Neel Kovelamudi

Results 25 comments of Neel Kovelamudi

I have added a section called "Batch-level logging" to a tutorial in the Tensorboard Github: https://github.com/tensorflow/tensorboard/blob/master/docs/scalars_and_keras.ipynb I think the more common use case is addressed in the section titled "Instantaneous...

I will defer to @G4G on this, as he is the owner of the TF graphics repo.

Sorry for the delay from my end. One thing I would like clarification on is why layers are being instantiated in the `build()` method instead of the `__init__()`? If those...

Thank you for the deep dive @akeeman ! This issue is tricky, since that commit [0454a40](https://github.com/keras-team/keras/commit/0454a403867cd4ba23bf2604f1feaa36c357db17) is a actually a fix for layer index naming. The naming should be `lstm`,...

@fchollet Yes, this is fixed. Closing the issue, thanks!

Hi @innat, this would be a good addition as a Callback! Would you be willing to contribute a PR to add this Profiler Callback?

Hi @innat, this sounds like an exciting idea! Would you be willing to contribute a PR to add this op to Keras 3?

@fchollet Do you have a preference for video decoding? ([decord](https://github.com/dmlc/decord), opencv, [av](https://github.com/PyAV-Org/PyAV))

Hi @lbortolotti, It seems that your workflow with jax2tf and TF SavedModel is exactly what the `model.export()` API is useful for. You can simply do `model.export("my_model")` or have more customization...

Yes, `model.export()` does turn a JAX-backend Keras model into a TF SavedModel. The API is not as mature for JAX-backend as it is for TF (hence why docs have not...