https://github.com/JerryYann/DPI

tmp DPI
https://github.com/JerryYann/DPI

Last synced: 15 days ago
JSON representation

tmp DPI

• Host: GitHub
• URL: https://github.com/JerryYann/DPI
• Owner: JerryYann
• Created: 2023-07-24T06:21:36.000Z (over 1 year ago)
• Default Branch: master
• Last Pushed: 2024-12-18T05:50:57.000Z (about 2 months ago)
• Last Synced: 2024-12-18T06:43:16.451Z (about 2 months ago)
• Language: Python
• Size: 47.9 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-diffusion-categorized - [Code

README

        
# Diffusion Prior Interpolation for Flexibility Real-World Face Super-Resolution (DPI)

Aaai25 paper, experimental code. We will further improve the code in the future.

This is the codebase for [Diffusion Models Beat GANS on Image Synthesis](http://arxiv.org/abs/2105.05233).

# Download pre-trained models

From the [link](https://drive.google.com/drive/folders/1jElnRoFv7b31fG0v6pTSQkelbSX3xGZh?usp=sharing), download the checkpoint "ffhq_10m.pt" and paste it to ./weights/

# Sampling from pre-trained models

cd scripts

python -m torch.distributed.launch --nproc_per_node=1 --use_env --master_port=29522 sr_sample.py --model_path [pre_trained_path] --CRT_path [CRT_path] --h_sets [GT/Real-World Path] --downsampling 8 --out_path [out_path]

## Training your CRT

cd scripts

python -m torch.distributed.launch --nproc_per_node=1 --use_env --master_port=29522 image_train.py --data_dir [Your_training_sets] --batch_size 32 --lr_anneal_steps 100000