minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[CI] Pin the rust frontend parity test to eager prefill (#34146)

Browse Source
This commit is contained in:
Liangsheng Yin
2026-08-08 20:07:19 -07:00
committed by GitHub
parent f6cbdc1dd1
commit 4792ab1e90
1 changed files with 7 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/registered/openai_server/basic/test_openai_completion_rust.py
+7 -1
View File
@@ -26,13 +26,19 @@ class TestOpenAICompletionRustParity(CustomTestCase):
api_key = "sk-123456" api_key = "sk-123456"
def _get_logprobs(self, *, rust_frontend): def _get_logprobs(self, *, rust_frontend):
# Prefill CUDA graph pads the batch, so numerics follow whichever
# requests share the forward pass; the assertions below need equality.
process = popen_launch_server( process = popen_launch_server(
self.model, self.model,
DEFAULT_URL_FOR_TEST, DEFAULT_URL_FOR_TEST,
timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH, timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
api_key=self.api_key, api_key=self.api_key,
env={"SGLANG_RUST_SERVER": "1" if rust_frontend else "0"}, env={"SGLANG_RUST_SERVER": "1" if rust_frontend else "0"},
other_args=["--random-seed", "42"], other_args=[
"--random-seed",
"42",
"--disable-prefill-cuda-graph",
],
) )
try: try:
response = requests.post( response = requests.post(