Skip to content

Leaderboard

PyTorch

Public profiles only, and only solves made without revealing the reference. Set your profile to private in Settings to stay off these boards.

Fastest submissions

ratio against the reference, lower is faster

Everyone runs on their own hardware, and this board does not adjust for it. A faster machine places higher for the same code.

  1. 1 banutha Self-Attention Layer 0.71x
  2. 2 banutha Implement Scaled Dot-Product Attention 2.1x
  3. 3 Anupa Adikary Matrix Multiplication 2.42x
  4. 4 Anupa Adikary SwiGLU Activation 4.18x

Most solved

PyTorch
  1. 1 Anupa Adikary 2
  2. 2 banutha 2
  3. 3 testadmin 1

Longest streaks

all modes, running now

No streaks running right now.