03_pytorch_computer_vision FashionMNIST

[kydaly]

I had a question regarding video 96 at time 14:48:38. When he uses .classes and .class_to_idx how can I look up what attributes the class has? I know the dropdown menu shows different attributes associated with the class but the descriptions are very vague. The FashionMNIST documentation does not provide the attribute information either. I appreciate any further details anyone can provide.
Thanks!

[mrdbourke]

Hi @kydaly ,

You're right, sometimes the PyTorch documentation can be a little lacking.

However, you can often find tidbits about things by reading the source code:

See here: https://pytorch.org/vision/stable/_modules/torchvision/datasets/mnist.html#FashionMNIST

I often find myself trawling through docs/source code like this.

[kydaly]

Thank you Daniel! I always wondered how to see the source code and always missed seeing that link. Thank you again for making this free course I have really been enjoying it! I am completely self taught with Python and really appreciate you putting such a thorough intro course up for free.

[arpit2167407]

Check the Dataset Class: Look at the FashionMNIST class in the source code of your library (like PyTorch) to see how self.classes and self.class_to_idx are defined.

Print Attributes: Load the dataset and print the attributes:

python
Copy code
from torchvision.datasets import FashionMNIST

dataset = FashionMNIST(root='path/to/data', train=True, download=True)
print(dataset.classes) # Class names
print(dataset.class_to_idx) # Mapping of class names to indices
Refer to Library Documentation: Check the documentation for the library you're using for additional context.

Community Forums: Look for similar questions on forums like Stack Overflow or GitHub.

Visualize Samples: Use matplotlib to visualize samples from each class for better understanding.