How to add my own trained BERT + Tensorflow model to RASA

Hello everybody! I have such a question: I want to add my Burt model trained in Russian to the pipeline, but… I don’t understand how to do it (

My model looks like this: Снимок экрана 2020-11-18 в 16.51.04

I would be very grateful for your answer!!

Hi @do8rolyuboff, you first need to convert these tensorflow checkpoints to hugging face compatible models. Once you do that you can pass the converted weights to model_weights parameter of LanguageModelFeaturizer .

Thank you very much for the answer! So I can’t add this model to HFTransformersNLP?

@dakshvar22 Could you please evaluate my config.yml, otherwise it seems to me that I am doing something wrong and LanguageModelFeaturizer does not affect the bot operation in any way