Issue with Diet Classifier

[Asma-droid]

Hi all,

I would like to use Diet Classifier with rasa version 2.0.0a1 (docker container). If i put 100 epochs, there is no problem generated. If i put epochs equal to 200 or 200. I obtain this message.

?it/s]/opt/venv/lib/python3.7/site-packages/rasa/utils/tensorflow/model_data.py:389: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray

rasa | final_data[k].append(np.concatenate(np.array(v)))

[sara-tagger]

Thanks for raising this issue, @chkoss will get back to you about it soon✨

Please also check out the docs and the forum in case your issue was raised there too 🤗

[chkoss]

@tabergma Can you have a look what causes this warning, please?

I could reproduce the issue independent of the number of epochs. Simply running rasa init --no-prompt gave me the warning mentioned above plus the same warning at a different place:

/Users/christina/rasa/rasa/nlu/classifiers/diet_classifier.py:683: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray
  X_dense_sequence = np.array(X_dense_sequence)