[Feature Request]: Warn user directly when custom loss is not differentiable

[Frightera]

System information.

TensorFlow version (you are using): TF 2.11
Are you willing to contribute it (Yes/No) : Yes

Describe the feature and the current behavior/state.
Some people write custom loss functions from scratch that is not being differentiable. Then this leads getting None gradients while training. Warning users directly could be a useful feature instead of saying no gradients are provided.

There are way too many Stackoverflow posts about that, I'll include some of them:
https://stackoverflow.com/questions/63874265/keras-custom-loss-function-error-no-gradients-provided
https://stackoverflow.com/questions/73197501/raise-valueerror-no-gradients-provided-for-any-variables-custom-loss-function
https://stackoverflow.com/questions/59292992/tensorflow-2-custom-loss-no-gradients-provided-for-any-variable-error
https://stackoverflow.com/questions/65619581/no-gradients-provided-for-any-variable-for-custom-loss-function
https://stackoverflow.com/questions/70537503/custom-loss-function-error-valueerror-no-gradients-provided-for-any-variable
https://datascience.stackexchange.com/questions/116645/custom-loss-function-for-binary-classificatio-in-keras-gets-error-no-gradients
https://stackoverflow.com/questions/74074934/error-no-gradients-provided-for-any-variable-while-using-custom-loss
https://stackoverflow.com/questions/75738678/gradienttape-returning-none-with-custom-csi-loss-function
https://stackoverflow.com/questions/72259489/valueerror-no-gradients-provided-for-any-variable-custom-loss-function
...

Will this change the current api? How?
This will change the current API by adding some checks on loss function before starting to training, an error/warning can be thrown.

Contributing

• Do you want to contribute a PR? (yes/no): Yes
• Briefly describe your candidate solution(if contributing): Probably add some checks using GradientTape in: https://github.com/keras-team/keras/blob/17af3fcb1d21f950fff097e0534a6ae56bd25a46/keras/engine/compile_utils.py#L104

[tilakrayal]

@Frightera,
Thank you for reporting the issue. We would prefer a PR for the validation, Could you please feel free to submit a PR for the requested change to be made. Thank you!

[Frightera]

The PR is in progress.

[tilakrayal]

@Frightera,
Once PR was raised for the issue, please provide the PR number here and it helps us to track. Thank you!

[github-actions]

This issue is stale because it has been open for 14 days with no activity. It will be closed if no further activity occurs. Thank you.

[Frightera]

The PR number is keras-team/keras#17753

[haifeng-jin]

@Frightera After a brief look at the PR, I wonder if we could generalize this feature to custom metrics and custom losses as well?

[Frightera]

@haifeng-jin Right now it checks the custom losses provided by the user, I am not sure how it would be useful for custom metrics?

[haifeng-jin]

Sorry, only custom layers and losses. Custom metrics do not need to be differentiable.

[Frightera]

@haifeng-jin I'll extend this to support custom layers aswell.

[Frightera]

Hi @haifeng-jin,

Are there any updates on this issue & PR?

Thanks

[tilakrayal]

@Frightera,
I tried to execute the code which was gist provided on the related PR keras-team/keras#17753 on tf-nightly and the keras 3.0 version and it was executed without any error/issue. Kindly find the gist of it here. Thank you!

[github-actions]

This issue is stale because it has been open for 14 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 28 days. Please reopen if you'd like to work on this further.