dorukkarinca/keras-buoy

Keras wrapper that autosaves what ModelCheckpoint cannot.

keras-buoy's Issues

Make it possible to also use save_weights() instead of save()

Using model.save and the h5 extension, results in an arror for subclass based models.

NotImplementedError: Saving the model to HDF5 format requires the model to be a Functional model or a Sequential model. It does not work for subclassed models, because such models are defined via the body of a Python method, which isn't
safely serializable. Consider saving to the Tensorflow SavedModel format (by setting save_format="tf") or using `save_weights`.

Could you offer an alternative option to use model.save_weights() ?

In the resumable model how to call custom_objects.

resumableModel=ResumableModel(model,save_every_epochs=1,to_path='checkpoint.h5')
value error:Unknown loss function:dice_coef_loss
How do make this work?