<div data-youtube-video><iframe width="640" height="480" allowfullscreen="true" autoplay="false" disablekbcontrols="false" enableiframeapi="false" endtime="0" ivloadpolicy="0" loop="false" modestbranding="false" origin playlist src="https://www.youtube.com/embed/9Dkd3JwkJWo" start="0"></iframe></div> ## East Asian Facial Fidelity LoRA for LTX-2.3This LoRA is designed to improve the natural appearance and identity consistency of East Asian female faces in LTX-2.3 Image-to-Video generation.It focuses on reducing facial drift during large camera movements and correcting exaggerated eye size, visible sclera, iris proportions, and double-eyelid geometry in full side-profile views. Its effects are especially noticeable in front-to-profile transitions and Camera Orbit shots.### Key Features- Improved facial consistency during large camera-angle changes- More natural eye geometry in left and right full side profiles- Better stability of the nose, lips, chin, and facial silhouette- Helps preserve the source subject, hairstyle, and clothing- A general East Asian female face prior, not a character-specific LoRA- No Trigger Word required### Usage- Base Model: <code>LTXV 2.3</code>- Recommended Task: <code>Image-to-Video</code>- Recommended LoRA Strength: <code>1.0</code>- Trigger Word: NonePlace the file in <code>ComfyUI/models/loras</code> and apply it through a LoRA Loader in your existing LTX-2.3 workflow. You can use ordinary subject phrases such as <code>woman</code> or <code>young woman</code> without a special activation token.The published <code>LORA OFF / LORA ON</code> comparisons use identical source images, prompts, seeds, sampling settings, Image Guide Strength, resolution, and camera movement. The application of this LoRA is the only intended difference between each pair.### Training InformationThe LoRA was trained with the official LTX Trainer and the LTX-2.3 22B Dev model. Following the initial general face-prior stage, an additional correction stage was completed using 1,127 manually selected full side-profile images.- Final Step: <code>11,019</code>- Rank / Alpha: <code>16 / 16</code>- Learning Rate: <code>5e-5</code>- Precision: <code>BF16</code>The most thoroughly verified use case is LTX-2.3 Image-to-Video generation featuring East Asian female subjects.This is a standard LoRA, not an IC-LoRA, ControlNet, full checkpoint, or character-specific identity model.### LicenseThis is a derivative model based on LTX-2.3 and is subject to the LTX-2 Community License Agreement.<a target="_blank" rel="ugc" href="https://huggingface.co/Lightricks/LTX-2.3/blob/main/LICENSE">https://huggingface.co/Lightricks/LTX-2.3/blob/main/LICENSE</a> ㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡㅡ

East Asian Facial Fidelity | LTX-2.3 I2V

First public release.Profile-correction pass on top of the base face-prior run. This checkpointspecifically improves side and three-quarter angles, where earlier steps stilllet the jawline and nose bridge drift during camera movement.Recommended strength: 1.0.

Files with SHA256: 9dc02c2033073cb2b31d37b972dfe443c769550ebef3f41a02ad9bfcb7fafbb4