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

[NPU] Set DEEPEP_HYBRID_DEPLOYMENT for new DeepEP tests; switch glm5_2 to w8a8; tune nightly timeouts (#38775)

Browse Source
This commit is contained in:
pllimax
2026-09-10 23:32:54 +08:00
committed by GitHub
parent 12771786f2
commit 92dffebe16
7 changed files with 37 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/sglang/test/ascend/e2e/test_npu_performance_utils.py
+1
View File
@@ -166,6 +166,7 @@ GLM5_TOP64_PRUNED_GSM8K_MODEL_PATH = (
)
GLM_5_1_W4A8_MODEL_PATH = "/root/.cache/modelscope/hub/models/Eco-Tech/GLM-5.1-w4a8"
GLM_5_2_W4A8_MODEL_PATH = "/root/.cache/modelscope/hub/models/Eco-Tech/GLM-5.2-w4a8"
GLM_5_2_W8A8_MODEL_PATH = "/root/.cache/modelscope/hub/models/Eco-Tech/GLM-5.2-w8a8"
MINIMAX_M2_5_W8A8_MODEL_PATH = (
"/root/.cache/modelscope/hub/models/Eco-Tech/MiniMax-M2.5-w8a8-QuaRot"
)