depthMultiplier not saved for separableConv2d

[ersimont]

Please make sure that this is a bug. As per our
GitHub Policy,
we only address code/doc bugs, performance issues, feature requests and
build/installation issues on GitHub. tag:bug_template

System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow.js): yes
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Windows 10
• Mobile device (e.g. iPhone 8, Pixel 2, Samsung Galaxy) if the issue happens on mobile device: n/a
• TensorFlow.js installed from (npm or script link): yarn add @tensorflow/tfjs-node
• TensorFlow.js version (use command below): 3.13.0
• Browser version: n/a
• Tensorflow.js Converter Version: n/a

Describe the current behavior

When saving a model I created that includes tf.layers.separableConv2d, it does not include the depthMultilpier in the resulting model.json. This causes an error when trying to load the file, e.g.:

Error: Shape mismatch: [5,5,3,1] vs. [5,5,3,6]

If I manually update model.json to have "depthMultiplier": 6 (in this example), the models loads and behaves correctly.

Describe the expected behavior

Models that are saved can also be loaded.

Standalone code to reproduce the issue

const model = tf.sequential();
model.add(
  tf.layers.separableConv2d({
    inputShape: [96, 96, 3],
    filters: 9,
    kernelSize: 5,
    depthMultiplier: 6,
  }),
);
await model.save('file://bad-model');
await tf.loadLayersModel('file://bad-model/model.json');

Other info / logs Include any logs or source code that would be helpful to
diagnose the problem. If including tracebacks, please include the full
traceback. Large logs and files should be attached.

[ersimont]

Maybe the cause of #1149?

[shmishra99]

Hi @ersimont
Apologies for the delay in my response. I was able to replicate the issue you reported using the latest version of @tensorflow/tfjs@4.4.0, and I encountered the same error.
I have included the code below for reference:
Code :

Screenshot:

@mattsoulanille Do we have any update on the issue ? Thank You!