The description and the behavior don't match

[yongchanghao]

In the following code, the comment says weight A is init as usual (kaiming init as in other places in the code) and B as zeros. However, the behavior is quite the opposite:

LoRA/loralib/layers.py

Lines 58 to 60 in 3f5c193

	# initialize A the same way as the default for nn.Linear and B to zero
	nn.init.zeros_(self.lora_A)
	nn.init.normal_(self.lora_B)

Is this intended? Thanks.

[edwardjhu]

I fixed the comment. Thanks!

[yongchanghao]

Thank you. May I know the reason why the matrix A is set to 0 here, unlike other parts (and the paper) where B is 0?