@stefanopini In the code of the file Coco.py under datasets folder, the data augmentations of training are done to the original image and no additional data is generated, hence the total number of samples of training remain unchanged even if data augmentations is applied. Can you give clarity on this issue.
@stefanopini In the code of the file Coco.py under datasets folder, the data augmentations of training are done to the original image and no additional data is generated, hence the total number of samples of training remain unchanged even if data augmentations is applied. Can you give clarity on this issue.