NeuralNetworkBuilder

{ **********************
{ Autor: @GabriM97
{ Telegram: @GabriM97
{ Email: marino.gabri97@gmail.com
{ *********************

Neural Network Builder is a website built using Laravel (PHP Framework) and Keras (ML Python Library), that allows you to easily build your Neural Network Models and train it on your personal dataset.
I promise you will not write a single line of code, just Import a Dataset -> Build the Model -> Train your Model -> Download your Trained Model. It's easy.

Requires:

• root permissions (sudo)
• MySQL (create database: neural_network_builder admin:admin)
• Apache --> read deploy-on-apache-server file
• PHP
• python3
• pip3

Dependences:

pip3 install (install them globally: `sudo -H pip3 intall ):

• tensorflow
• keras
• pandas
• supervisor

Supervisor:

Into the directory utils/supervisor/ you can find supervisor configuration files. After installing supervisor via pip3, copy the config files (both laravel-worker.conf and supervisord.conf) into /etc/supervisor/ directory.

NOTE: Edit the config file laravel-worker.conf with your correct paths and username (if you got errors on supervisor, you probably have to edit the /path/to/supervisor.sock path, for supervisor socket, into the supervisord.conf file)

To deploy the app read 'deploy-on-apache-server file' first. Then run the following commands:

• composer install
• npm install
• ./start.sh
• php artisan migrate:fresh

NOTE: Edit the database settings (database name and user) into .env file.

NOTE: Modify the upload_max_filesize and post_max_size on your php.ini file if you want to upload a bigger dataset. (Linux: /etc/php.ini)

NOTE: Edit the utils/start.sh script with your correct paths.

NOTE: If running the script start.sh you got an error (such as unix:///path/to/supervisor.sock no such file) on supervisroctl tab then don't worry, it will work the same! You can test if everything it's okay running status command on supervisorctl tab. If it shows the laravel-workers are in RUNNING or STARTING status, then it works. If not, try running reload command and then status again. If you still got this issue or others, then check your supervisor config files or try to unistall and install again.