Home / Tools / AnimeGANv2

AnimeGANv2

Open-source anime stylizer for photos and videos

AnimeGANv2 preview
github.com Open Source

Upvotes

0

Pricing

Open Source

Category

Cartoon Generators

In the database since

Oct 2026

Overview

About AnimeGANv2

AnimeGANv2 is a research project on GitHub that turns photos and videos into anime-style art. It was made by Xin Chen as a newer version of AnimeGAN. The AI is a GAN, a type of AI that learns art styles from examples. It was trained on frames from famous anime films. You run it on your own computer from the command line, not through an app or website. It needs Python 3.6 and TensorFlow 1.15, plus an NVIDIA GPU. Anime styles come ready to use.

Key Features

1. Three Film-Based Anime Styles

Styles come from Hayao Miyazaki, Makoto Shinkai, and Satoshi Kon. Each style was trained on frames from their films. Output looks like real anime screenshots.

2. Handles Photos and Videos

One command processes a whole folder of pictures. A separate script handles video clips too. It covers both still images and short films.

3. Small, Fast Model

The generator is only 8.17 MB, smaller than most image models. It saves disk space and runs fast. A lighter version exists for weaker machines.

4. Ready-Made Colab Notebooks

Two notebooks run on Google Colab with no local install. One handles photos, the other videos. You use a cloud GPU and download your results.

5. Trainable on Your Own Styles

Training code is included for new styles. Collect cartoon frames, run edge smoothing, and start training. It takes longer but gives a custom look.

Frequently Asked Questions (FAQ)

Can I use AnimeGANv2 output in a commercial project?

The license allows non-commercial use only, like research and teaching. For commercial use you need a written letter from the author. Contact him by email to ask.

Why do my results show strange artifacts?

Version 2 was built to fix the high-frequency artifacts of version 1. Make sure you use the v2 weights, and try the lite version if your GPU is weak. Lower input resolution can also help.

What is the difference between AnimeGANv2 and AnimeGANv3?

AnimeGANv3 is a later release with a new network design and its own paper. AnimeGANv2 is older but simpler to run, with ready weights and Colab notebooks. Try v3 if you want the newest quality.

Conclusion

AnimeGANv2 is a solid choice if you want anime-style photos or videos and do not mind running code. It suits developers and researchers who enjoy command-line tools, and the Colab route helps beginners test first. Just remember it is a research project, not an app, so check the GitHub page for setup guides and sample results before you start.

Newsletter

Tools like this change fast

New Cartoon Generators tools launch every week — the newsletter keeps you ahead of what's worth trying.

Keep exploring

Similar AI tools

Ranked by the job they do, their capabilities, and how closely they fit this workflow.