Expand description
Fastfood Transform for efficient random feature approximation
This module implements the Fastfood transform, which approximates Gaussian random projections using structured matrices, reducing computational complexity from O(d²) to O(d log d).
Structs§
- Fastfood
Kernel - Fastfood kernel approximation for multiple kernels
- Fastfood
Transform - Fastfood Transform for efficient random Fourier features
Enums§
- Fastfood
Kernel Params - Kernel parameters for Fastfood approximation FastfoodKernelParams