Skip to content

a tensor flow code to learn a classifier on SVHN dataset

Notifications You must be signed in to change notification settings

scne/svhn_classifier

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

19 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

SVHN classifier

Initialization steps

You have to a system with pyhton3 and pip. I recommend install virtualenv to set an environment on your project folder, then, with command install -r requirements.txt, all dependencies will be installed. For some graph pyhton3-tk is required so I need to install it with apt-get install pyhton3-tk

Dataset

You need to download the dataset (in .mat format) from here and copy all files in svhn folder.

How to

This code is composed by three python files:

  1. load.py reshape and format dataset to using in training and testing mode
  2. dp.py describe network configuration and compute performance
  3. main.py build the network and run training or test phase

be careful

If you run this code with combination of training and extra dataset you need a more then 16gb RAM and 8 core CPU to complete training in about 2 hours

TensorBoard

You can visualize your network trough your browser. You have to activate your virtualenv and run tensorboard command

~/$ source bin/activate
(ProjectName) ~/$  python -m tensorflow.tensorboard --logdir=/path/to/board/folder

About

a tensor flow code to learn a classifier on SVHN dataset

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages