RuntimeError: The size of tensor a (1204) must match the size of tensor b (1201) at non-singleton dimension 2

[Liampour]

Hello, I trained a model and I saved it with the name model2.pth but when I try to evaluate the results for an inputted .png image (with dimensions (1201x1159)) the following error appears

Traceback (most recent call last):
  File "colorize.py", line 188, in <module>
    main()
  File "colorize.py", line 184, in main
    evaluate(args)
  File "colorize.py", line 138, in evaluate
    utils.save_image_preserv_length(rec_img[0], img[0], save_path)
  File "/home/username/White-Balance-By-Deep-learning/utils.py", line 50, in save_image_preserv_length
    tensor = tensor * orilen
RuntimeError: The size of tensor a (1204) must match the size of tensor b (1201) at non-singleton dimension 2

The command that I used for the evaluation is the following
python3 colorize.py eval --input-dir input --output-dir output --model model/model2.pth

Any suggestions??