-
Notifications
You must be signed in to change notification settings - Fork 611
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
Register keras objects in the function register_all
, not at import time
#1567
Comments
@gabrieldemarmiesse , I tried calling |
@ashutosh1919 , thanks for giving a hand. I don't believe it's the right approach. The right one would be to call |
@gabrieldemarmiesse , I have taken a good look at |
Yes, notably, we need to adapt it to addons. This function should look for bugs in addons (the layer implementation), not tensorflow. For example, I see that layer_test runs the layer once in a sequential model and once in a functional model. We shouldn't do that. We shouldn't even need to put it in a model to check that it works correctly. We can do |
TensorFlow Addons is transitioning to a minimal maintenance and release mode. New features will not be added to this repository. For more information, please see our public messaging on this decision: Please consider sending feature requests / contributions to other repositories in the TF community with a similar charters to TFA: |
Currently, it's possible to do
which is not readable and will throw an error with all main linting tools (import unused).
Being explicit and readable come first in python, and similarely to how we do it with custom ops,
register_all()
should be called before loading a model.See background #1151
The text was updated successfully, but these errors were encountered: