Stanislav Chistyakov comments

Repositories
Issues
Comments

Results 2 comments of


                                            Stanislav Chistyakov

tfrs.models.Model "fit" and "evaluate" Return Loss Values for Last Batch Only

Yes, I agree, there should be created something like `self.loss_tracker = keras.metrics.Mean(name="loss")` for each additional loss in `__init__` of the Model and in `train_step` and `test_step` they should be updated...

LargeList type not supported by tensorflow_io.arrow

please add large_list support it is very popular now a days