AttributeError: Exception encountered when calling MonoDense.call(). can't set attribute 'kernel'

[anthonybritto]

Hi,

Thanks for your work on this package. I keep running into the error indicated in the title. To make sure it wasn't just my installation, I ran the collab on your homepage:

https://colab.research.google.com/github/airtai/monotonic-nn/blob/main/nbs/index.ipynb

and the error popped up there too (see below). Would be grateful for a quick fix / any ideas on how to proceed. Thanks!

Epoch 1/10
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
[<ipython-input-6-a9d306e6c0b4>](https://localhost:8080/#) in <cell line: 12>()
     10 model.compile(optimizer=optimizer, loss="mse")
     11 
---> 12 model.fit(
     13     x=x_train, y=y_train, batch_size=32, validation_data=(x_val, y_val), epochs=10
     14 )

3 frames
[/usr/local/lib/python3.10/dist-packages/airt/_components/mono_dense_layer.py](https://localhost:8080/#) in replace_kernel_using_monotonicity_indicator(layer, monotonicity_indicator)
    170     old_kernel = layer.kernel
    171 
--> 172     layer.kernel = apply_monotonicity_indicator_to_kernel(
    173         layer.kernel, monotonicity_indicator
    174     )

AttributeError: Exception encountered when calling MonoDense.call().

can't set attribute 'kernel'

Arguments received by MonoDense.call():
  • inputs=tf.Tensor(shape=(None, 3), dtype=float32)