Hello everyone, I have a question “How can I predict parallelly intent and entities for a batch of text”, not sentence by sentence. I tried to use an Interpreter but it failed. Can someone show me solutions? Thank you a lot!
You can use this endpoint.
Hello @stephens, thanks for your reply, and wish you a nice day! But is there any way to help us predict a batch of text in the local server by Python script? That means I don’t need to run Rasa and listen to sentences through a port. At university, I have a project which demands me to build a system to manage a lot of chatbots for many stores and restaurants. If each chatbot for each store uses one port to get texts from customers, I have to open a lot of ports in the server. My teacher gives me a piece of advice that I should use an Interpreter in Python script to load the NLU model (from rasa.nlu.model import Interpreter) to parse the messages. And she asks me to parse these messages in batches. Can you give me other solutions for this?
I don’t follow the use case you’ve described. This bot doesn’t stay online all of the time? You run do batch runs?
If that’s the case, run the rasa command with the --enable-api option, separately make the API call I references above, then stop the rasa instance.
1 Like