RASA: An Open source conversational AI, which made our work a lot easier. Many say it doesn’t require any programming skills. It is up to you to decide whether you need programming skills or not.
Rasa mainly has 2 main modules,
- Rasa NLU → which understands messages
- Rasa Core → it just decides what to do next.
I guess like me, you must also be not interested to look into theory bla bla bla stuffs, so let’s get started.
Here are other things you need to know Rasa works with the help of TensorFlow.
Here I am using Python version 3.7.3.
Alright the first thing is you need to check whether the Microsoft C++ Build Tools are available in your Local Machine, if not I have given the link for downloading it below,
Next you need to create your virtual environment, if you have any doubt or you don’t remember how to get it working no worries you can refer the link given below,
Link → https://medium.com/@narenltk/why-do-we-need-a-virtual-environment-ec5f4cbb9dc0
If you successfully installed Microsoft C++ Build Tools and created you Virtual Environment then you have reached the final stage, just type in the following command,
pip install rasa
This will install Rasa NLU and Rasa Core with all the required versions. Here you will see you will have the latest version of TensorFlow 2.x.
1. Chatbot Best Practices for designing a Conversational Experience
3. Chatbot software: Dominate the market (The Ultimate Guide)
4. How To Grow Your Business With The Help Of Artificial Intelligence
Now you can verify all this with a pip freeze / pip list command that displays all the lib files that have been installed.
Now that all the work is done, follow the following commands are something to be remembered by everyone,
- Rasa init
- Rasa train
- Rasa run actions
- Rasa run
- Rasa shell
When you run rasa init you will be asked to enter the directory i.e. something you can change or else just press the “enter” button it will just be your current working directory, as shown below.
After selecting the working directory it will ask whether you need to train or not,
Once you’re done with this, you need to give the next command rasa train, by default it will have the model training with 100 Epochs, and few classifiers and you can also use spacy.
Once you train your model, you can open up your own shell and chat with your bot by running rasa shell command,
This will host by default, with a local host with the port number 5005.
You can see there may be few errors dll, i.e. it will be just warnings, still it will run.
If you need any working example to work on then, you can feel free to use my Covid-19 Tracker Bot built using RASA and link has been given below,
Link → Git Repo
Well the output will look like this
If you wish to stay connected, LinkedIn Or just drop a mail to: narenltk@gmail.com, Happy to help..!!!