1、I run your code, and the hyperparameters remain unchanged,but But the result is very poor.
[FINAL] TRAINING COMPLETED
Best VALIDATION Accuracye 0039 IoU 0. 004 Mean Accuracy: 0. 026
Orgnal. FuseNet Accuracy: 066 IoU: 0.327 Mean Accuracy: 0. 434
Sat Nov 24 16:32:49 2018 Wed Nov 28 20:35:07 2018
Process finished with exit code 0
2、When i run ‘FuseNet_Class_Plots.ipynb’ in thie line ‘checkpoint = torch.load(model_path)’ I got the following errors:
UnicodeDecodeError: 'ascii' codec can't decode byte 0x9e in position 1: ordinal not in range(128)
I know the reason is your checkpoint_class_27 encoding of python2.7,my default encoding of python is utf8, but i don’t konw hoe to sovle it.And i run the checkpoint is too poor to use.
thanks!
Thank you very much for your reply!
I also tried out your previous code:https://github.com/zanilzanzan/FuseNet_PyTorch
First ,i want to know which one is better to run,this one or https://github.com/MehmetAygun/fusenet-pytorch
Second, if you Recommended me to run https://github.com/zanilzanzan/FuseNet_PyTorch,can you help me solve the following problems:
Process finished with exit code 0
2、When i run ‘FuseNet_Class_Plots.ipynb’ in thie line ‘checkpoint = torch.load(model_path)’ I got the following errors:
UnicodeDecodeError: 'ascii' codec can't decode byte 0x9e in position 1: ordinal not in range(128)
I know the reason is your checkpoint_class_27 encoding of python2.7,my default encoding of python is utf8, but i don’t konw hoe to sovle it.And i run the checkpoint is too poor to use. thanks!