I'm trying to convert a custom Keras model to ml model, via this code below:
##package versions
Keras version: 2.1.2
CoreML = 2.01b
tensorflow -1.10
Import coremltools from keras.models
import load_model
model = load_model('dummy.hdf5')
print(model.summary())
Coreml_model = coremltools.converters.keras.convert(model)
However I'm running into this error:
RuntimeError: Error compiling model: "compiler error: Invalid data blob: 'block1_conv1_output' shape (C,H,W = 32,0,0) for output of layer: 'block1_conv1'
If I run the model (for inference etc) there are no errors, it's just converting it to a coreml model that's outputting the error. So there doesn't seem to be a problem with the model itself.
Here is a dummy model I've created with which I'm having the same problem:
https://drive.google.com/open?id=1gchckYmIeDFCwUAEDXKM-An3K1VQ4DuW
Thank you.
Any suggestions?
Thank you.