AAAEEEE / Compressed_Image

Install

pip install -r requirements.txt

Pretrained Model

prepare two floders: weigth/dqt_hinet and weigth/swinir

Please downlaod two pretrained models:

hinet: https://drive.google.com/file/d/1bhjUDbq-TYp7GN146ghwEnk4BLM_ufV8/view?usp=sharing
swinir: https://drive.google.com/file/d/1OY8BV2A9VeIUsA2aPpxUCcpLxsxvyMVX/view?usp=sharing

Put hinet model:dqt_hinet_flikr2k_p512_epoch57.pth in path: weight/dqt_hinet

Put 001_classicalSR_DF2K_s64w8_SwinIR-M_x4.pth in weight/swinir

Inference

python src/eval_qm.py --realdatapath <your_image_val_path> --model <hinet_pretrained_model_path>
python swinir/main_test_swinir.py 

The final result images will be saved in path: results/swinir_classical_sr_x4