/code

Sample code for COS495 homework assignments

Primary LanguageJupyter Notebook

Sample code for COS495 homework assignments

We are providing Jupyter notebooks in Julia and Python.

  • MNIST-Intro. Basic operations with the MNIST dataset of handwritten digits.
  • DeltaRule. Training an LT neuron with the delta rule to recognize "two"s.
  • Backprop. Training a multilayer perceptron with backpropagation to classify digits.
  • Minibatch. Minibatch training of MLP on MNIST
  • ConvolutionBasics. Simple coding exercises for understanding convolution
  • LeNet. Training a convolutional net on MNIST
  • MaxPoolBackprop. Illustration of max-pooling forward and backward passes
  • ToyConvolutionPooling. Toy model illustrating convolution and pooling.
  • CIFAR10. Training a convolutional net on CIFAR-10 using TensorFlow