TFlearn classification example using MNIST dataset

TFlearn is a python deep learning library which is running on top of the Tensorflow. This library provides really easy to way implement deep neural networks and visualization with Tensorboard.

1
pip install tflearn

More information: http://tflearn.org/

The downloadable jupyter notebook has been provided below. The notebook can be downloaded by clicking on the “view raw” text at the end of the notebook.

Tensorboard visualization can be viewed using the following command. The log file is saved in “/tmp/tflearn_logs” in Linux systems.

1
tensorboard --logdir=/tmp/tflearn_logs/

The following figure shows the structure of the deep neural network

 

Variation of accuracy can be seen as follows.

Add Comment

Required fields are marked *. Your email address will not be published.

Loading Facebook Comments ...