Showing 301-320 of 351 projects
MeZO: A novel fine-tuning method for language models that requires just forward passes, ideal for vibe coders.
A text-to-video generation model that combines pixel-level and latent diffusion approaches.
A PyTorch-based implementation of the PaddleOCR optical character recognition model, enabling developers to leverage state-of-the-art AI for text detection and recognition.
A zero-shot image-to-image translation library for developers building AI-powered creative applications.
A Python library for creating and performing exact inference on Bayesian Belief Networks.
A Jupyter Notebook project for precise control and experimentation with diffusion models.
Recipes to scale inference-time compute of open models for AI/ML developers.
Neural Network Compression Framework for enhanced OpenVINO™ inference
The simplest way to serve AI/ML models in production, with support for popular models like Stable Diffusion and Whisper.
Pruna is a model optimization framework that helps developers deliver faster, more efficient AI models with minimal overhead.
Official PyTorch implementation of a prompt-based distillation method for on-device deployment of the Segment Anything Model (SAM).
Official PyTorch implementation of Cold-Diffusion, a technique for different transformations.
A tool for structurally pruning large language models like LLaMA, BLOOM, and Vicuna to reduce their size and inference time.
A Python benchmark suite for evaluating text-to-3D generation models and techniques.
A self-hosted web UI for over 30 generative AI tools, including Stable Diffusion, GFPGAN, and Whisper.
A comprehensive library for sentiment analysis, text classification, and text adversarial defense, tailored for AI-powered developers.
JoyCaption is an open, uncensored image captioning Visual Language Model (VLM) for training Diffusion models.
A powerful open-world object detection model for computer vision tasks, leveraging the DINO framework.
A minimal implementation of the Flash Attention algorithm in CUDA for efficient AI model inference.
Paddle.js is a web project for the Baidu PaddlePaddle deep learning framework, enabling browser-based inference.
Get weekly updates on trending AI coding tools and projects.