Image
ControlNet
ControlNet — the original conditional control for diffusion models: steer Stable Diffusion with edges, depth, pose and scribbles instead of prompts alone, Apache-2.0
★ 34.1k starsApache-2.0
View guide →// image · Stability-AI/generative-models
Generative Models
Generative Models is an open-source image project by Stability-AI (Stability-AI/generative-models on GitHub). It has 27,318 GitHub stars and 3,090 forks, is released under the MIT license, and was last updated on GitHub on 16 December 2025.
Generative Models — Stability AI reference code for Stable Diffusion, SV3D and SV4D video-to-4D models, MIT
★ 27.3k stars⑂ 3.1k forksMITupdated 2025-12
▦ visual — this project's README ships no embeddable screenshot,
so here's a placeholder. The guide below is complete regardless.
so here's a placeholder. The guide below is complete regardless.
// project facts
Project facts
- Repository
- Stability-AI/generative-models
- GitHub stars
- 27,318
- Forks
- 3,090
- License
- MIT
- Last updated on GitHub
- 16 December 2025
- Category
- Image
- Listing
- Full install guide
// how it works
How it works
- Set up the PyTorch environment
Create a Python 3.10 virtual environment, install Torch with the matching CUDA build, then the project requirements and package.
- Download checkpoints
Fetch the model weights you need from Hugging Face into a checkpoints directory — for example the SV4D 2.0 weights.
- Sample or open a demo
Run a sampling script for video-to-4D generation, or launch a community Gradio demo for the model you chose.
// features
Why people love it
Stable Diffusion reference code
The canonical repository for Stability image-generation models, alongside the video models.
Video-to-4D with SV4D 2.0
Generate novel-view videos and 4D assets from a single input video, at 576x576 with strong spatio-temporal consistency.
Sampling scripts included
Ready sampling scripts for SV3D, SV4D and SV4D 2.0, with low-VRAM options for smaller GPUs.
Community Gradio demos
Local demo apps for the models, for trying generation without writing sampling code.
Advertisement
// run it
Run it yourself
- Install dependencies
The README environment route: venv, Torch, requirements and package.
python3.10 -m venv .generativemodels source .generativemodels/bin/activate pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118 pip3 install -r requirements/pt2.txt pip3 install . - Download SV4D 2.0 weights
The README checkpoint route from Hugging Face.
huggingface-cli download stabilityai/sv4d2.0 sv4d2.safetensors --local-dir checkpoints - Generate video-to-4D
The README SV4D 2.0 quickstart sampling command.
python scripts/sampling/simple_video_sample_4d2.py --input_path assets/sv4d_videos/camel.gif --output_folder outputs
// use cases
What you can build
- Generating novel-view videos and 4D assets from a single video
- Running Stable Diffusion family models from reference code
- Researching video-to-3D and video-to-4D diffusion models
- Trying Stability video models locally via Gradio demos
Advertisement
// faq
Questions, answered
Who makes Generative Models?
Stability AI, as the reference repository for its generative model family.
Is it free?
The repository code is MIT licensed; model checkpoints on Hugging Face carry their own terms — check the checkpoint you download.
What is SV4D 2.0?
Stable Video 4D 2.0: a video-to-4D diffusion model generating 48 frames (12 video frames x 4 camera views) at 576x576 from a 12-frame input video.
Do I need a GPU?
Yes — generation needs a GPU with the documented PyTorch/CUDA environment; low-VRAM options exist in the sampling scripts.
How do I get started?
Set up the Python 3.10 environment, download a checkpoint, and run a sampling script or a Gradio demo.
// keep exploring
Similar tools
Image
lite.ai.toolkit
A lite C++ AI toolkit: 100+ models with MNN, ORT and TRT, including Det, Seg, Stable-Diffusion, Face-Fusion.
★ 4.4k starsGPL-3.0
View guide →Image
SimpleTuner
A general fine-tuning kit geared toward image/video/audio diffusion models.
★ 2.9k starsAGPL-3.0
View guide →Image
Stable-Diffusion
FLUX, Stable Diffusion, SDXL, SD3, LoRA, Fine Tuning, DreamBooth, Training, Automatic1111, Forge WebUI, SwarmUI, DeepFake, TTS, Animation, Text To Video, Tut…
★ 2.8k starsGPL-3.0
View guide →Image
diffusers
🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.
★ 34.6k starsApache-2.0
View guide →Image
VAE-BlessUp
Easily modify and tune a Stable Diffusion VAE
★ 84 starsCustom
View guide →// alternatives
Alternatives in Image
The most-starred other Image projects in the LibreAI catalogue — the realistic shortlist if Generative Models is not the right fit.
- Stable Diffusion Webui★ 165.2k
Stable Diffusion web UI
- ComfyUI★ 136.0k
The node-graph powerhouse for precise, professional AI image and video generation.
- Immich★ 115.5k
High performance self-hosted photo and video management solution
- PhotoCraft★ 73.2k
PhotoCraft — a clean-room Photoshop reimplementation in pure Rust: layers, masks, adjustment layers, type, vectors, brushes and real PSD files, offline, MIT / Apache-2.0
- Fooocus★ 53.3k
Focus on prompting and generating
// more in image
More in Image
Nearby Image projects in the catalogue, closest in stars to Generative Models — keep exploring the shelf either side of it.
- InvokeAI★ 28.3k
A polished creative studio for professionals who want power without the node spaghetti.
- flux★ 26.0k
Black Forest Labs' official inference repo for the FLUX.1 image generation models
- Insightface★ 29.9k
State-of-the-art 2D and 3D Face Analysis Project
- squoosh★ 26.0k
Make images smaller using best-in-class codecs, right in the browser.
- czkawka★ 33.9k
Multi functional app to find duplicates, empty folders, similar images etc.
- pytorch-CycleGAN-and-pix2pix★ 25.3k
The classic PyTorch implementation of CycleGAN and pix2pix for image-to-image translation
Get Generative Models
Free and open — grab the source, star the repo, and start creating.