From a892b259b4503526511d9ec17d5f8961e46969a6 Mon Sep 17 00:00:00 2001 From: Wentao Ye <44945378+yewentao256@users.noreply.github.com> Date: Thu, 11 Sep 2025 15:25:47 -0400 Subject: [PATCH] [Doc] Remove Useless Comments (#24687) Signed-off-by: yewentao256 --- vllm/model_executor/layers/quantization/fp8.py | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/vllm/model_executor/layers/quantization/fp8.py b/vllm/model_executor/layers/quantization/fp8.py index 65e0b7062153..d2616da84a00 100644 --- a/vllm/model_executor/layers/quantization/fp8.py +++ b/vllm/model_executor/layers/quantization/fp8.py @@ -757,10 +757,9 @@ class Fp8MoEMethod(FusedMoEMethodBase): layer.w2_weight = torch.nn.Parameter(shuffled_w2, requires_grad=False) - # DeepGemm scales need to be transposed and aligned. We try to do + # DeepGemm scales need to be transposed and aligned. We try to do # it ahead of time for performance reasons. if self.allow_deep_gemm and not is_deep_gemm_e8m0_used(): - # Lazy import to avoid CUDA initialization problems. if _is_col_major(layer.w13_weight_scale_inv): layer.w13_weight_scale_inv = \ get_col_major_tma_aligned_tensor(layer.w13_weight_scale_inv).contiguous()