How have pretrained models from modelzoo been trained?

vision
1

Are there any information about the training process (which part of ImageNet, which transformations, etc.) of the models in PyTorch modelzoo?

Thanks a lot.