tfann 0.5.0

Tiny Fast Artificial Neural Network Library (Based on SIMD computations)

example/README.md

sine.dart

'''dart import 'dart:convert'; import 'dart:math';

import 'package:tfann/tfann.dart';

void main() { { Random r = Random(); final sine_net = TfannNetwork.full([1, 32, 16, 1], [ActivationFunctionType.fastBell, ActivationFunctionType.uscls, ActivationFunctionType.uscsls]); for (int i = 0; i < 400000; ++i) { var x = r.nextDouble() * 6.29 - 3.145; sine_net.train(TrainSetInputOutput.lists([x], [sin(x)]), learningRate: 0.02); } var x = -3.145; for (int i = 0; i < 1000; ++i) {

  x += 6.29 / 1000;
  print(
      "x= $x : error= ${sine_net.calculateMeanAbsoluteError(
        TrainSetInputOutput.lists([x], [sin(x)]))}");
}

} } '''