mirror of
https://github.com/huggingface/pytorch-image-models.git
synced 2025-06-03 15:01:08 +08:00
add notes from author
This commit is contained in:
parent
899cf84ccc
commit
3a55a30ed1
@ -229,8 +229,7 @@ class ClassAttentionBlock(nn.Module):
|
||||
else:
|
||||
self.gamma1, self.gamma2 = 1.0, 1.0
|
||||
|
||||
# (note from official code)
|
||||
# FIXME: A hack for models pre-trained with layernorm over all the tokens not just the CLS
|
||||
# See https://github.com/rwightman/pytorch-image-models/pull/747#issuecomment-877795721
|
||||
self.tokens_norm = tokens_norm
|
||||
|
||||
def forward(self, x):
|
||||
@ -309,6 +308,7 @@ class XCABlock(nn.Module):
|
||||
def forward(self, x, H: int, W: int):
|
||||
x = x + self.drop_path(self.gamma1 * self.attn(self.norm1(x)))
|
||||
# NOTE official code has 3 then 2, so keeping it the same to be consistent with loaded weights
|
||||
# See https://github.com/rwightman/pytorch-image-models/pull/747#issuecomment-877795721
|
||||
x = x + self.drop_path(self.gamma3 * self.local_mp(self.norm3(x), H, W))
|
||||
x = x + self.drop_path(self.gamma2 * self.mlp(self.norm2(x)))
|
||||
return x
|
||||
|
Loading…
x
Reference in New Issue
Block a user