Add enable_flashinfer_mxfp4_bf16_moe for higher precision and slower moe backend (#9004)

2025-08-23 15:38:40 +08:00
parent 127d4b0d5e
commit 0374304a2c
3 changed files with 37 additions and 5 deletions
--- a/python/sglang/srt/managers/schedule_batch.py
+++ b/python/sglang/srt/managers/schedule_batch.py
@@ -87,6 +87,7 @@ GLOBAL_SERVER_ARGS_KEYS = [
    "disable_flashinfer_cutlass_moe_fp4_allgather",
    "disable_radix_cache",
    "enable_dp_lm_head",
+    "flashinfer_mxfp4_moe_precision",
    "enable_flashinfer_allreduce_fusion",
    "moe_dense_tp_size",
    "ep_dispatch_algorithm",