In notebook 001a_nn_basics.ipynb the mnist dataset is downloaded into the data folder.
But in notebook 002_images.ipynb we are using cifar10, or ‘cifar10_dog_air’, but there does not seem to be any code for downloading this dataset. Is this missing or am I missing something?
1 Like
I had the same problem and could not find the code downloading the data either.
As a workaround, downloading http://files.fast.ai/data/cifar10.tgz to data and renaming to the unzipped folder to cifar10_dog_air seems to make the notebook runnable.
Please note that those are development notebooks that aren’t tutorials or examples. They correspond to the stages of development of the library. Jeremy will comment them in the next part 2 of the course, but they’re not supposed to be used as is for now.
001a has been converted to a pytorch tutorial, which is why it’s fully documented, but don’t expect that in the next ones. If you delve into them, you’ll have a few things to figure out by yourself.
Thank you I will try that.
Ok, thanks. I kinda thought they were supposed to be ready to go as-is.
But I’m really learning a lot by going through them nonetheless.