2025年3月11日 · V1.0: ComfyUI Workflow: LTX IMAGE-to-TEXT-to-VIDEO Using Florence2 Caption. This workflow transforms the input images into a prompt (Florence2 for captioning) and uses the LTX Text to Video model for video generation (Image -> Prompt -> Video)

LTX-VideoQ8 is an 8-bit optimized version of LTX-Video, designed for faster performance on NVIDIA ADA GPUs. Repository: LTX-VideoQ8; Features: 🚀 Up to 3X speed-up with no accuracy loss; 🎥 Generate 720x480x121 videos in under a minute on RTX 4060 (8GB VRAM) 🛠️ Fine-tune 2B transformer models with precalculated latents

2024年12月24日 · LTXV是一个仅有20亿参数(2B)的基于DiT架构的视频生成模型，能够实时生成高质量视频。它以768x512的分辨率，每秒24帧的速度生成视频，最多生成4秒，支持文生图，图生图。另外，LTX Video 基于开发人员的反馈和真实世界

2025年3月29日 · V3.0: Introducing STG (Spatiotemporal Skip Guidance for Enhanced Video Diffusion Sampling). Included a SIMPLE and an ENHANCED workflow. Enhanced Version has additional features to upscale the Input Image, that can help in some cases.

STG 是一种用于增强扩散器的采样指导方法，无需下载额外的模型，只需修改模型中特定层的采样方法，即可得到更高质量的视频效果。 亲测有效，感兴趣的可以去试下。 3. LTX-Video API 调用. 本地部署毕竟太吃资源了。 目前，硅基流动 SiliconCloud 已上线 LTX-Video，关键是 免费调用。

A portion of Spatiotemporal Skip Guidance for Enhanced Video Diffusion Sampling (STG) has been implemented. Planning to add the "Restart" feature when time allows. Use this example workflow to see how it works. Left is without STG, and right is with STG. Deprecated, since this functionality has been integrated into core ComfyUI.

今天我们就来介绍一种可以增强 LTXV 视频效果的新技术： STG，STG 是一种用于增强基于变换器的视频扩散模型的简单、无需训练的采样指导方法。 不需要下载额外的模型，就可以实现更加高效高质量的视频效果。 好了，话不多说，我们直接开整。 我们直接来对比下原版的 LTXV 和使用了 STG 技术的文生视频效果和图生视频效果。 先来对比下文生视频： 提示词：A misty morning envelops a serene mountain valley, with a thin veil of fog hovering just above the dew-kissed …

[2024.12.19] 🖥️ ComfyUI STG support for LTXVideo! 🎬 Implemented in ComfyUI, enhancing LTXVideo support. Below are example videos showcasing the enhanced video quality achieved through STG: 🧪 Diffusers-based codes To run the test script, refer to the inference.py file in each folder. Below is an example using Mochi:

2025年3月25日 · LTX Video 是一个仅有 2B 参数的 DiT 架构视频生成模型,具有以下特点: 实时生成：能以超过实时播放速度生成视频 高质量输出：768x512 分辨率,24FPS 的流畅视频

2025年3月15日 · This is an entirely new flow that started as a test flow for new sampling methods for ltx. No notes, very raw. Should be easy enough to follow. Single image, 2 insert ways, upscale test. THe main things to use are CFG and STG levels. lower stg for more movement higher for less. and like perturbed rescale tells it when to stop.