为什么使用--chop和不使用--chop得到的结果不一样

[237014845]

(EDSR) yjl@gxh:~/SR_work/adaptive_distill/EDSR-PyTorch-master/src$ python main.py --model RCAN --chop --pre_train /home/yjl/SR_work/adaptive_distill/FAKD/teacher_checkpoint/RCAN_BIX4.pt --test_only --data_test Urban100
Making model...
Load the model from /home/yjl/SR_work/adaptive_distill/FAKD/teacher_checkpoint/RCAN_BIX4.pt

Evaluation:
100%|█████████████████████████████████████████| 100/100 [00:49<00:00, 2.02it/s]
[Urban100 x4] PSNR: 26.819 (Best: 26.819 @epoch 1)
Forward: 49.59s

Saving...
Total: 49.59s

(EDSR) yjl@gxh:~/SR_work/adaptive_distill/EDSR-PyTorch-master/src$ python main.py --model RCAN --pre_train /home/yjl/SR_work/adaptive_distill/FAKD/teacher_checkpoint/RCAN_BIX4.pt --test_only --data_test Urban100
Making model...
Load the model from /home/yjl/SR_work/adaptive_distill/FAKD/teacher_checkpoint/RCAN_BIX4.pt

Evaluation:
100%|█████████████████████████████████████████| 100/100 [00:36<00:00, 2.75it/s]
[Urban100 x4] PSNR: 26.745 (Best: 26.745 @epoch 1)
Forward: 36.40s