Skip to content

mariusdanciu/snn

Repository files navigation

SNN - Simple Neural Network

A Neural Network library written in Rust.

Test 1 Training on MNIST data set with following configuration:

  1. Two hidden layers with 50 activation functions (RELU)
  2. Softmax classifier
  3. Minibatch size 200
  4. Learning rate 0.05
  5. No regularization
  6. Minibatch optimizer

train-test curve

Test 2 Training on MNIST data set with following configuration:

  1. Two hidden layers with 50 activation functions (RELU)
  2. Softmax classifier
  3. Minibatch size 200
  4. Learning rate 0.01
  5. No regularization
  6. Adam optimizer

train-test curve

Test results

  • With Adam optimizer we see ~95% accuracy on ~385 iterations
  • Without Adam we see ~ 93% accuracy after 1200 iteration.

Thus with Adam optimizer we observe a reach to a slightly higher accuracy ~3.1 times faster.

The library supports:

  1. L2 regularization
  2. Minibatch, Momentum, RMSProp and Adam optimizations
  3. Only Dense layers so far.
  4. No GPU acceleration yet

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages