Skip to content

Conversation

Cyber-Machine
Copy link
Contributor

Partially fixes #830

@Cyber-Machine Cyber-Machine changed the title Default compilation for Debertav3 model Default compilation for Debertav3MaskedLM model Mar 11, 2023
@mattdangerw
Copy link
Member

Code looks good. Have we tried this out with a few different presets to make sure we are converging generally?

self.compile(
loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True),
optimizer=keras.optimizers.Adam(5e-5),
metrics=keras.metrics.SparseCategoricalAccuracy(),
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I believe this should be weighted_metrics as we pass sample weights in the preprocessing.

@mattdangerw mattdangerw merged commit 3e2589c into keras-team:master Mar 17, 2023
kanpuriyanawab pushed a commit to kanpuriyanawab/keras-nlp that referenced this pull request Mar 26, 2023
* Default compilation for Debertav3 model

* Formatted deberta_v3

* Minor fix

---------

Co-authored-by: Matt Watson <1389937+mattdangerw@users.noreply.github.com>
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.

Add compilation defaults for the MaskedLM task models
2 participants