Thomas Capelle

Thank you for this awesome work, Training the model works, with some tweaks. The frozen test set file is missing. I forked the repo and corrected this issues, but a...

AOI losses

1

How does radiance deals with AOI losses (IAM modifiers)? It comes down to material definitions?

How to load single channel (or 4 channel) images

1

I am trying to load a dataset consisting of 1 channel images, and don't know how to do it. I have modified the loading to open them as grayscale: ```python...

Great work !

Big Refactor

7

Hey, this code has made me play and have so much fun =) I refactored the code to do multiple things: - It uses mixed precision training (2x speed up)....

Hello, Congratz on the win. I am trying to inspire myself from this model to make predictions on sequence data. I am having a hard time getting to understand the...

I see a lot of jupyter notebooks! You may want to try [nbdev](https://github.com/fastai/nbdev). It is a delight to work with, and enable full CICD/ docs /deploy of jupyter to conda/pip....

Log the config params to wandb

2

This PR logs the parameters living on the global scope as a config to W&B. This is a greedy solution, as the parameters live on the global scope. I do...

Make wandb training logs public

1

Hello, again Andrej, Would you mind making the training logs public so we can follow your progress in reproducing GPT2? You can do this by clicking on the lock on...

Compile gives weird error about CUDA on MPS device (apple Macbook)

3

### 🐛 Describe the bug In PyTorch 2.0 on an Apple Macbook running latest OS. ```python from torch import nn from torchvision.models import resnet50 torch.set_default_device('mps') img = torch.rand(1,3,64,64) r50 =...