HTTP 403 returned from http://yann.lecun.com/exdb/mnist/ when downloading .gz files

[mahtin]

Sadly the yann.lecun.com website is returning HTTP 403 errors when downloading the .gz files. Even after adding headers to the urlfetch() to make it look more like a browser, it still fails. (It fails in a browser also!).

This one line fix is all that I needed to get the code working. I used a different URL to download from.

< datasets_url = 'http://yann.lecun.com/exdb/mnist/'
> datasets_url = 'https://storage.googleapis.com/cvdf-datasets/mnist/'

This simple fix solves the issue until the website responds cleanly. It would be better to fix the website as this code is in a lot of places.

[headingy]

You save my day. Thanks

[JoshInLisbon]

Thanks for this!

For a bit more clarification (e.g. when using images = mnist.train_images() to check things are working), you want:

mnist.datasets_url = 'https://storage.googleapis.com/cvdf-datasets/mnist/'