House_Lee

main_test_rrdb.py

3

用main_train_rrdb_psnr.py训练出model之后，用test代码不能正常超分图片。 用原model = net(in_nc=3, out_nc=3, nc=64, nb=23, gc=32, upscale=4, act_mode='L', upsample_mode='upconv') 得出的输出会变黄色，后面按照option改成了 model = net(in_nc=3, out_nc=3, nc=64, nb=23, gc=32, upscale=4, act_mode='R', upsample_mode='upconv') 得出的输出会变黑色 训练过程中的验证是能正常超分出来的

Using the default setting, the kernel size is 17x17(X2) and 33x33(X4)

2

 But in your paper the kernel size should be 13x13. I look forward to your answer! Thanks!

What is your test result on NTIRE dataset?

3

Thanks for your excellent work! Did you test your result? (e.g., PSNR, LPIPS) . I want to know the differences between the paper and yours. Thanks!

Can you provide the option of PSRT-Sliding?

2

Thanks for your great work! I wonder to reproduce the result of PSRT-Sliding in the paper. Do you have a plan to release the option of PSRT-Sliding?

Can not test SSL_Vimeo_BD.pt

1

` model.load_prune(args,opt) [0/1920] File "/data1/project/SSL/basicsr/models/video_recurrent_prune_model.py", line 245, in load_prune prune_info = torch.load(load_path) File "/home/miniconda3/envs/SSL/lib/python3.8/site-packages/torch/serialization.py", line 789, in load return _load(opened_zipfile, map_location, pickle_module, **pickle_load_args) File "/home//miniconda3/envs/SSL/lib/python3.8/site-packages/torch/serialization.py", line 1131, in _load result...

The script of runtime/Flops testing

1

Thanks for the great work! I wonder to know how to accurately test the runtime and Flops (reported as 18ms and 63.9G in the paper). Can you provide the script...