/how_to_image_classifier

How to design, implement, test, and deploy image classifiers, with deep neural networks.

Primary LanguageShellMIT LicenseMIT

README

How to design, implement, test, and deploy image classifiers, with deep neural networks.

What is this repository for?

This is project based learning, not research

Current version 0.1 (20220707)

How do I get set up? (Linux)

  1. Go to your home directory, make Apps directory, change directory

    $ cd ~; mkdir Apps && cd Apps

  • Clone this repo

    $ git clone git@bitbucket.org:daniel_walther_berns/how_to_image_classifier.git

  • Change directory

    $ cd how_to_image_classifier/

  • Run start.sh

    $ ./start.sh

The bash script start.sh

  1. creates a python3 environment ditdic (design, implement, test and deploy image classifiers);
  2. installs some dependences;
  3. installs the ditdic package;
  4. creates directory ~/Data/how_to_image_classifier/v_0_1

Expected results

Created directories

~/Apps/how_to_image_classifier

~/.venvs/ditdic

~/Data/how_to_image_classifier/v_0_1

Documents

In the ~/Apps/how_to_image_classifier/docs directory,

$ cd ~/Apps/how_to_image_classifier/docs

you will find documents related to data, software, hardware, installation, and operations of this project.

Source

In the ~/Apps/how_to_image_classifier/source directory,

$ cd ~/Apps/how_to_image_classifier/source

you will find a set of jupyter notebooks, the package ditdic and a README.md.

You can use the notebooks and the package ditdic following the next instructions:

  1. Make sure you have executed ~/Apps/how_to_image_classifier/start.sh
  2. Activate the environment $ . ~/.venvs/ditdic/bin/activate
  3. Change directory $ cd notebooks/
  4. Run jupyter $ jupyter notebook

Configuration

Write

Who do I talk to?