测试问题

[luobin97]

从谷歌云盘下载2倍的超分辨率预训练模型，测试Set5结果PSNR只有30.460 --model PAEDSR --data_test Set5 --save_results --rgb_range 1 --data_range 801-900 --scale 2 --n_feats 256 --n_resblocks 32 --res_scale 0.1 --pre_train D:\PANSR\pretrain_model\model_x2.pt --test_only --chop

[HarukiYqM]

Hi, sorry for the late reply.

It seems that your command is correct but I did not observe such effects. It maybe some weights are not correctly loaded. Please use this code base to avoid any version problem. You can also double check the results by first inference the outputs and evaluate the results using third party code like (https://github.com/yulunzhang/RNAN/blob/master/SR/experiment/PP_Evaluate_SR.m). Here is my result and my comand:

python main.py --model PAEDSR --save_results --data_test Set5 --dir_data ~/workspace --rgb_range 1 --data_range 801-900 --scale 2 --n_feats 256 --n_resblocks 32 --res_scale 0.1 --pre_train ../model_x2.pt --test_only --chop