This repository provides Keras implementations of numerous vision transformer architectures including BEiT, ConvNeXt, EfficientNet, Swin, DINO, and others. It also includes implementations of object detection models (YOLO, EfficientDet), diffusion models (DDPM, Stable Diffusion), CLIP training, and segmentation models. Supports both TensorFlow and PyTorch backends with training pipelines for ImageNet, COCO, and other datasets.