Ajinkya Tejankar

I think a simple solution would be to use the pre-trained MoCo model and get the loss for Imagenet. That will give you the final loss achieved by the model.

Looks great! A minor suggestion (feel free to ignore it): the high contrast use of white is somewhat cumbersome to read. Specifically, the code section also has a small font...

I didn't understand what you mean by debug them, but here is the command to run eval_linear.py: ``` python eval_linear.py \ -j 16 \ -b 256 \ --arch resnet50 \...

Hi @18456432930, Sorry, I didn't understand the problem. Can you be more specific? For instance, could you reproduce the numbers in our paper with our models? What augmentation does your...

I see. Our linear evaluation code uses a trick of normalizing the features by subtracting mean and std over the entire dataset. The mean and std are calculated on the...

Hi Chen, Unfortunately, no. We believe CIFAR is too small to be helpful for comparing SSL methods. You could use the ImageNet-100 subset (~125k) from [Contrastive Multiview Coding](https://github.com/HobbitLong/CMC/blob/master/imagenet100.txt). I have...

It seems the linter is failing on multiple pull requests not just mine. Do you know if it is the linter or I need to fix my pull request?

Sure, we can integrate this later. Also, thanks for the help with linter. For now, I will update the pull request after running `black`.