omarsar

test_acc = 0.0
for i, data in enumerate(testloader, 0):
    inputs, labels = data
    inputs = inputs.view(-1, 28, 28)

    outputs = model(inputs)

    test_acc += get_accuracy(outputs, labels, BATCH_SIZE)
        
print('Test Accuracy: %.2f'%( test_acc/i))

omarsar

tensor([[0.6741],
        [0.9680],
        [0.9973],
        [0.9993]])
tensor([[ 0.9889],
        [-0.4191],
        [ 0.8347],
        [-0.0384]])

omarsar

Computing Resources

• Operating system: Preferably Linux or MacOS. If you have Windows, things may crash unexpectedly (try installing a virtual machine if you need to)
• RAM: Minimum 8GB
• Disk space: Mininium 16GB

---

Software Requirements

In this lab session we are going to be using Python as our main programming language. If you are not familiar with it, I recommend you start with this free Python course offered by Codecademy.

Here is a list of the required programs and libraries necessary for this lab session. (Please install them before coming to our lab session on Tuesday; this will save us a lot of time, plus these include some of the same libraries you may need for your first assignment).

omarsar

pip install medicaltorch

omarsar

from collections import defaultdict
import time
import os

import numpy as np

from tqdm import tqdm

from tensorboardX import SummaryWriter

omarsar

!pip3 install http://download.pytorch.org/whl/cu80/torch-0.4.0-cp36-cp36m-linux_x86_64.whl 
!pip3 install torchvision
!pip install medicaltorch
!pip3 install numpy==1.14.1

omarsar

ROOT_DIR_GMCHALLENGE = "/gdrive/My Drive/DAIR RESOURCES/PyTorch/medical_imaging/train/"
mri_input_filename = os.path.join(ROOT_DIR_GMCHALLENGE,
                                          'site1-sc01-image.nii.gz')
mri_gt_filename = os.path.join(ROOT_DIR_GMCHALLENGE,
                                       'site1-sc01-mask-r1.nii.gz')

pair = mt_datasets.SegmentationPair2D(mri_input_filename, mri_gt_filename)
slice_pair = pair.get_pair_slice(0)
input_slice = slice_pair["input"]
gt_slice = slice_pair["gt"]

omarsar

# transformer
composed_transform = transforms.Compose([
            mt_transforms.Resample(0.25, 0.25),
            mt_transforms.CenterCrop2D((200, 200)),
            mt_transforms.ToTensor(),
])

# load data
train_dataset = mt_datasets.SCGMChallenge2DTrain(root_dir=ROOT_DIR_GMCHALLENGE, transform=composed_transform)

omarsar

def threshold_predictions(predictions, thr=0.999):
    thresholded_preds = predictions[:]
    low_values_indices = thresholded_preds < thr
    thresholded_preds[low_values_indices] = 0
    low_values_indices = thresholded_preds >= thr
    thresholded_preds[low_values_indices] = 1
    return thresholded_preds

omarsar

# training dataset
train_transform = transforms.Compose([
        mt_transforms.Resample(0.25, 0.25),
        mt_transforms.CenterCrop2D((200, 200)),
        mt_transforms.ElasticTransform(alpha_range=(28.0, 30.0),
                                       sigma_range=(3.5, 4.0),
                                       p=0.3),
        mt_transforms.RandomAffine(degrees=4.6,
                                   scale=(0.98, 1.02),
                                   translate=(0.03, 0.03)),