diff --git a/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/patches/sglang-kimi-image-sm120-flashinfer-mxfp4.patch b/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/patches/sglang-kimi-image-sm120-flashinfer-mxfp4.patch index a7adada..c01a093 100644 --- a/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/patches/sglang-kimi-image-sm120-flashinfer-mxfp4.patch +++ b/experiments/pro6000/kimi3_pro6000_sglang_tp32ep32_moe_backend_prefill/patches/sglang-kimi-image-sm120-flashinfer-mxfp4.patch @@ -36,7 +36,12 @@ x = dispatch_output.hidden_states topk_output = dispatch_output.topk_output -@@ -384,7 +388,7 @@ +@@ -343,3 +347,4 @@ ++ x = x.contiguous() + x, input_sf = mxfp8_quantize( + x, + is_sf_swizzled_layout=True, +@@ -384,7 +389,7 @@ ep_rank=quant_info.moe_ep_rank, use_w4_group_scaling=not use_mxfp8_act_scaling, use_mxfp8_act_scaling=use_mxfp8_act_scaling,