README.md

metadata

title: LTX V2V
emoji: 🎥
colorFrom: purple
colorTo: blue
sdk: gradio
sdk_version: 6.8.0
app_file: app.py
pinned: false

LTX V2V - Video to Video and Generation

This Hugging Face Space runs a distilled LTX-2.3 (22B parameter) Diffusion Transformer (DiT) model. It supports two primary functions:

1. Video to Video (V2V): Upload a video, specify a prompt, and choose a denoising strength (higher strength = more change) to stylize or edit the video. Includes options to keep the original audio, generate new audio, or output silent video.
2. Text/Image to Video: The classic distilled LTX-2.3 generation from a text prompt and/or static image.

Setup & Running

This app is designed to run on Hugging Face ZeroGPU. At startup, it automatically clones the LTX-2 codebase and installs its dependencies. It dynamically falls back to downloading models if the Hugging Face pre-mounts are not available.