CoDiVSR: Rethinking What to Condition and What to Disentangle in Diffusion-based Video Super-Resolution

Paper

News

• March 22, 2026: We have released our code and checkpoints.

To-Do List

• Release training code
• Release testing code
• Release pre-trained models
• Release demo

Quickstart

Environment Setup

# 1. Clone the repository and create a conda environment: 
git clone https://github.com/yyang181/CoDiVSR.git
cd CoDiVSR
conda create -n codivsr python=3.10 -y
conda activate codivsr

# 2. Install PyTorch. We recommend PyTorch 2.5.1 with CUDA 12.4:
pip3 install torch==2.5.1 torchvision torchaudio --index-url https://download.pytorch.org/whl/cu124

# 3. Install other requirements:
pip install -r requirements.txt

Download Checkpoints

pip install modelscope

# Pre-trained CogVideoX1.5-5B-I2V
modelscope download ZhipuAI/CogVideoX1.5-5B-I2V --local_dir ./checkpoints/CogVideoX1.5-5B-I2V

# CoDiVSR checkpoints
modelscope download yyang181/CoDiVSR --local_dir ./checkpoints/codivsr/transformer

Inference

The inference code has been tested on:

• Ubuntu 20.04
• Python 3.10
• PyTorch 2.5.1
• 1 NVIDIA H20 GPU with CUDA 12.4 (Requires ~40GB VRAM for a video with 1080x1920x100 resolution; inference takes approximately 3 minutes and 40 seconds).

We use PLLaVa-13B to extract captions.

# NTIRE 2026 SUGC-VR
CUDA_VISIBLE_DEVICES=0 python inference.py \
    --input_dir data/input \
    --input_json data/csv/input_text.csv \
    --output_path data/output/codivsr \
    --model_path checkpoints/codivsr \
    --is_vae_st \
    --save_format yuv420p \
    --upscale 1 \
    --load_skipconv1d \
    --use_low_pass_guidance \
    --enable_midresidual 

Acknowledgements

This project builds upon several excellent open-source projects:

• CogVideo - A large-scale video generation model developed by Tsinghua University that provides the foundational architecture for this project.
• DiffusionAsShader - Provides the foundational architecture for this project.

We thank the authors and contributors of these projects for their valuable contributions to the open-source community!