIDOut,
GetMemPoolSizeReqOutput,
UpdateWeightReqOutput,
) )
from sglang.srt.managers.schedule_batch import FINISH_MATCHED_STR, FINISH_MATCHED_TOKEN from sglang.srt.managers.schedule_batch import FINISH_MATCHED_STR, FINISH_MATCHED_TOKEN
from sglang.srt.server_args import PortArgs, ServerArgs from sglang.srt.server_args import PortArgs, ServerArgs

4
test/srt/test_pytorch_sampling_backend.py
View File

@@ -51,7 +51,7 @@ class TestPyTorchSamplingBackend(unittest.TestCase):
response_single = requests.post( response_single = requests.post(
self.base_url + "/generate", self.base_url + "/generate",
json={ json={
"text": "The capital of France is", "text": "The capital of Germany is",
"sampling_params": { "sampling_params": {
"temperature": 0, "temperature": 0,
"max_new_tokens": 32, "max_new_tokens": 32,
@@ -69,7 +69,7 @@ class TestPyTorchSamplingBackend(unittest.TestCase):
response_batch = requests.post( response_batch = requests.post(
self.base_url + "/generate", self.base_url + "/generate",
json={ json={
"text": ["The capital of France is"] * 10, "text": ["The capital of Germany is"] * 10,
"sampling_params": { "sampling_params": {
"temperature": 0, "temperature": 0,
"max_new_tokens": 32, "max_new_tokens": 32,