Normalization for object detection

Migrated from discuss.pytorch.org. Requests were made by @mattans.

📚 Documentation

The reference implementations for classification, segmentation, and video classification all use a normalization transform. In contrast, object detection does not use any normalization.

1. Consider explaining why the pretrained detection models are the only ones that don’t require image normalization (I understand that the training set was not normalized. But again, why?)
2. Worth mentioning that no normalization is needed. The classification, segmentation and detection pretrained models are trained on ImageNet, so one may think all of them require ImageNet normalization, when in fact only the classification and segmentation models require normalization. Perhaps it’s best to put this info in a table, since the pretrained video models also have a normalization, but different.

Compare the reference implementations for classification, segmentation, video classification, and object detection, focusing on their normalization transforms. Update the documentation to explain why pretrained detection models need no image normalization and clearly state the normalization expectations, optionally using a comparison table. Done means the contrast with the other pretrained models is unambiguous.