TypeError: unhashable type: 'list' when calling .iloc()

Question

I'm currently doing some AI research for a project and for that I have to get used to a framework called "Pytorch". That's fine and all but following the official tutorial (found here) the code doesn't run properly.

The idea is that I analyse a set of facial features from a prepared dataset and then do something with it (haven't gotten to that part yet). But when I run this piece of code:

img_name = os.path.join(self.root_dir, self.landmarks_frame.iloc([index, 0]))  # At this point 'index' is 0

The dataset is initialized like this:

face_dataset = fDataset(csv_file='faces/face_landmarks.csv', root_dir='faces/')

And here is where the error pops up:

for i in range(len(face_dataset)):
    sample = face_dataset[i]  # <-- right there

That leads to the getter function:

def __getitem__(self, index):
    img_name = os.path.join(self.root_dir, self.landmarks_frame.iloc([index, 0]))
    image = io.imread(img_name)
    landmarks = self.landmarks_frame.iloc[index, 1:].as_matrix()
    landmarks = landmarks.astype('float').reshape(-1, 2)
    sample = {'image': image, 'landmarks': landmarks}

Found in my FaceLandmarksDataset(Dataset): class I simply get the error of the title. This is strange I find, because I can read the dataset just fine as a frame in PyCharm:

Where the first picture is clearly visible. I have checked as well that it is in the folder that I'm looking in.

Can anyone help out? :)

Answer 1

You don't need the parentheses with iloc:

self.landmarks_frame.iloc[index, 0]