The largest collection of PyTorch image encoders / backbones. Including train, eval, inference, export scripts, and pretrained weights -- ResNet, ResNeXT, EfficientNet, NFNet, Vision Transformer (ViT)...
Implementation of modern data augmentation techniques in TensorFlow 2.x to be used in your training pipeline.
Pytorch Implementation and Performance Analysis of the Popular Vision Architectures from Scratch.