Replies - Activity - asokolow - Rasa Community Forum

How to call Rasa 1.0 via RESTful API?

Be careful it will call the NLU only and the NLU is working with the format ‘{"text":"hello"}’ if you want the complete pipeline (nlu + core), you will have to use the rest/webhook path mentioned by @JiteshGaikwad.

Rasa_nlu interpreter - how to load latest model?

Hi, If you target the parent directory of your models, normally it should load the latest version. Could you try this solution ?

Why my interactive concert bot training only stop when the next action is ActionListen?

For rasa-core, I run python3.6. I’m usually working with Pycharm but for interactive learning, I work in the terminal.

Yes you are right! Is it happening all the time ?

Intent is not being identified

Hi Leonardo, You could probably improve your intent featurizer by using regex features! You should have a look at rasa-nlu.

How to use tracker in a custom class and call custom function in an action class?

Hi there, Does not mean that Action is the parameter of ActionCalculation! it means that ActionCalculation inherits from Action class. The entry point of your ActionCalculation is run(), inside this function you can call your custom function sum().

What do you mean by “stop” ? Ending the dialog flow ?

From my perspective, it seems logical that the bot triggers action_listen after an utter_greet. What does it mean :

You said something like “Hello” (action_listen) The bot replies with its own greeting (utter_greet) The bot is then wait…

Text-to-Speech

If your targeted application is web-based, I suggest to have a look at TensorflowJS instead of Keras! Tensorflow is a bit more complex than Keras but it is worth every penny (a.k.a time)!

Details about conversational-ai-workshop-18

Thank you for the answer … What would be the impact on the training/accuracy if the slots were set to “text” ?

Does the value of an entity influence the action policy?

I agree with you but … I’ve already a severity slot which is set when a user “says” : inform{"severity":"mild"} … In this case, I want to register the information. I had to set autofill to None and rose an event when the intent “inform” is detected … I guess I will have to do the magic trick inside …

Embedding Policy Results

Hi! Great feature! Did you benchmark the training time? I’m currently training a model by using this policy on a GTX 1080ti (12go) + 32Go ram + 32 core CPU and each epoch takes about ~10 min … I’m using :

policies: - name: EmbeddingPolicy epochs: 2000 attn_shift_range: 5

Rasa NLU 0.13.0 is released!

Hey, are the changes available on the github ?

Slot filling "asynchronously"

Thank you, I will have a look now! It seems to be a good approach.

I agree with you, I thought about this too! Thank you, we are on the same track.

Introduce yourself!

Hey there, I’m Alex, I’m working as a R&D engineer in a research center in Belgium (Multitel). I’ve been working for 7 years as a speech processing engineer (recognition and enhancement). I also worked in other fields like biomedical signal processing and image processing. I discovered RASA 6 months ago.