## Rasa.core.policies.ted_policy predicting wrong action

I’m trying rasa with example on [Generating NLU Data (rasa.com)](/content/docs/rasa/generating-nlu-data/#splitting-on-entities-vs-intents/index.html) and I’ve written appropriate nlu data and domain yml to go along with stories. So, intents and entities (there is only one for now: product) are correctly classified and identified, however, wrong action is being predicted by TEDPolicy.

## Extracting Entities from Intents

Is this forum helpful to anybody?

## Introduce yourself

Hi, this is Jawad from Pakistan. I am ML/DL researcher and developer and build applications for audio and text analysis. But for a personal project I am learning rasa, perhaps I can deploy it on a GPU poor environment if I’m comfortable enough with it. For now however am having trouble with implemen…