Get up to speed on GPU programming with Mojo
All skill levels are welcome in this hackathon, but if you want to get a leg up on the competition, start exploring how to make GPUs go brrr with Mojo:
- Mojo GPU Puzzles: hands-on challenges to sharpen your skills
- Optimize custom ops for GPUs: a focused tutorial to get up and running
- Mojo GPU documentation: a great starting point if you’re new to Mojo on GPUs
Serve custom model architectures using MAX
Build and deploy your own model graphs with MAX:
- Get started with MAX graphs: quick guide to building MAX Graphs in Python
- MAX graph Python API reference: full API docs for MAX Graph
- Build an MLP block as a module
- Serve custom model architectures
- New code example that shows how to define, register, and serve your own model with MAX. This opens up the ability for you to bring your own custom LLM design to MAX, and take advantage of the hardware portability and high-throughput serving capabilities of the framework.
Write hardware-agnostic custom ops for PyTorch
Extend PyTorch with portable, high-performance ops written in Mojo:
We’ve also started open-sourcing some of our internal engineering documents, including a guide that helps map common PyTorch patterns to MAX graph operations and layers.