Settings used for ImageNet pre-training

[MucManu]

Summary
This is more a general question that applies to all models where ImageNet pre-trained weights are available.

Is there any source or documentation what settings were being used for pre-training the particular weights that are being provided, e.g. when using the pre-trained MobileNetV2 weights is there any source where I can get some information about the settings for pre-training?
When asking for settings I refer to parameters like number of epochs, number of classes, decay params, used augmentations, size of the pretraining dataset etc.

Would be happy if somebody has more information since my web search didn't yield useful results!