Lightricks Releases Multiple LoRA and IC-LoRA Adapters for LTX-2.5

September 18, 2026

Lightricks Releases Multiple LoRA and IC-LoRA Adapters for LTX-2.5

At a Glance

Item Value
Repository Lightricks/LTX-2.5-22b-IC-LoRA-Ingredients
Published 2026-09-11
License other
Formats safetensors
Access Gated on Hugging Face (license agreement required)
Source type Primary source (the publisher itself)

Values determined by this site’s code at collection time. Dates are JST.

Overview

Lightricks has released multiple LoRA and IC-LoRA adapters for the open-weight video generation model Lightricks/LTX-2.5. These are additional weights designed to extend the capabilities of the base Lightricks/LTX-2.5 model, enabling video editing, character consistency maintenance, and the application of special visual effects.

Specifications

Targeting the base model Lightricks/LTX-2.5, the released adapters are each specified with an “other" license. Using them requires license agreement (gated access) on Hugging Face. The specifications of each newly released repository are as follows:

  • Lightricks/LTX-2.5-22b-IC-LoRA-Ingredients: A video-to-video model supporting reference sheets and character consistency (character-consistency)
  • Lightricks/LTX-2.5-22b-LoRA-Cinemagraph: An image-to-video model supporting cinemagraph production, motion graphics, and selective motion (selective-motion)
  • Lightricks/LTX-2.5-22b-IC-LoRA-Colorization: A video-to-video model supporting video colorization
  • Lightricks/LTX-2.5-22b-IC-LoRA-Decompression: A video-to-video model supporting video restoration and decompression (video-restoration)
  • Lightricks/LTX-2.5-22b-IC-LoRA-Deblur: A video-to-video model supporting video blur reduction (deblur)

Additionally, the base model Lightricks/LTX-2.5 broadly supports cross-modal conversion between image, video, and audio as well as multimodal inputs (text-to-video, image-to-video, video-to-video, audio-to-video, text-to-audio, etc.), and includes supported languages such as English (en) alongside German, Spanish, French, Japanese, Korean, Chinese, Italian, and Portuguese.

Strengths and Use Cases

These are all adapters based on Lightricks/LTX-2.5, provided with the aim of specializing the base model’s exceptional generation capabilities for specific processing and editing tasks.

  • Lightricks/LTX-2.5-22b-IC-LoRA-Ingredients: Suited for maintaining character consistency using reference images and generating video leveraging reference sheets
  • Lightricks/LTX-2.5-22b-LoRA-Cinemagraph: Can be utilized for cinemagraph effects that animate only a portion of a still image and motion design that applies movement to intended areas
  • Lightricks/LTX-2.5-22b-IC-LoRA-Colorization: Specialized for colorization tasks on monochrome or color-deprived videos
  • Lightricks/LTX-2.5-22b-IC-LoRA-Decompression and Lightricks/LTX-2.5-22b-IC-LoRA-Deblur: Effective for video restoration use cases such as repairing compression-degraded footage and clarifying blurred videos

Hardware Requirements

Estimated requirements (calculated by Local Model Watch)

Your VRAM Quantization File size Est. memory needed
4GB (laptop iGPU / phone class) そのままの精度 1.2GB 1.5GB

Memory estimates add a 20% runtime overhead (KV cache, etc.) to the actual size of the distributed files. Actual usage varies with context length, batch size and inference engine. These figures are computed by this site from file sizes, not published by the model’s authors. Compare with other models in our VRAM quick reference. What the quantization names mean: glossary.

How to Get It

Each model weight can be obtained from Hugging Face in the diffusion-single-file format. Since accessing each repository requires agreeing to the terms of use, users must agree to the terms on the respective page using a Hugging Face account prior to downloading.

Sources

Image, Video and AudioIC-LoRA,Lightricks,LoRA,LTX-2.5,動画生成,確認済み