Mochi
Mochi is a text-to-video generation model based on the GGUF quantized version, supporting video content generation from text descriptions.
Downloads 140
Release Time : 12/24/2024
Model Overview
This model is a quantized version of genmo/mochi-1-preview, focusing on text-to-video generation tasks, optimized for memory usage and speed with the GGUF format.
Model Features
GGUF Quantization
Utilizes GGUF format for quantization, significantly reducing model size and memory usage.
Memory Optimization
Speed improvement of approximately 50% through revised workflow and adoption of fp8_e4m3fn files, addressing memory shortage issues.
Efficient Text Encoding
Uses T5XXL as the text encoder, performing well and supporting multiple quantized versions.
Model Capabilities
Text-to-Video Generation
Supports English text input
Generates dynamic video content
Use Cases
Creative Content Generation
Natural Scene Generation
Generates dynamic videos of natural scenes based on text descriptions, such as a fox in a winter forest.
Generates dynamic video files (e.g., .webp format)
Featured Recommended AI Models
Š 2025AIbase