Rasa data validate found no errors. But then when I train I get a conflict between a rule and story. Can this somehow be ignored because it does not matter. The rule always wins so it won’t break anything. Similarly if there are conflicts in stories can these just be warnings. Big data projects often have data conflicts but they don’t fail on training but just ignore them. Also if I fix it then it starts all the training again including NLU training even though none of the NLU data has changed. Is there some way to avoid that?

Can the story conflicts be ignored somehow?

simonm3 (simon) January 17, 2022, 10:20pm 2

Under RulePolicy you just set “check_for_contradictions: False”

1 Like