[Solved] Errors with sklearn.fetch_openml in Python Machine_Learning Neaural_Networks

Question

When I was trying to run the code of the lesson on my PC, an error occurred indicating a ‘urllib.error.URLError’ and stating that ‘certificate verify failed’ when the program was trying to go through the function ‘fetch_openml’.

I don’t know what’s wrong. Somebody plz help!

Here is my code…

https://code.sololearn.com/c23U26sb9zu0/?ref=app

Answer 1

You might be using a wrong version of Python or wrong version of sklearn.
Find some ideas to debug here:

https://github.com/tensorflow/tensorflow/issues/10779

https://stackoverflow.com/questions/54365045/scikit-learn-cannot-load-mnist-original-dataset-using-fetch-openml-in-python

Answer 2

Tibor Santa Thanks! I’ll try~

Answer 3

The question is solved rather MIRACULOUSLY when I removed ‘version=1’ in my code, using ‘X, y = fetch_openml(‘mnist_784’, return_X_y=True)’ instead of ‘X, y = fetch_openml(‘mnist_784’, version=1, return_X_y=True)’, but the code runs really slowly…

Think I’ll just download the data from openml.org and use the .csv file directly… Thanks guys for your attention