PyTorch-BYOL
PyTorch-BYOL copied to clipboard
sthalles
PyTorch implementation of Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning
This line of code self.checkpoint_interval = params['checkpoint_interval'] had no impact on the code, therefore I think it's better to allow the code to be clean by eliminating the unused variables.
Hello, Thank you very much for the program you shared, but I want to test it on the CIFAR-10 data set. How should the relevant parameters be set?
I am not sure, but I think the loss function is a bit different from the original paper [https://github.com/sthalles/PyTorch-BYOL/blob/f007ed6b3619a3ae00eb9c95cde342eac5b98923/trainer.py#L35](here). In the paper the loss function equation is for unnormalized tensors(loss...
Hi this is really good application. I would like to have your advice to use my own dataset. I have images that are very different with widely known dataset such...
Hello, Thank you for this excellent repository! Do you have any suggestions of changes to make to train BYOL on the CIFAR10 dataset? The way I am doing this (in...
Hi, thanks for your share. But I have a question about your STL10 top-1 accuracy. Is the accuracy '70.1' fine-tuning from your pre-trained model 'resnet-18_40-epochs' or training from scratch for...
First of all, thank you so much for making this work open source! My question isn't directly related to your implementation but rather a question about the paper -- I...
Thanks a ton for making this open-source. I have a question about replicating the 75% accuracy for 80 epochs. I just changed the config file to read "max_epochs: 80" and...
