I did a clean install of rasa on ubuntu and I cannot train the sample model.
rasa init works but only if I choose not to train the model. When I choose to train the model I get the warning,
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
and then a tensor flow error
/usr/local/lib/python3.8/dist-packages/tensorflow/python/framework/ops.py:845 __array__
raise NotImplementedError(
NotImplementedError: Cannot convert a symbolic Tensor (strided_slice_6:0) to a numpy array. This error may indicate that you're trying to pass a Tensor to a NumPy call, which is not supported
My system is as follows;
Rasa Version : 2.5.1 Minimum Compatible Version: 2.2.0a1 Rasa SDK Version : 2.5.0 Rasa X Version : None Python Version : 3.8.5 Operating System : Linux-5.8.0-50-generic-x86_64-with-glibc2.29 Python Path : /usr/bin/python3
Hi, I didnt install tensorflow. It seemed to come with the rasa install. Looks like 2.3.2
python3 -c 'import tensorflow as tf;print(tf.version.VERSION)'
2021-05-10 13:59:26.246762: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library 'libcudart.so.10.1'; dlerror: libcudart.so.10.1: cannot open shared object file: No such file or directory
2021-05-10 13:59:26.246799: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.
2.3.2
@amn41 what should I do from here? I can uninstall rasa and start again from scratch. Do you have a build which you know will work on Ubuntu to train an initial model ?
thanks
Philip
@amn41 OK, good news. I uninstalled everything including any tensorflow libraries I could find. I then created a virtual python enviroment and installed rasa there. This worked as per your steps here,