For AI agents: a documentation index is available at the root level at /llms.txt and /llms-full.txt. Append /llms.txt to any URL for a page-level index, or .md for the markdown version of any page.
The LTX-2 Trainer enables researchers and advanced users to fine-tune LTX-2 models on custom datasets. You can train LoRAs, perform full fine-tuning, or create video-to-video transformations (IC-LoRA).
Who Should Use the Trainer?
The trainer is designed primarily for:
Researchers exploring video generation model training techniques
Advanced users building specialized video generation models
Requirements
Before starting, ensure you have:
LTX-2.3 Model Checkpoint - Local .safetensors file
Nvidia H100 GPU with 80GB+ VRAM - Lower VRAM may work with gradient checkpointing and reduced resolutions
Additional Documentation
For comprehensive training documentation, including setup instructions, dataset preparation guides, configuration options, and troubleshooting, visit the LTX-2 Trainer GitHub repository.
The GitHub documentation includes:
Quick Start Guide - Get training in minutes
Dataset Preparation - Video preprocessing and captioning
Training Modes - Detailed explanations of each training approach
Configuration Reference - All available training parameters
