Skip to content

Showing full TensorBoard support in Tensorflow for a CNN using MNIST data.

License

Notifications You must be signed in to change notification settings

krisfur/TensorBoard-CNN-Visualization-Example

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 

Repository files navigation

TensorBoard CNN Visualization Example

This example Convolutional Neural Network using MNIST data was written to showcase usage of the TensorBoard visualisation tools. No functions are defined so code can be easily read from top to bottom, step by step. Made for learning purposes.

Getting Started

Prerequisites

Python 3.x
TensorFlow 1.2.1 or higher 

Doesn't require GPU support, can be done on CPU only. If using GPU support you will need the CUPTI library for TensorBoard to work.

Running the program

Simply execute:

python CNN_TB_Example.py

To train the CNN and get logs saved to mnist_TB_logs folder. To view logs simply run:

tensorboard --logdir=mnist_TB_logs

In the command line/terminal while in the folder with the program. You can naturally replace "mnist_TB_logs" with path to any folder that contains TensorBoard logs.

Note: In Tensorflow 1.5 it is a known issue of TensorBoard not launching properly. Instead use command:

python -m tensorboard.main --logdir=mnist_TB_logs

Built With

Contributing

Feel free to contribute by forking and submitting a merge request.

Authors

  • Krzysztof Furman - Initial work - krisfur

License

This project is licensed under the MIT License - see the LICENSE.md file for details

Acknowledgments

About

Showing full TensorBoard support in Tensorflow for a CNN using MNIST data.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages