Disable flaky eagle tests (#5753)
This commit is contained in:
@@ -1,5 +1,4 @@
|
||||
import json
|
||||
import multiprocessing as mp
|
||||
import os
|
||||
import random
|
||||
import threading
|
||||
@@ -8,7 +7,6 @@ import unittest
|
||||
from concurrent.futures import ThreadPoolExecutor
|
||||
from functools import partial
|
||||
from types import SimpleNamespace
|
||||
from typing import List, Optional
|
||||
|
||||
import numpy as np
|
||||
import requests
|
||||
@@ -18,7 +16,6 @@ import sglang as sgl
|
||||
from sglang.srt.hf_transformers_utils import get_tokenizer
|
||||
from sglang.srt.utils import kill_process_tree
|
||||
from sglang.test.few_shot_gsm8k import run_eval
|
||||
from sglang.test.runners import DEFAULT_PROMPTS, SRTRunner
|
||||
from sglang.test.test_utils import (
|
||||
DEFAULT_EAGLE_DRAFT_MODEL_FOR_TEST,
|
||||
DEFAULT_EAGLE_TARGET_MODEL_FOR_TEST,
|
||||
@@ -541,36 +538,5 @@ class TestEAGLEServerTriton(TestEAGLEServer):
|
||||
)
|
||||
|
||||
|
||||
class TestEAGLEServerPageSize(TestEAGLEServer):
|
||||
@classmethod
|
||||
def setUpClass(cls):
|
||||
cls.base_url = DEFAULT_URL_FOR_TEST
|
||||
cls.process = popen_launch_server(
|
||||
DEFAULT_EAGLE_TARGET_MODEL_FOR_TEST,
|
||||
cls.base_url,
|
||||
timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
|
||||
other_args=[
|
||||
"--speculative-algorithm",
|
||||
"EAGLE",
|
||||
"--speculative-draft-model-path",
|
||||
DEFAULT_EAGLE_DRAFT_MODEL_FOR_TEST,
|
||||
"--speculative-num-steps",
|
||||
5,
|
||||
"--speculative-eagle-topk",
|
||||
1,
|
||||
"--speculative-num-draft-tokens",
|
||||
6,
|
||||
"--mem-fraction-static",
|
||||
0.7,
|
||||
"--chunked-prefill-size",
|
||||
128,
|
||||
"--max-running-requests",
|
||||
8,
|
||||
"--page-size",
|
||||
8,
|
||||
],
|
||||
)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
unittest.main()
|
||||
|
||||
Reference in New Issue
Block a user