Doubt about the cosine_angular_loss

[ireneMsm2020]

Hello,
I have doubts about cosine_angular_loss. After permute operation, the channel dim is in dim 3, but in normalize operation, you set dim=1. So is there a mistake? Can you please explain it ?

---

def masked_cosine_angular_loss(preds, target, mask_valid):
preds = (2 * preds - 1).clamp(-1, 1)
target = (2 * target - 1).clamp(-1, 1)
mask_valid = mask_valid[:,0,:,:].bool().squeeze(1)
preds = preds.permute(0,2,3,1)[mask_valid, :]
target = target.permute(0,2,3,1)[mask_valid, :]
preds_norm = torch.nn.functional.normalize(preds, p=2, dim=1)
target_norm = torch.nn.functional.normalize(target, p=2, dim=1)
loss = torch.mean(-torch.sum(preds_norm * target_norm, dim = 1))
return loss

[alexsax]

Hi your concern looks right to me, but the models are quite good. So I think the released code might just be the problem (and the evaluation code we borrowed right from the OASIS paper). @Ainaz99 what do you think?

[limacv]

isn't the mask operation flatten the two tensor into [N, 3] shape? In that case dim=1 should be correct.