Running nicMSlesions in Google Colaboratory or with other env

[ghofranGh2022]

Hi @isabellameds @mbrown0294 @isabellameds @colpark
I need your help please about the running of the code nicMSlesions of SergiValverde
Can you help me please! Thanks.

[Sergi Valverde]

@ghofranGh2022 what is your actual problem?

[ghofranGh2022]

Hi @sergivalverde
Thank you
When i execute the file "nic_infer_segmentation_batch.py" to do the test, i find the problem: "valueError: attempt to get argmin of an empty sequence"
The input data is CHB_train_Case which contains FLAIR, t1, T2, lesion.
I use the model pretrained baseline_2ch.




I hope that I have explained my problem well Mr.Sergi!
Thanks

[Sergi Valverde]

Is your inference folder pointing to the data to infer? Looks like that it is pointing to a empty or inexistent folder.

[ghofranGh2022]

@sergivalverde I used the data from the CHB_train_case database and I put them in a folder called images, and I copied the path of these data in the configuration.cfg file and the default.cfg file, ?! I don't know how to make this data read by the program Mr.sergi.
Mr.Sergi, I have another question regarding the input data, the input data: flair and t1,? so what is the role of the mask (label)? also in the file description there are files like preprocess.py described by files without lesions!! what does that mean?
Thanks

[Sergi Valverde]

If you are running the code inside a collab instance then your inference folder should point to (as I see in your figure):

/content/gdrive/MyDrive/nicMSlesions/images

then inside images, you should have a folder for each subject. Please see how to run it:

https://github.com/sergivalverde/nicMSlesions#how-to-run-it

So inside images folder, each patient should be inside a different folder.

[ghofranGh2022]

Thank you Mr.Sergi @sergivalverde
I found the same error when i did folder organization, Mr.Sergi, is the problem related to google colab which only accept python 3 ?