I want to train my own dataset, which consists of small patch size and the number is more than 1e6, then I find the process just evaluation but training, for example:
python main.py --model EDSR --scale 2 --save EDSR_x2 --n_resblocks 32 --n_feats 256 --res_scale 0.1 --dir_data ~/code/dataset --batch_size 1 --patch_size 64 --data_range 1-20000/20001-20010
20000
10
Making model...
Preparing loss function:
1.000 * L1
[Epoch 1] Learning rate: 1.00e-4
I want to train my own dataset, which consists of small patch size and the number is more than 1e6, then I find the process just evaluation but training, for example: python main.py --model EDSR --scale 2 --save EDSR_x2 --n_resblocks 32 --n_feats 256 --res_scale 0.1 --dir_data ~/code/dataset --batch_size 1 --patch_size 64 --data_range 1-20000/20001-20010 20000 10 Making model... Preparing loss function: 1.000 * L1 [Epoch 1] Learning rate: 1.00e-4
Evaluation: 100%|███████████████████████████████████████████| 10/10 [00:01<00:00, 9.68it/s] [DIV2K x2] PSNR: 10.102 (Best: 10.102 @epoch 1) Total time: 1.04s
Please tell me why this happened?Thank you.