/Deeplearning-in-Mnist

Primary LanguageJupyter Notebook

Deeplearning-in-Mnist

Here deep learning algorithm is applied on MNIST dataset for image classification.

MNIST Dataset

The dataset consists of 70000 images, 60000 images are taken as traning set remaining 10000 images as testing set. All the images are 28*28 with 10 classes (0-9).

Libraries

  • tensorflow
  • keras
  • matplotlib

Activation function

  • relu
  • softmax

Optimizer

  • adam
  • rmsprop
  • nadam

Summary

Here the CNN alogorithm evaluated with different filter values, optimizers. The CNN with specified parameteres, optimizers are effectively increases the accuracy of the model.