[sgl-kernel][1/2] Fused qk_norm_rope for Qwen3-MoE (#14036)

Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
2025-11-28 12:25:15 +08:00
parent bce40fa217
commit e12c78aab6
7 changed files with 696 additions and 0 deletions
--- a/sgl-kernel/python/sgl_kernel/init.py
+++ b/sgl-kernel/python/sgl_kernel/init.py
@@ -84,6 +84,7 @@ from sgl_kernel.moe import (
    apply_shuffle_mul_sum,
    cutlass_fp4_group_mm,
    fp8_blockwise_scaled_grouped_mm,
+    fused_qk_norm_rope,
    kimi_k2_moe_fused_gate,
    moe_align_block_size,
    moe_fused_gate,