Image
pytorch-CycleGAN-and-pix2pix
The classic PyTorch implementation of CycleGAN and pix2pix for image-to-image translation
★ 25.3k starsNOASSERTION
View guide →Unofficial DragGAN implementation: drag points on a GAN image to edit it interactively
DragGAN is an open-source image project by OpenGVLab (OpenGVLab/DragGAN on GitHub). It has 4,936 GitHub stars and 473 forks, is released under the NOASSERTION license, and was last updated on GitHub on 17 July 2023.
Use the Colab or InternGPT online demo for a quick look, or install locally for custom images and models.
DragGAN is published on PyPI, so a local install is a single pip command (a GPU runtime is recommended).
Start the Gradio interface and pick a StyleGAN2 or StyleGAN2-ada checkpoint; checkpoints download automatically.
Set handle and target points on the image, optionally define a movable region, and run the manipulation.
Interactive handle/target point manipulation on the generative image manifold, with multiple handle points supported.
Higher-quality generation across more image types by selecting models whose names start with 'ada'.
Upload your own image and edit it through GAN inversion (the README notes inversion can distort some images).
A Gradio interface controls the generation process, plus Colab and InternGPT-hosted demos.
StyleGAN2 checkpoints download automatically, and you can download the generated image and generation trajectory.
Define a movable region so drags affect only the part of the image you intend to change.
Clone the project and enter its directory.
git clone https://github.com/OpenGVLab/DragGAN.git
cd DragGANThe README documents DragGAN on PyPI; version 1.1.0 is recommended if you want to upload custom images.
pip install draggan==1.1.0Follow the repository's INSTALL.md for the full local deployment steps, then launch the Gradio interface.
# see INSTALL.md in the repo for the exact local run commandOpenGVLab hosts this unofficial implementation (originally by Zeqiang-Lai); the underlying DragGAN paper is by Xingang Pan and co-authors, and an official implementation is linked from the README.
The code is publicly available on GitHub and PyPI, but the README states no license, so its license is unasserted — check the repository before commercial reuse.
No. It runs locally (or in the free Colab demo); a GPU is strongly recommended for usable speed.
This is an unofficial reimplementation aiming at full features, including StyleGAN2-ada and custom-image inversion; the official code, released in June 2023, is linked from the README if you prefer the authors' version.
Run `pip install draggan` (or `pip install draggan==1.1.0` for custom-image upload), then follow INSTALL.md — or open the Colab demo and select a GPU runtime.
The classic PyTorch implementation of CycleGAN and pix2pix for image-to-image translation
Advanced Nano Banana image studio — generate, mask-edit and track versions
Extract clean line sketches from anime art, illustrations, and manga with a pretrained GAN model.
AI image descriptions for Immich — Ollama analyzes, database updates
CUT — contrastive unpaired image-to-image translation in PyTorch, faster and lighter than CycleGAN (ECCV 2020)
Real-ESRGAN image and video upscaler app with GPU acceleration
The most-starred other Image projects in the LibreAI catalogue — the realistic shortlist if DragGAN is not the right fit.
Stable Diffusion web UI
The node-graph powerhouse for precise, professional AI image and video generation.
High performance self-hosted photo and video management solution
Focus on prompting and generating
🆙 Upscayl - #1 Free and Open Source AI Image Upscaler for Linux, MacOS and Windows
Free and open — grab the source, star the repo, and start creating.