Overview

This is a video-to-video IC-Lora for LTX 2.3, which is designed to remove credits and other text overlays from anime. It generalizes to live-action content quite well, too.

Click the image to view demo video Demo

Usage

The model can be used through tools like ComfyUI or WanGP. I highly recommend WanGP as an easy way to use the model. I have included a settings file for WanGP with everything configured. I don't have a comfyui workflow, but if anyone wants to provide one, I will be glad to add it.

Trigger phrase: removetext

Lora strength 1.5 seems to work well in most cases. Highly stylized or animated text, as well as rendering at higher resolutions like 1080p may need higher strengths.

LTX 2.3 tends to struggle with fast motion, so you may see a reduction in quality during fast moving parts.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Zarxrax/removetext

Finetuned
(93)
this model