Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.
And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.
Save the trainability status of layers in JSON
See original GitHub issueAm I correct in saying that currently when the configuration of the model is saved as JSON, the trainable config parameter is always saved as true even if the layer is actually frozen (i.e. layer.isTrainable == false)? If so, shouldn’t it be fixed?
On a related note, similar to Keras, isn’t it better to do some refactoring and introduce a new method like getConfig for the base Layer (and possibly overridden by subclassed layers to add additional layer-specific configs) which then could be used for obtaining config of a layer (instead of creating isolated functions in ModelSaver.kt)?
Issue Analytics
- State:
- Created 2 years ago
- Comments:6 (2 by maintainers)
Top Results From Across the Web
Top Related Medium Post
Top Related StackOverflow Question
Troubleshoot Live Code
Top Related Reddit Thread
Top Related Hackernoon Post
Top Related Tweet
Top Related Dev.to Post
Top Related Hashnode Post
@dosier Feel free to take this ticket. This project is not mature enough, and some important parts like guidelines on contributing are missed, so I’ll add them in the nearest future; thanks for the advice.
I assigned this ticket for you
Regarding the ticket: I suggest saving trainable status for all layers when we serialize the model.
Great job @dosier congrats you with the first contribution