Main Script of the study¶
Created on Thu Dec 3 10:07:36 2020
This script is for detection IDH and TERTp mutation in gliomas. 1D-CNN is the approach that we use in this study. This study was submitted to ISMRM 2021
Abdullah BAS abdullah.bas@boun.edu.tr BME Bogazici University Istanbul / Uskudar @author: abas
-
main.get_data_loaders(train_batch_size, test_batch_size)¶ This function is for initializing data_loaders
- Parameters
train_batch_size (int) – Can be manipulated from config.py
test_batch_size (int) – Can be manipulated from config.py
- Returns
Returns train and test dataloaders.
- Return type
dataloaders
-
main.test(model, test_loader)¶ Testing loop for optuna
- Parameters
model (torch.model) – model to input
test_loader (dataloader) – dataloader for testing
- Returns
accuracy acc1 [float]: accuracy of class1 acc2 [float]: accuracy of class2 losses [list]: loss list
- Return type
accuracy [float]
-
main.train(log_interval, model, train_loader, optimizer, epoch)¶ This is for training loop of optuna
- Parameters
log_interval (int) – Logging interval
model (torch model) – 1D-CNN model that we created by using model1DCNN.py
train_loader (dataloader) – train data_loader for training
optimizer (torch.optimizer) – you can select the optimizer
epoch (int) – Epoch number
- Returns
Returns training loss
- Return type
[float]
-
main.train_optuna(trial)¶ This function is for optimizing the hyperparameters.
- Parameters
trial (optuna.study) – This comes from optuna. Initialized study
- Returns
Test loss value. In this study optuna tries to minimize this value.
- Return type
test_loss [float]