LTX-Video 0.9.1 image2video and text2video workflow

Try LTX-Video 0.9.1 image2video and text2video online
https://ai.mimicpc.com/LTXVnew

Key Updates in LTX-Video 0.9.1

1. Bug Fixes:

   • Resolved an issue where short negative prompts caused artifacts when using the native CLIP Loader node.

2. Improved Model:

   • Eliminated "strobing texture" artifacts, resulting in smoother and more realistic video motion.

   • Enhanced motion generation quality.

   • Added STG (Spatial Temporal Guidance) support for better control over video content.

   • Integrated an image degradation system for improved motion generation in videos.

   • Introduced optional initial latent input, enabling latent chaining for high-resolution video generation.

   • Added image captioning functionality in image-to-video workflows for better content alignment.

Installation Guide

1. Open ComfyUI Manager.

2. Search for ComfyUI-LTXVideo in the list of nodes.

3. Follow the installation instructions.
   

1. Download the model file ltx-video-2b-v0.9.1.safetensors from Hugging Face and place it in the models/checkpoints folder of your ComfyUI directory.

2. Install a T5 text encoder, such as google_t5-v1_1-xxl_encoderonly, which can be managed through ComfyUI Model Manager.
   Image2Video workflow：
   Text2video workflow：