[NVIDIA] Fix cutedsl backend of MoE (#12353)

This commit is contained in:
Kaixi Hou
2025-11-04 18:54:55 -08:00
committed by GitHub
parent 09938e1f82
commit 0711d1509b
8 changed files with 579 additions and 28 deletions
+1 -1
View File
@@ -57,7 +57,7 @@ DEFAULT_MODEL_NAME_FOR_TEST_MLA = "lmsys/sglang-ci-dsv3-test"
DEFAULT_MODEL_NAME_FOR_TEST_MLA_NEXTN = "lmsys/sglang-ci-dsv3-test-NextN"
# NVFP4 models
DEFAULT_DEEPSEEK_NVFP4_MODEL_FOR_TEST = "nvidia/DeepSeek-R1-0528-FP4"
DEFAULT_DEEPSEEK_NVFP4_MODEL_FOR_TEST = "nvidia/DeepSeek-V3-0324-FP4"
# FP8 models
DEFAULT_MODEL_NAME_FOR_TEST_FP8 = "neuralmagic/Meta-Llama-3.1-8B-Instruct-FP8"