Home › Image Models › Stability AI – Generative Models

Stability AI – Generative Models - AI Image Models Tool

Overview

Stability AI – Generative Models is an open-source repository that collects implementations and sampling scripts for state-of-the-art image, 3D and video generative models. According to the GitHub repository, the project includes reference implementations for model families such as SV3D and SV4D and provides utilities for sampling, evaluation and training workflows. The codebase is intended as a research-to-production bridge: example configs, model definitions and sampling scripts make it easier to reproduce published experiments and integrate models into pipelines. The repository is oriented toward researchers and engineers who want direct access to model code, configurable sampling, and compatibility with GPU-accelerated PyTorch runtimes. It emphasizes reproducibility (configs and checkpoints) and extensibility (modular model/sampler code), and is maintained alongside Stability AI’s other open-source projects. For the latest commits, release notes, and specific pretrained checkpoints, consult the GitHub repository and linked documentation on the project page.

Installation

Install via pip:

git clone https://github.com/Stability-AI/generative-models.git
cd generative-models
pip install -r requirements.txt
python scripts/sample.py --config configs/<model>.yaml --ckpt /path/to/checkpoint --device cuda

Key Features

  • Reference implementations for SV3D and SV4D volumetric/video generative models.
  • Sampling scripts for text-to-image, image-to-image and video generation workflows.
  • Config-driven experiments: YAML configs to reproduce training and sampling setups.
  • Compatibility with PyTorch and CUDA for GPU-accelerated training and inference.
  • Utilities for checkpoint loading, evaluation metrics, and dataset preprocessing.

Community

The project is hosted on GitHub and maintained by Stability AI with community contributions via issues and pull requests. Users discuss usage and troubleshooting in the repo issues and in external forums (e.g., Hugging Face and Stability AI community channels). For up-to-date activity, check the repository commit history, open issues, and release tags on the project page.

Last Refreshed: 2026-01-09

Key Information

  • Category: Image Models
  • Type: AI Image Models Tool
Visit Official Website

Related Tools

Recraft V3

A text-to-image generative AI model capable of generating images with long text, available via Replicate’s API.
CodeFormer

A robust face restoration algorithm designed to repair old photos or improve AI-generated faces, delivering improved image quality.
GFPGAN

A practical AI tool for face restoration, capable of enhancing and restoring old and AI-generated faces, available for self-hosting via Docker.
FLUX.1-dev

High-quality image generation model with ComfyUI and Diffusers support, available under a non-commercial license.
OmniGen

OmniGen is a unified image generation model that can generate a wide range of images from multi-modal prompts, simplifying the image generation process without the need for additional network modules or preprocessing steps. It supports various tasks such as text-to-image generation, identity-preserving generation, image editing, and more.
FLUX1.1 [pro]

A new text-to-image AI model capable of generating images six times faster than its predecessor, with higher quality, better prompt adherence, and more diversity in outputs. It includes a prompt upsampling feature that utilizes a language model to enhance prompts for improved image generation.