The accepted answer says to set the environment variables which made me curious but the problem is I do not know the rasa installation path to set the environment variables.
Step 1:
So I'll write down how I figured this out. First, if you don't have the Anaconda package manager install it from the official website. (While installing click the checkbox to add Anaconda to your PATH environment variable.)
Step 2:
Now open up the anaconda prompt and go to the directory where you want to run rasa
.
Step 3:
Then we can create a new conda
environment by running conda create --name installingrasa python==3.8.5
to keep all of our dependencies together in a centralized place. Finally activate the environment by conda activate installingrasa
Step 4:
Install UJSON and Tensorflow that will help us to work with rasa
.
conda install ujson
conda install tensorflow
Step 5:
Ultimately we can install rasa
. Here we are going to install it via pip
rather than conda
. (there is no conda
version fr rasa at the moment I'm writing this)
pip install rasa
Step 6:
In order to run Tensorflow on windows, we need to download visual c++ separately. Find the executable from the official website. And now we can run rasa init
without errors and initialize new bot.