What should I build?

Hack Weekend
1

We’re so excited to have you join us for Modular Hack Weekend!

This hackathon is focused on Mojo kernels and MAX Graph model architectures.

Looking for ideas of what to build? You’ve come to the right place.

We would particularly love to see hackers implement:

Any of the following kernels in Mojo:

  • Batched matrix multiplication (BMM)
  • Multi-head latent attention (MLA)
  • Mixture of experts (MOE)
  • Non-maximum suppression (NMS)
  • Grouped matrix multiplication
  • 2D convolutions
  • General matrix-vector multiply (GEMV) on Hopper

Any of the following models using the MAX graph Python API:

5 Likes