Trending Projects

A deep reinforcement learning library for training robotic agents to plan pushing and grasping actions for manipulation tasks.

A Swift library for optical character recognition (OCR) of Chinese ID cards, using machine learning.

This is a repository about the true story behind the development of the Pangu large language model, a cautionary tale for AI researchers.

A mobile robotics project focused on bimanual manipulation and teleoperation using low-cost hardware.

A fast and simple framework for building neural data processing pipelines using Python.

Sample CUDA programming codes for GPU-accelerated molecular dynamics simulations

A command-line tool to search and download full transcripts of YouTube videos for semantic search and analysis.

This is a large language model (LLM) focused on mental health, with pre/post-training, datasets, evaluation, and deployment tools.

A Go-based RSS reader and aggregator that uses AI to enhance the feed experience.

WhisperFusion builds upon WhisperLive and WhisperSpeech to provide seamless conversational AI.

Official repo for Pai-Megatron-Patch, a large language model and visual language model training framework developed by Alibaba Cloud.

This is a dataset of character animation and motion capture data for developers working on AI-powered animation tools.

A collection of pre-trained StyleGAN 2 models for AI-powered generative art and image synthesis.

A TensorFlow implementation of a deep learning-based self-driving car model.

An open-source, feature-rich web UI for image matching and visual localization tasks using AI models like SIFT, SuperPoint, and SuperGlue.

A Go-based library for developers building AI-powered tools and applications.

A fast and differentiable model predictive control (MPC) solver for PyTorch.

A Python library for generating synthetic datasets and tools for computer vision applications.

MIRAI is a Rust mid-level IR Abstract Interpreter for analyzing and optimizing Rust code.

A real-time and accurate full-body multi-person pose estimation and tracking system written in Python.

A Python library that allows developers to interact with ChatGPT and other large language models using a Xiaomi AI speaker.

An AI-powered tool for automated test generation and code coverage enhancement.

Auto detecting, masking, and inpainting tool for stable diffusion models, built with Python.

An AI-driven local automation assistant that uses natural language to make computers work by themselves.

Improved AnimateDiff for ComfyUI and Advanced Sampling Support for AI-powered animation and image generation.

A collection of optimized TensorFlow binaries with SIMD instructions for improved performance.

A Python library for deep and online learning with spiking neural networks (SNNs) using PyTorch.

A Python library that projects the motion of pixels to a voxel representation, useful for vibe coders working with AI tools.

Bailing is an open-source AI voice assistant built with ASR, LLM, and TTS, supporting low-latency response on low-end devices.

A C++ library for fast and controllable 3D editing using Gaussian splatting, presented at CVPR 2024.

A computer vision package that makes it easy to run image processing and AI functions using OpenCV and Mediapipe.

A Python library for speech restoration, including tasks like declipping, denoising, and dereverberation.

Real-time photorealistic talking-head animation system built with Python and deep learning.

A Python library for efficient autonomous driving using vectorized scene representation.

Crab is a flexible, fast recommender engine for Python that integrates classic information filtering recommendation algorithms.

Helpful tools and examples for working with flex-attention in PyTorch

A curated collection of must-read papers and blogs on speculative decoding techniques for developers.

Official PyTorch implementation of a scalable transformer-based generative model for image generation and manipulation.

Unofficial PyTorch implementation of the Conformer model for speech recognition tasks.

A PyTorch implementation of the paper 'All are Worth Words: A ViT Backbone for Diffusion Models'.

A real-time, expressive chatbot using LLM and TTS, with support for QQ robot and various media types.

Lumina-mGPT 2.0 is a stand-alone autoregressive image modeling tool powered by Python.

A unified toolkit for deep learning-based document image analysis and layout parsing.

DiffSinger is a singing voice synthesis system using a shallow diffusion mechanism, enabling efficient TTS and SVS.

MAGI-1 is a powerful autoregressive model for generating high-quality video at scale, built for vibe coders who work with AI tools.

TorchIO is a Python library for efficient medical image preprocessing and data augmentation for AI applications.

A video foundation model and dataset for multimodal understanding and video understanding tasks.

Notebooks for a course on Deep Learning with TensorFlow 2 and Keras, focused on AI and machine learning.

MotionGPT is a unified motion-language generation model that can generate human motion using large language models.

A set of pre-trained machine learning models for the spaCy NLP library.