why should use pre_train?

sanghyun-son / EDSR-PyTorch

PyTorch version of the paper 'Enhanced Deep Residual Networks for Single Image Super-Resolution' (CVPRW 2017)

MIT License

2.43k stars 669 forks source link

why should use pre_train? #235

Closed weishuoycbb closed 4 years ago

weishuoycbb commented 4 years ago

I run python3 main.py --model EDSR --scale 2 --patch_size 96 --save edsr_baseline_x2 --reset and run python3 main.py --data_test Demo --scale 2 --test_only --save_results get this wrong output I see some anwser ,say use pre_train Why?

sanghyun-son commented 4 years ago

It's because you are using a randomly initialized model. Please load the pre-trained models to get a proper result. Thank you!