Skip to content

Conversation

vfrico
Copy link

@vfrico vfrico commented Oct 25, 2016

This is just a pull request because I'm using this library and is a feature I needed. Sometimes, when evaluating a model while it is training, the performance obtained with some hyperparameters can be really low. If so, it is better to stop the trainer friendly, adding this small lines of code, and test out another trainer with other hyperparameters.

@vfrico
Copy link
Author

vfrico commented Mar 27, 2017

Support for array indexing with non integers indexes has been dropped on 1.12 version of numpy. [https://github.com/numpy/numpy/blob/master/doc/release/1.12.0-notes.rst#deprecationwarning-to-error](See release notes)

The latest commit (e2600b7) fixes that error

@vfrico vfrico changed the title Trained can be stopped from outside NumPy 1.12 support Mar 27, 2017
@vfrico vfrico force-pushed the master branch 2 times, most recently from e968702 to e2600b7 Compare March 27, 2017 12:09
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant