How to evaluate core Model

Rasa Open Source
1

HI

I have trained 2 models with memorization and keras policy with max histoty as 10 and max history as 5 but some of the stories working in model 1 is not working in model 2 and vice-versa so i got to know about core model evaluation from which i can find optimum value for hyper parameters.But when i give following command it tells me it will support only one policy (i have both keras and memorization policy) in my config file .should i have to remove any one of them.Please let me know.How can i find proper values for hyper parameters using model evaluation.

rasa train core -c config.yml config_2.yml
-d domain.yml -s data --out comparison_models --runs 3
–percentages 0 5 25 50 70 95

2

what version are you using? We have fixed it in latest version, please take a look at migration guide, cli commands have changed

3

Thanks

My rasa version is Rasa 1.1.3

Can you please provide that migration guide link

Also please let me know commands from which i can find version for NLU and core separately

4

here is migration guide: Migration Guide

Rasa NLU and Core are merged into one package now