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

[NPU] Remove temperature/top_p from Qwen3.5-397B-A17B perf test (#39047)

Browse Source
This commit is contained in:
pllimax
2026-09-14 21:48:50 +08:00
committed by GitHub
parent f81fbc749a
commit d5f1c593c1
1 changed files with 3 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/registered/npu/performance/qwen3_5_397b/test_npu_qwen3_5_397b_w4a8_8p_in3k5_out1k5_50ms.py
+3 -2
View File
@@ -125,8 +125,9 @@ class TestNPUQwen3_5_397B_A17B_3K5_1K5_50ms(TestNpuPerformanceTestCaseBase):
tpot = 50 tpot = 50
output_token_throughput = 5415 output_token_throughput = 5415
request_rate = float("inf") request_rate = float("inf")
temperature = 0.6 # Disabled to work around DTS2026091033366; re-enable after the fix.
top_p = 0.95 # temperature = 0.6
# top_p = 0.95
def test_npu_qwen3_5_397b_a17b_3k5_1k5(self): def test_npu_qwen3_5_397b_a17b_3k5_1k5(self):
"""Run NPU performance test for Qwen3.5-397B-A17B in3k5 out1k5""" """Run NPU performance test for Qwen3.5-397B-A17B in3k5 out1k5"""