Error in U-Net_2D_Multilabel_ZeroCostDL4Mic.ipynb: IndexError: list index out of range

[jazberna1]

I get the following error in section 5.1 Generate prediction(s) from unseen dataset when running the notebook U-Net_2D_Multilabel_ZeroCostDL4Mic.

IndexError: list index out of range

The error happens in function_29, in the following line:

print('Model input size: '+str(Input_size[0])+'x'+str(Input_size[1]))

I believe what is happening here is that Input_size is [(None, 256, 256, 1)], where Input_size[0] is the None (the batch size)

If comment the Input_size everywhere in the function I am able to get the predictions:

To Reproduce
I have run the following sections of the notebook:

1. Install U-Net dependencies
   1.1. Load key dependencies
2. Select your parameters and paths
   2.1. Setting main training parameters
   2.2. Data augmentation
   Here I disable Use_Data_augmentation
   2.3. Using weights from a pre-trained model as initial weights
3. Using the trained model
   5.1 Generate prediction(s) from unseen dataset
   Here in setion 5.1 is where I get the error.

Describe the bug
See beginning of the message

Expected behavior
Get model predictions

Screenshots

Desktop (please complete the following information):

• OS: macOS Monterey
• Browser: chrome
• Version: chrome version 127.0.6533.122

My version of the notebook is:

6. Version log
v2.1.2:

Correct for data loading to avoid .DS_Store or similar
v2.1.1:

Replaced all absolute pathing with relative pathing
v2.1:

Updated to TensorFlow 2.11
Updated to fpdf2 and add lines to ensure a proper format. Correct keras package version parsing.

Smartphone (please complete the following information):
Does not apply

[IvanHCenalmor]

Hello @jazberna1,

Sorry for the delay, it has been a really busy week 😅
Seems like the code it's failing when getting the input size, I will take a look to the notebook in ZeroCostDL4Mic.
I will see if I can reproduce the error and fix it, when I manage to solve the issue I will let you know here with another comment 🤗

One quick question to try to reproduce the error, when using section 2.3. Using weights from a pre-trained model as initial weights. Did you used any previous weights?

Thank you for reporting the bug! ❤️

[jazberna1]

Hello Ivan,

Thanks for coming back to me.
To your question:
when using section 2.3. Using weights from a pre-trained model as initial weights. Did you used any previous weights?

In section 2.3 I have:

Then I go straight to 5.1 Generate prediction(s) from unseen dataset where I have:

Inside the results folder I have unet_2d_multlilabel/weights_best.hdf5 so like this:
results/unet_2d_multlilabel/weights_best.hdf5

Then I comment out the line to avoid the error:
print('Model input size: '+str(Input_size[0])+'x'+str(Input_size[1]))

click 'Load and run' and get the predictions:

Hope this helps!
Jorge