Glossary · Term

Performer

← all terms

Definition

A Transformer variant that approximates attention with random feature kernels for linear-time scaling.

A kernel-approximation efficient-attention architecture using positive random features, recombined as a building block in agent-designed LRA solutions.

Mentioned in 1 episode

  1. 053
    An AI Agent Swapped In Focal Loss And Beat A Human-Tuned Training Script