Saving custom tensor fails

[awav]

Hello everyone,

There is a bug with saving and loading of custom tensors (variables) in TF. It was discovered at GPflow/GPflow#1127. I have created the MWE in colab here. There are two errors when tf.save_model applies to custom tensors:

• The loading for custom tensor (variable) restores broken tf.Variable object
• Saving of tf.Module with custom tensors (variables) fails

Details in colab notebook

PS: referencing @alextp and @dynamicwebpaige, because they know what I mean by custom tensor.

Thanks,
Artem

[alextp]

@k-w-w as far as I can tell those module-visible variables should be correctly handled by savedmodel. Do you know what is going on here?

[awav]

@k-w-w, @alextp ping ^^^

[k-w-w]

This is an interesting use case. SavedModel should be correctly handling this, but we hadn't considered this case yet. I'll see what changes are required to resolve the issue.

[awav]

@k-w-w, @alextp any progress on the issue?

[awav]

@k-w-w, @alextp ping

[awav]

@k-w-w, @alextp Hello! Because of this issue, in GPflow we can't save models. My suspicion that it also affects TensorFlow Probability project. Can you make an update about it? Can you propose any workaround? Can you advise on how we can contribute to TensorFlow to make this work?
Thanks!

[awav]

hello @k-w-w, @alextp! any updates?

[marcoadurno]

Hello @k-w-w @alextp , any updates on this?

[alextp]

Pinging @rohan100jain for FYI on something that might need a priority bump

[awav]

@alextp thanks!

[nathanhack]

@alextp any idea when this might be fixed?

[alextp]

@ccrusius FYI

[Venkat6871]

Hi,

Thank you for opening this issue. Since this issue has been open for a long time, the code/debug information for this issue may not be relevant with the current state of the code base.

The Tensorflow team is constantly improving the framework by fixing bugs and adding new features. We suggest you try the latest TensorFlow version with the latest compatible hardware configuration which could potentially resolve the issue. If you are still facing the issue, please create a new GitHub issue with your latest findings, with all the debugging information which could help us investigate.

Please follow the release notes to stay up to date with the latest developments which are happening in the Tensorflow space.

[github-actions]

This issue is stale because it has been open for 7 days with no activity. It will be closed if no further activity occurs. Thank you.

[github-actions]

This issue was closed because it has been inactive for 7 days since being marked as stale. Please reopen if you'd like to work on this further.