y0ast/Faster MNIST.md

Last active May 30, 2021 21:01

Star () You must be signed in to star a gist
Fork () You must be signed in to fork a gist

Learn more about clone URLs
Clone this repository at <script src="https://gist.github.com/y0ast/f69966e308e549f013a92dc66debeeb4.js"></script>
Save y0ast/f69966e308e549f013a92dc66debeeb4 to your computer and use it in GitHub Desktop.

Download ZIP

Raw

Faster MNIST.md

Now available here: https://github.com/y0ast/pytorch-snippets/tree/main/fast_mnist

AlexPasqua commented May 22, 2021

Unfortunately that does not give the correct behavior: you're not randomizing your batches at each epoch which leads to significant reduced performance.

That's true, the shuffling should then be done manually. I think this should work:

train_dataset.data = train_dataset.data[torch.randperm(train_dataset.data.shape[0])]

(assuming the first dimension of train_dataset.data to be the batch size)