EasyAnimate
aigc-apps/EasyAnimateReviewed by AI
Checked by RepoPicks Radar on
What is EasyAnimate?
An Alibaba research project that generates videos of roughly 6 seconds at up to 1024x1024 from a text prompt, a starting image or a reference video, with camera and pose controls, and can also train your own style models. It has a browser page (Gradio), ComfyUI nodes, a Docker image and scripts you edit and run. You need Python 3.10 or 3.11, an NVIDIA card (12 GB works for small videos, 24 GB or more for the bigger ones) and about 60 GB of disk for the model files. It is a project for people comfortable with the command line.
EasyAnimate is free and open source (Apache-2.0) and can replace Runway, Pika, and Kling.
How easy is it to install EasyAnimate?
D · terminal, several tools or a specific graphics card · Windows
- Windows D
- Linux D
Both Windows and Linux need Python 3.10 or 3.11 with CUDA, a pip install of a long list of packages, a 30 to 39 GB model download and editing a Python file or starting the Gradio page by hand; the README says it was tested with an NVIDIA 3060 12 GB on Windows and V100/A10/A100 cards on Linux. Linux also offers a ready Docker image, but Docker plus an NVIDIA driver is no simpler.
Coming soon: install in 1 click with the RepoPicks app.
EasyAnimate on GitHub
- Stars
- 2,265on Nov 10, 2026
- Downloads
- —all releases
- Growth
- —no gain in the last reading
Stars over time
Facts
- Runs on
- Windows, Linux
- You need to install
- Python, NVIDIA graphics card
- To build from source
- Python, Graphics card, FFmpeg
- Disk space
- 1,010 MB–4.5 GB · estimate
- Accounts or keys
- None
- Cost
- Free
- License
- Apache-2.0 (Permissive)
- Last change on GitHub
- Mar 6, 2025
EasyAnimate risk signals (5)
- Attentionstale releases
Last release on 2024-05-29 (892 days ago)
automatic check
- Attentionno updates for over a year
Last push 2025-03-06 and no commits for 611 days, with a single release from 2024-05-29. The last README entry says the model moved into the diffusers library ("EasyAnimate-V5.1 is now supported in diffusers", 2025.03.06), so the repository looks parked while work continues elsewhere.
AI review
- Infoheavy gpu and disk use
README.md: "We need about 60GB available on disk (for saving weights)", model files of 30 to 39 GB, and a table where 768x1344 videos only run on 24 GB cards or more. The deterministic disk estimate (1 to 4.5 GB) covers only the code and packages, not the models.
AI review
- Infodocker image from vendor registry
README.md Quick Start: "docker pull mybigpai-public-registry.cn-beijing.cr.aliyuncs.com/easycv/torch_cuda:easyanimate". The ready-made image comes from Alibaba's own registry with no published build recipe in the repository.
AI review
- Infomodel files hosted elsewhere
The Model zoo section links every weight to huggingface.co/alibaba-pai or modelscope.cn. The Apache-2.0 license in the repository covers the code; the terms that come with those downloaded model files are not shown here.
AI review
Evidence is shown in its original language.
What Radar could not check about EasyAnimate
- Whether it still installs today, since the package list has not been touched in over a year and torch, diffusers and Gradio have moved on.
- Real disk use after downloading models, beyond the 60 GB the README asks for.
- How good the videos are on a 12 GB card with the memory-saving options, and how long each one takes.
- Whether the Hugging Face and ModelScope online demos linked in the README are still running.
- What licence the downloaded model files carry and whether they allow commercial use.