Refactor fp8 nextn layer for DeepSeek nvfp4 checkpoint (#15353)

This commit is contained in:
Baizhou Zhang
2025-12-28 11:57:09 +08:00
committed by GitHub
parent 8e43980ebb
commit 656f4d69a1
6 changed files with 115 additions and 112 deletions
+1
View File
@@ -253,6 +253,7 @@ class Envs:
SGLANG_MOE_NVFP4_DISPATCH = EnvBool(False)
SGLANG_NVFP4_CKPT_FP8_GEMM_IN_ATTN = EnvBool(False)
SGLANG_PER_TOKEN_GROUP_QUANT_8BIT_V2 = EnvBool(False)
SGLANG_NVFP4_CKPT_FP8_NEXTN_MOE = EnvBool(False)
# Flashinfer
SGLANG_IS_FLASHINFER_AVAILABLE = EnvBool(True)