-
Notifications
You must be signed in to change notification settings - Fork 19.7k
New pull request -- much cleaner #327
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why the camel case here only?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I don't know ... I think it's because of habit when using the unittest framework, I think it magically recognised that particular name, so it "has" to be camel case to be discovered appropriately. My knowledge could be stale, but that's why.
|
Cool! So are some tests still not passing? Which? |
|
All three tests in tests/auto/test_embeddings.py and tests/auto/test_regularizers.py are failing. Haven't investigated yet, but these exist on master and "someone" (possibly you) should be running py.test before committing to master. There is another failing test in our work in test_constraints.py which seems to be caused by an actual problem in the method implementation which is using the symbol 'e' but it's not a variable and not imported from anywhere. I tried math.e in its' place, but it caused another issue. I think the method got refactored to match the new "instance" style for constrain but not checked in any way, and needs to be rewritted to used theano.exp() functions or something. |
|
I have fixed the constraints tests. The loss weighting test succeeds when run manually but fails on py.test; is there something about pytest I am not getting? |
New pull request -- much cleaner
|
It's possible to get unexpected interactions with pytest and your main That helps keep things straight between the current release, the master If you're using the 'system' py.test, you might (for example) be picking up After your changes, for me, using py.test, everything except On 3 July 2015 at 10:48, François Chollet [email protected] wrote:
Tennessee Leeuwenburg |
|
All tests should pass now. |
* Add dtype to keras_core.operations.one_hot * Default dtype to None and use `backend.floatx()` instead * Use `backend.floatx()` in the __init__ itself
* Add dtype to keras_core.operations.one_hot * Default dtype to None and use `backend.floatx()` instead * Use `backend.floatx()` in the __init__ itself
Weird file deletions have gone. Rebase done. Sorry for before :)