Convering my 2D CNN model to 3D

vision
1

class Skyrmion_or_Not(Dataset):

def __init__(self, csv_file, root_dir, transform=None):
        self.annotations = pd.read_csv(csv_file)
        self.root_dir = root_dir
        self.transform = transform

def __len__(self):
    return len(self.annotations)

def __getitem__(self, index): # return specific image and target for image
    
    # Some aspect of the code ignores the first row of the csv file and treats as the header
    
    img_path = os.path.join(self.root_dir, self.annotations.iloc[index, 0])
    # pytorch chooses index. idea sytem looks up file name in the labels file 
    # and then finds corresponding image/file in the relevant folder
    image = np.load(img_path)
    y_label = torch.tensor(int(self.annotations.iloc[index, 1])) # reading column 1 as the label

    if self.transform:
        image = self.transform(image)

    return (image, y_label) # the y_label is it being assigned 0, 1.

This is my current custom dataset for my 2D case. I now have 3D images (dimensions 30 x 30 x 30 x 3), however I am unsure what to adjust here.

Folloeing on from this, how does one visualise a pure 3D image on matplotlib. I am aware of how I can do it on plotly, however, I want to visualise batches (previously used plt.imshow, however this only has 2d inputs). Any help would be much appreciated!

2

Can’t think of any reason why you’d need to change anything in the Dataset/Dataloader to go from 2D to 3D.

The model, however, is another story.