How to install the Rasa NLU chatbot framework on windows 11.
What is Rasa?
Rasa NLU is a natural language understanding framework for building interactive chatbots. The special thing about Rasa is that it lets you define actions that should be taken based on user input, effectively letting you make your own voice assistants.
Chances are that if you’re here looking at this installation guide, you already know what Rasa is, whether you should use it or not, and the things it lets you do. In case you don’t know, I have an upcoming story planned where I’ll go in depth on Rasa’s features and build some chatbots with it.
To start off, Rasa has an installation guide on their website but it didn’t work very well for me, hence me making this guide. However, if you’d like to view Rasa’s online installation guide, I’ll link it here. If you’d like to view the video installation guide, which this tutorial will be based on, I’ll link that here.
Anyway, now that I’ve got all that out of the way, let’s get to what you’re actually here for.
Step 1: Install TensorFlow redistributable
First off, since Rasa is based off of TensorFlow, you’ll need to have the TensorFlow C++ Redistributable installed on your machine. Just go to this link and download the version that is compatible with you’re CPU architecture.
Once you’ve downloaded that, follow the installation instructions by clicking on the new application file in your downloads folder. This will include restarting your computer before the software becomes active.
Once you’ve done that, you’re ready for the next step.
Step 2: Install Anaconda
This step isn’t technically required, but for the purposes of this tutorial we’re gonna be using the Anaconda interpreter. If you’d like to see how to complete this process without having to do that, the aforementioned Rasa tutorial is the place to go.
Anyway, go to Anaconda’s download page and get the individual edition.
Once you’ve downloaded that, install it using the install tool in your downloads folder.
Step 3: Creating a Virtual Environment
Creating a virtual environment is also not required, however it is very useful. Rasa is a big package and requires a lot of external dependencies. Initializing a Rasa project in your local computer could cause conflicting versions of dependencies to be downloaded and cause obscure errors. However, if you’d like to install Rasa directly on your machine, go ahead and skip this step.
First, we’re going to use Anaconda to make our lives easier. Now that it’s set up, you should have an Anaconda prompt terminal application.
Run this prompt and use it to navigate to a directory you’re happy creating a virtual environment in. Then run this command, replacing [env_name] with whatever you’d like to call the environment.
After that builds, activate the virtual environment with the following command, replacing [env_name] once again with the name of your virtual environment.
After running this command, you’ll notice on the left of the command prompt that (base) has changed to whatever your [env_name] was, this means you’re in your virtual environment. If you’re in a different prompt then the indication will be different, but it should still work.
Step 4: Installing Rasa
Now that you are in your virtual environment, we can finally install Rasa’s python package. First we need to make sure that we’re using the latest version of Pip by running this command.
This will update Pip to the latest version if it needs to be updated at all. Now that we’ve done that, run this command to actually install the Rasa dependency.
To verify that it installed, run this command.
Once it’s done downloading, you can run this command to initialize a new project in the current directory.
You’re all done!
After completing all of those steps, you should be able to make new chatbots and cool assistants using Rasa. If you want more guidance on how to actually get started after installation, check out Rasa’s YouTube playlist on using Rasa Open Source 3.x for some in-depth tutorials.
I’d be very interested to see what kind of projects people have come up with to utilize Rasa, so share them with me if you’d like! Stay tuned for some more material on Rasa NLU as I further explore it’s capabilities and build some projects with it. As always, if anything needs correction or clarification, let me know.
Happy Coding Everyone!