OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Image & Video AI

Videocolorgrading

Videocolorgrading is a Python research implementation that uses a diffusion model to generate look-up tables for aligning the color attributes of an input video with a reference scene.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Videocolorgrading — OpenViST project profile in Image & Video AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use the inference workflow when you have a reference image, an input video, pretrained weights and a configured Python environment, or use the training and evaluation scripts for research workflows.

Who it is for

  • Video creators seeking reference-based color grading through a Python command-line workflow.
  • Computer-vision researchers who want to train or evaluate the project's two-stage model.

Documented uses

  • Generate a graded video from an input clip and reference image.
  • Train GS-Extractor and L-Diffuser using video and LUT datasets.
  • Evaluate results using video and LUT dataset paths.
Before you choose

Requirements & limits

The repository is public and specifies the Apache-2.0 license. Inference requires a separate download of pretrained weights.

  • The documented interface is a Python command-line script rather than a video-editor plugin.
  • Inference depends on separately downloaded pretrained model weights.
  • Training requires environment-specific configuration, video data, LUT data and a first-stage checkpoint.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Run inference on one reference and video

Generate a graded video using a reference image and an input video.

Before you start

  • A Linux/Python environment with dependencies installed through fast_env.sh or manually.
  • Pretrained model weights placed in pretrained/, or their path updated in configs/prompts/video_demo.yaml.
  • A reference image and an input video.

Steps

  1. Clone the repository with git clone https://github.com/seunghyuns98/VideoColorGrading.git.

  2. Set up dependencies by running source fast_env.sh, or install them manually using fast_env.sh as a reference.

  3. Download pretrained weights and place them in pretrained/; update configs/prompts/video_demo.yaml if using another location.

  4. Run python video_demo.py --ref_path examples/reference1.jpg --input_path examples/video1.mp4 --save_path output/example1.mp4, replacing the example paths as needed.

Expected outcome

The inference script writes a graded video to the specified save path.

Guide limitations

  • The documented inference example processes one reference-image and input-video pair.
  • Weights are a separate download from the repository.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
12 October 2026
Original website title
seunghyuns98/VideoColorGrading · Official homepage title · retained in its source language
How we review sourcesSuggest a correction