isnet:imagelevel.py

[shujunyy123]

imagelevel.py : 47: feats_il = self.correlate_net(x, torch.cat([x_global, x], dim=1))

isanet.py： 47：context = super(SelfAttentionBlock, self).forward(x, x)

is there any problem？ bug？

[CharlesPikachu]

Sry, I cannot get you point. Could you describe your problem clearer?

[shujunyy123]

(x, torch.cat([x_global, x], dim=1)) != (x, x) In debug mode, there’s a error .

[CharlesPikachu]

./models/base/selfattention.py stands for the self-attention-like structure, not self-attention in a narrow sense.