package tiny_libs

  1. Overview
  2. Docs
From-scratch libraries for teaching: graphics, audio, compression, crypto, networking and more

Install 

dune-project
 Dependency

github.com Readme Edit opam file Versions (1)

Authors

  1. Y
    Yoann Padioleau <yoann.padioleau@gmail.com

Maintainers

  1. Y
    Yoann Padioleau <yoann.padioleau@gmail.com

Sources

0.3.6.tar.gz
md5=7c636383d146d30ac6f2fa234a6253c8
sha512=c79f3823c5f8f57e5038eb640d487c61168b84aa07c61999d6622ef9fd0c890e2b03b4c6a7cdbbe9352a49e25dda00ac7bb14693cee8e3d7beeed251351a2af0

doc/tiny_libs.ai_learning/Train/index.html

Module TrainSource

Sourcetype history = {
  1. epoch : int;
  2. training : float;
  3. held_out : float;
}
Sourceval split : ?part:float -> Backprop.example list -> Backprop.example list * Backprop.example list
Sourceval epoch : ?seed:int -> ?rate:float -> ?batch:int -> Net.t -> Backprop.example list -> Net.t
Sourceval run : ?epochs:int -> ?rate:float -> ?decay:float -> ?batch:int -> ?seed:int -> ?held:Backprop.example list -> Net.t -> Backprop.example list -> Net.t * history list
Sourceval accuracy : Net.t -> Backprop.example list -> answer:(float array -> int) -> float
Sourceval best : float array -> int
Sourceval one_hot : int -> int -> float array