Parameter understanding of quad dataloader in yolov5

The original question github Address ：what is the new quad dataloader? · Issue #1898 · ultralytics/yolov5 · GitHub

Usage mode ：python train.py --quad

effect ：This quad-collate function will reshapes a batch from 16x3x640x640 to 4x3x1280x1280, which does not have much effect by itself as it is only rearranging the mosaics in the batch, but which interestingly allows for 2x upscaling of some images within the batch (one of the 4 mosaics in each quad is upscaled by 2x, the other 3 mosaics are deleted).

It is to transfer the information on the channel to the image scale information

It can be seen that this method is --img 640 and --img 1280 A compromise of , In special project scenarios, both speed and precision are considered .