[fbsync] Fix bug by checking if norm_layer weight is None before init (#6082)

Reviewed By: NicolasHug

Differential Revision: D36760938

fbshipit-source-id: 366bbd3fa225431dd5a0871019bb0e70d348ee7c

Author: YosuaMichael

torchvision/models/resnet.py

@@ -217,9 +217,9 @@ def __init__(
         # This improves the model by 0.2~0.3% according to https://arxiv.org/abs/1706.02677
         if zero_init_residual:
             for m in self.modules():
-                if isinstance(m, Bottleneck):
+                if isinstance(m, Bottleneck) and m.bn3.weight is not None:
                     nn.init.constant_(m.bn3.weight, 0)  # type: ignore[arg-type]
-                elif isinstance(m, BasicBlock):
+                elif isinstance(m, BasicBlock) and m.bn2.weight is not None:
                     nn.init.constant_(m.bn2.weight, 0)  # type: ignore[arg-type]
 
     def _make_layer(