KaunSourcetype params = Ptree.t = | Tensor of Ptree.tensor| List of params list| Dict of (string * params) listtype module_ = Layer.module_ = {init : 'layout. rngs:Rune.Rng.key ->
dtype:(float, 'layout) Rune.dtype ->
Ptree.t;apply : 'layout. Ptree.t ->
training:bool ->
?rngs:Rune.Rng.key ->
(float, 'layout) Rune.t ->
(float, 'layout) Rune.t;}Efficient dataset handling for machine learning pipelines This module provides composable dataset transformations with support for:
Flax-compatible weight initializers for Kaun
module Checkpoint : sig ... endUnified training state for Kaun.
Optax-inspired optimisation algorithms with explicit checkpoint support.
Activation functions for neural networks.