Code help in running unet1 & unet2

[Jyothics]

Hello Sir,
I'm doing a capstone project using your codes. started facing issues in running unet1 & unet2.
Then I noticed these lines:
"We designed three neural networks accept different lengths of input at full, 1/2 and 1/8 resolutions, corresponding to "unet0.py", "unet1.py", and "unet2.py". To run the two lower 1/2 and 1/8 resolution models, the train and predict codes need to be modified."

Please help me in understanding what changes has to be made to train.py & predict.py for running unet1 & unet2.

[Hongyang449]

Hi,

A few places need to be modified for the 1/2 and 1/8 resolution (unet1 & unet2):
step 1. uniform_8m.py data preprocessing
line 26: path2='./data/feature_unet1/'
line 27: path3='./data/label_unet2/'
line 33: num_pool=1 for unet1 and num_pool = 3 for unet2
rerun and save the processed data for training and prediction next.

step 2. train.py e.g. unet1
line 43: import unet1
line 45: model = unet1.get_unet()
change the path to the processed data in step 1
line 51: path1='./data/feature_unet1/'
line 52: path1='./data/label_unet2/'
rename the weights e.g.
line 134: name_model='weights_unet1.h5'

step 3. predict.py
line 55: size = 4096*1024 for unet1 or size = 4096*256 for unet2
line 71: model01 = unet1.get_unet()
line 72: model01.load_weights('weights_unet1.h5')
line 99: num_pool=1 for unet1 and num_pool = 3 for unet2

Let me know if you still have issues after modifying the code.

Thanks,
Hongyang