-
Notifications
You must be signed in to change notification settings - Fork 617
Add subpackage for RNN related code. #189
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
1. The test case was missing test.main, which means the test wasn't executed at all. 2. Add initializer params for user to control how weights are initialized. 3. Update the tests with the fix.
seanpmorgan
left a comment
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.
Code looks very good and this is a nice addition to the repo. Some things need to be added though:
- Please import rnn in the global init
- Add rnn to the packaging BUILD file
- Add the subpackage to the project REAMDE
- Add a subpackage README describing the API style for RNN (inherit from keras.AbstractCell etc.)
1. Update the addon init/build/README 2. Add README for rnn.
|
Done. Thanks for the review. Btw, the build seems broken due to the recent release of tf estimator. |
seanpmorgan
left a comment
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.
Thanks! One small nit comment and then could you also modify .github/CODEOWNERS. I really like the explicit criteria for being accepted. This might be something we should apply to more subpackages. Thoughts @facaiy ?
Small changes and then we can merge this after our CI returns tomorrow.
|
Done. Updated doc and gitowner file. |
seanpmorgan
left a comment
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.
Looks great... thanks!
Slowly bring over the tf.contrib.rnn code to addons.
The selection criteria so far are:
The initial commit only contains one rnn cell so far (NASCell), expect more cells to come.