LTX Trainer Update Makes Character LoRA Creation Easier
The update lowers the technical barrier for creating character and style LoRAs with LTX-2.3, making the free, locally-run video generation AI more accessible to non-programmers.
Reporting from 1 sources: GIGAZINE.
Lightricks announced a major update to LTX Trainer, the official LoRA development tool for its local video generation AI LTX-2.3. The update allows users to create LoRA and IC-LoRA for multiple purposes by rewriting a single configuration file, eliminating the need for custom code. A SKILL.md file for AI agents was also released.
Lightricks announced on June 17 a major update to LTX Trainer, the official LoRA development tool for its LTX-2.3 video generation AI. The update lets users create LoRA and IC-LoRA for text-to-video, image-to-video, audio-to-video, video-to-video, video extension, inpainting, and outpainting by editing a single configuration file, replacing the earlier need to build custom code for each purpose.
LTX-2.3 is a free AI model that runs locally on personal PCs via tools like ComfyUI and can generate videos with audio including dialogue and background music. LoRA fine-tuning allows control over art style and consistent depiction of specific people or characters. The update also adds a SKILL.md file for instructing AI agents such as Claude Code to create LoRAs, further simplifying the process for users unfamiliar with video generation AI.
Hardware requirements remain high: LoRA training demands an NVIDIA GPU with at least 80GB VRAM, or a minimum 32GB VRAM GPU like the GeForce RTX 5090 with memory-saving features enabled. Lightricks released several demo LoRAs on Hugging Face, including ones for consistent character video generation, day-to-night scene conversion, and colorization of black-and-white footage.
Synthesized by Yomimono from the 1 cited source below, including Japanese-language reporting where cited, then editorially reviewed before publishing.