Scribd
Upload
45 views8 pages

MLP Scratch

The document implements a multilayer perceptron from scratch to classify images from the Fashion-MNIST dataset. It initializes the model parameters, defines the activation and loss functions, trains the model for 10 epochs and evaluates it on the test data.

Uploaded by

damasodra33
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
45 views8 pages

MLP Scratch

The document implements a multilayer perceptron from scratch to classify images from the Fashion-MNIST dataset. It initializes the model parameters, defines the activation and loss functions, trains the model for 10 epochs and evaluates it on the test data.

Uploaded by

damasodra33
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 8

Implementing a Multilayer Perceptron from Scratch

In [1]: %matplotlib inline


import d2l
from mxnet import nd
from mxnet.gluon import loss as gloss
Load the Fashion-MNIST data set
In [2]: batch_size = 256
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size)
Initialize Model Parameters
In [3]: num_inputs, num_outputs, num_hiddens = 784, 10, 256

W1 = nd.random.normal(scale=0.01, shape=(num_inputs, num_hiddens))


b1 = nd.zeros(num_hiddens)
W2 = nd.random.normal(scale=0.01, shape=(num_hiddens, num_outputs))
b2 = nd.zeros(num_outputs)
params = [W1, b1, W2, b2]

for param in params:


param.attach_grad()
Activation Function
In [4]: def relu(X):
return nd.maximum(X, 0)
The model
In [5]: def net(X):
X = X.reshape((-1, num_inputs))
H = relu(nd.dot(X, W1) + b1)
return nd.dot(H, W2) + b2
The Loss Function
In [6]: loss = gloss.SoftmaxCrossEntropyLoss()
Training
In [7]: num_epochs, lr = 10, 0.5
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, batch_size,
params, lr)

epoch 1, loss 0.7868, train acc 0.708, test acc 0.822


epoch 2, loss 0.4831, train acc 0.820, test acc 0.853
epoch 3, loss 0.4295, train acc 0.842, test acc 0.859
epoch 4, loss 0.3930, train acc 0.856, test acc 0.865
epoch 5, loss 0.3663, train acc 0.866, test acc 0.869
epoch 6, loss 0.3520, train acc 0.870, test acc 0.871
epoch 7, loss 0.3368, train acc 0.876, test acc 0.870
epoch 8, loss 0.3236, train acc 0.880, test acc 0.878
epoch 9, loss 0.3129, train acc 0.886, test acc 0.883
epoch 10, loss 0.3067, train acc 0.886, test acc 0.882
Evaluation
In [8]: for X, y in test_iter:
break

true_labels = d2l.get_fashion_mnist_labels(y.asnumpy())
pred_labels = d2l.get_fashion_mnist_labels(net(X).argmax(axis=1).asnumpy())
titles = [truelabel + '\n' + predlabel for truelabel, predlabel in zip(true_labels
, pred_labels)]

d2l.show_fashion_mnist(X[0:9], titles[0:9])

You might also like