RL lora for LTX2.3. It greatly increases coherence and quality while reducing artifacts.

r/StableDiffusion
Reinforcement Learning

Submitted by /u/Different_Fix_2217 [link] [comments]