Tool for high-resolution video generation from prompts using AI. It supports two stages: generating low-resolution video and enhancing it to 1080p. Implemented with PyTorch, it includes model weights and uses Jupyter notebooks for execution.
Features
High-Resolution Video Generation
Provides two-stage video generation: first stage generates 270P video and second stage enhances it to 1080P. It allows efficient high-resolution video creation from lower resolution inputs.
Text to Video Generation
Generates videos from text prompts, trained using long prompts to achieve detailed outputs. Supports both simple and complex video generation.
Inference Code and Model Weights
Includes inference code and model weights for users to perform video generation on their own systems.
Jupyter Notebook Integration
Allows video generation using Jupyter notebook, optimized for an 80G GPU, facilitating ease of use and experimentation.