Neural Tangents Reference

Neural Tangents is a set of tools for constructing and training infinitely wide neural networks.

Topics:

• Stax – infinite networks (NNGP, NTK)
  • Combinators
  • Branching
  • Linear parametric
  • Linear nonparametric
  • Elementwise nonlinear
  • Helper enums
  • For developers
• Empirical – finite NNGP and NTK
  • Kernel functions
  • Linearization and Taylor expansion
• Predict – inference with NNGP and NTK or linearized networks
  • Prediction / inference functions
  • Utilities
  • Helper classes
• Batching – using multiple devices
• Monte Carlo Sampling

Indices and tables

• Index

• Module Index

• Search Page