Slot filling not working for large number of slot values for an intent - Rasa

Hi everyone, I am new in RASA. And I have an issue with slot filling. I have created an intent in nlu file named “learncontent”. This intent will request for 3 slots, out of which slot “chapter” will contains 10k values. How can we manage them in utterances like:- 1- i want to learn 2- i want to learn from english 3- i want to learn algebra from mathematics …


  • Subject
  • Chapter

So, my concern is when i entered “i want to learn algebra from mathematics” the subject and chapter slots is not filling. Slots are suppose to be filled like this : {subject: mathematics} {chapter: algebra}

But again bot reply enter your subject and chapter…and if i enter the subject and chapter again then only it is working.

@ btotharye @ Juste @ arthurTemporim

Hello @rishabhsrivastava, welcome to the Forum!

Many issues could lead to your bot does not replying correctly - it’s not necessarily the fact that the slots are not recognised. Have a look at the docs page about slots: Slots. You may also try to run the NLU server by itself and see if the entities are picked up: Using NLU Only. If that doesn’t help, please let us know what Rasa version you are using.

Hi @j.mosig,

Somehow I have managed to do the slot filling but now I am getting stuck in the slot values. Do I have to give all the slots values (i.e 10k) in the utterances? if yes then how can i add all the slots values in my nlu.


It is working fine in case like:

But if i give same utterance with different slot values like:

Then it is not working because [chemistry] and [carbon] values is not added in utterances. So to handle this case i have to add above slot values in the utterances then only it will work. Do i need to add all 10k slots values in my nlu?

First of all, what you define in the NLU file are entities, not slots. Slots may be filled by entities (see Slots). If the user inputs are always of the same structure, i.e. the entities appear at the same positions, then there is a good chance that entities are picked up even if you don’t have an example for them. Moreover, if you have a list of examples for an entity, you can provide that list as a lookup table (see Training Data Format).

Alright, it’s working fine. Thank you

Hi @j.mosig,

I have successfully set up my bot locally but now as my nlu and models are evolving I want to move rasa to AWS EC2 but I can’t understand what’s the minimum requirements for rasa setup on EC2. Please help me out with the selection of correct EC2 instance type.

In short which EC2 instance is best for rasa setup?

Instance Type:



Network Bandwidth:

EBS Bandwidth:

Hi @rishabhsrivastava, that is a good question.

But could you please ask this question in a new thread? It is completely unrelated to the topic and people won’t find this later if they have the same problem.


1 Like