Jingzheng Li

In the SSL setting, the target data is not the same as test data, although they are in the same distribution. But in your paper, the target data is used...

Question about no folder of "ground truth" of SYNTHIA dataset

1

Sorry to bother you, when I downloaded the SYNTHIA dataset and decompressed it. I found there was no folder with "ground truth"(GT). I want to know if someone else is...

hi. this good work is difficult for me, especially the part of "Edge prior model". Could you provide a detail version of method in Chinese such as published in Zhihu....

this amazing performance improvments kill this area!!!!!!!!!!

1

this amazing performance improvements kill this area!!!!!!!!!!

Great work is not accepted~

1

I didn't see this work in the Neurips acceptance list. I can't believe that such a great job was not accepted. I think it is the reviewers who misunderstand the...

How to extend to base-to-novel classes task?

4

Hi, This method modifies the parameters of the text encoder, so it cannot extend to base-to-new classes tasks. I would like to know how to address this problem.

I think that the noisy label belongs to a kind of distribution shift (output shift). Will this library support detecting noisy labels? Thanks

某些生僻字不能正常显示，例如“犇”

2

某些中文的生僻字，显示不出来该怎么解决呢，例如“犇”

Hi, thanks for the nice code. I found the performance is poor when full fine-tuning the ResNet-based CLIP on ImageNet while for ViT-based CLIP the performance is good. Do you...

Hi, thanks for the nice code. How to implement Layer-wise learning rate decay on ResNet instead of ViT?