Features
- Text-to-video generation from scene descriptions
- Image-to-video that preserves the image's content, structure and style
- VideoCrafter1 and VideoCrafter2 models
- Training code, inference code and configuration files
- Local Gradio demo app
- Pre-trained models on Hugging Face
- Open source on GitHub, intended for research and non-commercial use
Pricing Details
The VideoCrafter tool is available through its open-source repository on GitHub, where users can download the project’s code and models and run them locally or use them within appropriate development environments. The tool provides the VideoCrafter1 and VideoCrafter2 models for generating video from text (Text-to-Video) and converting images to video (Image-to-Video), with access to the training and inference code. Powerful computing resources, such as GPUs, are recommended for running the models efficiently, while adhering to the licensing and usage terms specified by the project.