From 159376ff686ae8eb5cbc809decf121acd43063d1 Mon Sep 17 00:00:00 2001 From: RandNMR73 Date: Thu, 9 Apr 2026 05:41:19 +0000 Subject: [PATCH] fix --- fastvideo/attention/layer.py | 4 ---- fastvideo/layers/fp4linear.py | 2 +- 2 files changed, 1 insertion(+), 5 deletions(-) diff --git a/fastvideo/attention/layer.py b/fastvideo/attention/layer.py index b83d7019c..56a0911db 100644 --- a/fastvideo/attention/layer.py +++ b/fastvideo/attention/layer.py @@ -48,10 +48,6 @@ class DistributedAttention(nn.Module): **extra_impl_args) # Register attn_impl as submodule if it has learnable parameters (e.g., SLA's proj_l) # This ensures its parameters are included in state_dict() for saving/loading - if isinstance(self.attn_impl, nn.Module): - self.add_module('attn_impl', self.attn_impl) - # Register attn_impl as submodule if it has learnable parameters (e.g., SLA's proj_l) - # This ensures its parameters are included in state_dict() for saving/loading if isinstance(self.attn_impl, nn.Module): self.add_module('attn_impl', self.attn_impl) self.num_heads = num_heads diff --git a/fastvideo/layers/fp4linear.py b/fastvideo/layers/fp4linear.py index 133607deb..b70c3d15d 100644 --- a/fastvideo/layers/fp4linear.py +++ b/fastvideo/layers/fp4linear.py @@ -86,4 +86,4 @@ def fp4_linear_forward(self, x: torch.Tensor) -> tuple[torch.Tensor, torch.Tenso bias = self.bias if not self.skip_bias_add else None output = _LinearFWD4BWD16Fn.apply(x, self.weight, bias, "cutlass", 16, True) output_bias = self.bias if self.skip_bias_add else None - return output, output_bias + return output, output_bias \ No newline at end of file