UNetFormer: A UNet-like transformer for efficient semantic segmentation of remote sensing urban scene imagery, ISPRS. Also, including other vision transformers and CNNs for satellite, aerial image and UAV image segmentation.

Adan: Adaptive Nesterov Momentum Algorithm for Faster Optimizing Deep Models

Extract video features from raw videos using multiple GPUs. We support RAFT flow frames as well as S3D, I3D, R(2+1)D, VGGish, CLIP, and TIMM models.

Modern human action classification system using MediaPipe pose estimation and PyTorch deep learning, achieving 47x faster inference than the original TensorFlow implementation. Supports 100+ model architectures for classifying 40 human actions with real-time performance suitable for autonomous vehicles and video surveillance.

Holds code for our CVPR'23 tutorial: All Things ViTs: Understanding and Interpreting Attention in Vision.

Simple Implementation of Pix2Seq model for object detection in PyTorch

Multi-label classification based on timm.

Encoder-Decoder Cell and Nuclei segmentation models

PyTorch Volume Models for 3D data

A PyTorch-based Python library with UNet architecture and multiple backbones for Image Semantic Segmentation.

💡💡💡awesome compute vision app in gradio

Jittor Image Models is a library for pulling together a wide variety of SOTA deep learning models in the Jittor framework.

A library that includes Keras3 layers, blocks and models with pretrained weights, providing support for transfer learning, feature extraction, and more.

A simple wrapper library for binding timm models as detectron2 backbones

Attribution methods that explain image classification models, implemented in PyTorch, and support batch inputs and GPU.

Multi-label classification based on timm, and add SimCLR to timm.

Simple image classification for custom dataset (pytorch-lightning, timm)

Supercharge Your PyTorch Image Models: Bag of Tricks to 8x Faster Inference with ONNX Runtime & Optimizations

Clean, reproducible, boilerplate-free deep learning project template.

Jax/Flax implementation of DeiT and DeiT-III (ViT)