Thanks for open-sourcing this interesting work!
Although the pre-training script is released, it doesn't seem to work properly because the dataset code is still missing (https://github.com/FoundationVision/Liquid/blob/main/liquid/train/train.py#L30). I am wondering when will these details be released?
Thanks for open-sourcing this interesting work!
Although the pre-training script is released, it doesn't seem to work properly because the dataset code is still missing (https://github.com/FoundationVision/Liquid/blob/main/liquid/train/train.py#L30). I am wondering when will these details be released?