What is Mojo?
The next new Python? How do I use it?
“What are the best Python implementations?”, I type into Google. I was searching for a Python implementation that can cross-compile my programs to binaries. That was when I came across an implementation I have never came across - Mojo. I started researching on this new language…
What is Mojo?
“Mojo unifies high-level AI development with low-level systems programming. Write once, deploy everywhere - from CPUs to GPUs - without vendor lock-in.”
Mojo is a language developed by Modular that combines benefits of multiple languages such as:
- Python
- C++
- Rust
- Zig
It claims to provide bare-metal performance comparable to that of C++ while holding the flexibility and readability of Python, allowing for faster prototyping and development. Mojo also has native support for NVIDIA and AMD GPUs (Graphical Processing Units).
Applications
Mojo can be used in many places, but primarily is used in developing LLMs (Large Language Models) and other AI models. Its support for GPUs makes it highly suitable replacement for Python in AI training. Mojo also supports existing Python libraries such as Pytorch and Tensorflow. It’s bare-metal performance also gives it an edge over Python. However, Python is not likely to get fully replaced by Mojo. Rather, both Mojo and Python will be used together, just like how C++ and Python are currently used together.
Getting Mojo
Mojo currently only supports MacOS, Linux and WSL (Window Subsystem for Linux) platforms. You can refer to this page in their docs to download and install Mojo. Also, if you are a Windows user like me, you probably have to use WSL or some other virtualization platform. This guide will help you install WSL on your system. From then on, you can continue with the standard installation on bash.
I finally setup my Hyper-V Virtual Machine (Hyper-V is a platform similar to WSL). I run the installation command. “Installation failed. error: …” it returns. AARGH!!! Seems like the installation is not as easy as it seems…