As everyone knows, Rasa open source cannot be natively installed on an M1 Mac. I would like to know if there are any solutions now to install Rasa natively on an M1 or at least in Jupyter notebook or Google Colab or any virtual environment like VMware. Kindly help me with this.
Yes, you can use a Virtual Machine, like Ubuntu 20.04 to install Rasa.
Can you elaborate a bit on it? Like how to install it and stuff?
I have installed ubuntu on UTM. The issue persists. When I try to install Rasa, it tries installing 3.0.3 and it gets downgraded again and again and ends up installing rasa 0.0.5. The same happened for me in native installation
@Suganth - Downgrade your python from 3.9 to 3.8.x , the issue isnt your M1 mac but rather a different version of python. Rasa doesn’t yet support 3.9
I use python 3.8.10
Can you show me the output of pip -V
?
This is exactly what happens when you use Python 3.9 or 3.10.
You mean the one from UTM or from my native terminal?
Please use -V
not -v
What about your Ubuntu VM?
That’s weird…
How did you install Rasa on Ubuntu?
I created a python environment and used pip to install rasa
Please show me the commands.
It also looks like you’re not in your virtual environment in your pictures. Do pip -V
in the same environment in which you would use Rasa.
I will do that…but can you please tell me if rasa work on virtual environment like UTM even when on M1?
You mean virtual environment or virtual machine?
The virtual environment stays under your OS, the virtual machine is a whole other independent OS.
I mean a virtual machine. Because M1 is the blocker here, so my question is the virtual machine uses the same M1. So will it work?
@Suganth to answer your previous question, M1 is a microprocessor on your laptop so your virtual machine will also use the same.
I don’t think this is an issue as you described earlier of your M1 chip. Issues with M1 chip is related to tensorflow and there is a separate thread for that.
The fact that you are getting rasa 0.0.5 installed is because that’s the last package where strict Python constraints were not given.
Could you simply make a virtual environment of Python with pyenv or virtualenv, with Python 3.8.x and activate the same virtual environment and install rasa specifying a given version