model weights after Stage I

[fmu2]

Thanks for the great work! I am wondering if you are able to provide the encoder weights after Stage-I training. According to the paper, the encoder is finetuned in Stage II so that it adapts to corrupted images. However, the encoder weights before fine-tuning would make more sense for my use case. Thank you!