I have deployed rasa-x using:
curl -s get-rasa-x.rasa.com | sudo bash
then I connected my git repo having rasa assistant to rasa-x
everything completes successfully.
But when I train my model for the first time, it keeps on training and never finishes.
The same rasa bot was trained within 2 minutes on local machine.
@Sajjadmanal please tell are you using docker based installation? if yes please check cd /etc/rasa or if you created your own folder and in that folder you will see model, check is there any trained model?
@nik202 I just restarted my aws machine instance and then redeployed using curl -s get-rasa-x.rasa.com | sudo bash
This time, model was trained successfully but when I go to chat with bot, I get nothing on screen.
I checked the logs for the running rasa sdk and it showed no module named numpy.
Could this be a reason. How to install all requirements file in the running containers?
@Sajjadmanal mean you are using quick installation for rasa x. Well, I personally not used the quick install, did you see some folder in the root which is related to rasa?
@Sajjadmanal try to install in /etc/rasa itself and then check pip list later, even before installing check that requirement is there in the list or not.
@Sajjadmanal can you even share the docker-compose.override.yml to me?