Getting error : ValueError: bad input shape (1, 3)

[himanshuteotia]

These are my files screenshots :

[himanshuteotia]

Error file

[JustinaPetr]

@himanshuteotia Strange. Did you try retraining the models? (running nlu_model.py and dialogue_management_model.py scripts)?

[himanshuteotia]

Yes, I tried that

first I ran the 'train_nlu' function then 'run_nlu' as mentioned in docs

[datA2Z]

downgrade scikit-learn 0.20 to 0.19.1. and again train your model. That should solve the problem.

[cobusgreyling]

I had the same error, kept be busy for half a day. Solution for me was:

1. Install Anaconda
2. Created a virtual environment with Python 3.6 (I was 3.7) conda create -n rasa python=3.6.
3. conda activate rasa
4. All worked fine.

So it seem you nee to use Python 3.6 with this list of requirements. But run Anaconda virtual environment. Much easier to manage.

[JustinaPetr]

Yes, you should use Python 3.6 (or earlier) with Rasa Stack for now as some dependencies are not yet updates for Python 3.7. I will add this to the readme :)

[himanshuteotia]

Thanks, guys.. after too many changes it's working now :+1: