Stars
High quality training free inpaint for every stable diffusion model. Supports ComfyUI
[WACV2023] The official code implementation of paper "Pik-Fix: Restoring and Colorizing Old Photos".
A Deep Learning based project for colorizing and restoring old images (and video!)
[DEIMv2] Real Time Object Detection Meets DINOv3
CUDA Python: Performance meets Productivity
CUDA integration for Python, plus shiny features
One-Step Diffusion Transformer for Controllable Real-World Image Super-Resolution
Offical repo for "OMGSR: You Only Need One Mid-timestep Guidance for Real-World Image Super-Resolution"
Open-Sora: Democratizing Efficient Video Production for All
Code release for ConvNeXt V2 model
Tiny AutoEncoder for Hunyuan Video (and other video models)
A unified inference and post-training framework for accelerated video generation.
[AAAI 2026] Turbo-VAED: Fast and Stable Transfer of Video-VAEs to Mobile Devices
Official codebase for "Self Forcing: Bridging Training and Inference in Autoregressive Video Diffusion" (NeurIPS 2025 Spotlight)
A pytorch re-implementation of Real-time Scene Text Detection with Differentiable Binarization
「ICLR 2025」 A Sanity Check for AI-generated Image Detection
[CVPR2025] Any-Resolution AI-Generated Image Detection by Spectral Learning
Official PyTorch Implementation of "Latent Diffusion Model Without Variational Autoencoder".
An Open-Sourced LLM-empowered Foundation TTS System
Build and share delightful machine learning apps, all in Python. 🌟 Star to support our work!
[ICCV 2025] Official implementations for paper: VACE: All-in-One Video Creation and Editing
Learning Blind Video Temporal Consistency (ECCV 2018)
[CVPR 2025] Teaching Large Language Models to Regress Accurate Image Quality Scores using Score Distribution
[AAAI 2026] Official code release of our paper "Fine-grained Image Quality Assessment for Perceptual Image Restoration"
🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.
RepAlignLoss is a PyTorch loss function designed for guiding the training of a 'student' model by aligning its internal feature representations with those of a pre-trained 'teacher' model. It encou…
Official implementation of HYPIR: Harnessing Diffusion-Yielded Score Priors for Image Restoration (SIGGRAPH 2025)