Skip to main content
Cross Validated

You are not logged in. Your edit will be placed in a queue until it is peer reviewed.

We welcome edits that make the post easier to understand and more valuable for readers. Because community members review edits, please try to make the post substantially better than how you found it, for example, by fixing grammar or adding additional resources and hyperlinks.

Required fields*

7
  • 1
    $\begingroup$ +1, but "bloody Jupyter Notebook"? Care to comment on that? :) $\endgroup$ Commented Jun 28, 2018 at 14:38
  • 11
    $\begingroup$ Here's why I hate Jupyter Notebooks. TL;DR: hidden state, diffing is a pain, security issues and it encourages bad programming practices, such as not using unit/regression/integration tests. Training NNs is already hard enough, without people forgetting about the fundamentals of programming. $\endgroup$ Commented Jun 28, 2018 at 16:20
  • 2
    $\begingroup$ I'm possibly being too negative, but frankly I've had enough with people cloning Jupyter Notebooks from GitHub, thinking it would be a matter of minutes to adapt the code to their use case and then coming to me complaining that nothing works. For cripes' sake, get a real IDE such as PyCharm or VisualStudio Code and create a well-structured code, rather than cooking up a Notebook! Especially if you plan on shipping the model to production, it'll make things a lot easier. $\endgroup$ Commented Jun 28, 2018 at 16:24
  • 2
    $\begingroup$ Lol. 'Jupyter notebook' and 'unit testing' are anti-correlated. $\endgroup$ Commented Jun 28, 2018 at 16:38
  • 2
    $\begingroup$ (+1) This is a good write-up. The suggestions for randomization tests are really great ways to get at bugged networks. $\endgroup$ Commented Aug 1, 2018 at 4:05