jiro_nn 0.5.0

Neural Networks framework with model specification & data preprocessing features.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39

#![doc = include_str!("../README.md")]

#[macro_use]
extern crate assert_float_eq;

/// Monitoring performance of tasks and logging
pub mod monitor;
/// Activation functions and abstractions (sigmoid, relu, softmax...)
pub mod activation;
/// Model performance benchmarking utilities
pub mod benchmarking;
/// Dataset specification (metadata, preprocessing flags...)
pub mod dataset;
/// Wrapper around dataframes libaries
pub mod datatable;
/// Parameters initializers and abstractions (uniform, glorot...)
pub mod initializers;
/// Layers and abstractions (dense, full...)
pub mod layer;
/// Learning rate schedulers and abstractions (constant, exponential...)
pub mod learning_rate;
/// Basic linear algebra backends and wrappers (matrix, vector, scalar...)
pub mod linalg;
/// Loss functions and abstractions (mse, crossentropy...)
pub mod loss;
/// Model specification
pub mod model;
/// Neural network abstractions
pub mod network;
/// Optimizers and abstractions (sgd, adam...)
pub mod optimizer;
/// Preprocessing and pipelining utilities (normalization, one-hot encoding...)
pub mod preprocessing;
/// Training methodologies (k-fold, split...)
pub mod trainers;
/// Utilities for `Vec<Scalar>`, `Vec<Vec<Scalar>>`...
pub mod vec_utils;
/// Everything vision (CNNs, images operations & backends...)
pub mod vision;