Skip to main content

Deep Learning for everyone

Project description

Omnis

Deep Learning for Everyone


You have just found Omnis.

Omnis is an API of deep neural network applications, written in Python and capable of running on top of Keras. It was developed with a focus on enabling fast application of deep learning.

Use Omnis if you need a deep learning library that:

  • Is EASY to learn and use.
  • Allows for easy and fast prototyping.
  • Supports CNN, LSTM, GAN, RL application.(TBA)
  • Runs seamlessly on CPU and GPU.

Omnis is compatible with: Python 2.7-3.6.


Deep Block

Omnis has been developed as a backend library of Deep Block. Deep Block is a creative learning community where anyone can learn AI technologies with ease. If you want to learn more about AI or Omnis, try Deep Block.


Getting started: Implement a deep learning application with four lines of code!

The core data structure of Omnis is an application which is designed to be easy to use in each field. Each application consists of a Keras Functional Model.

Here is an Image Classification example with the Caltech 101 dataset:

from omnis.application.image_processing.image_classification.densenet import DenseNet121
import cv2
import numpy

Choose an application:

cnn = DenseNet121()

Prepare data:

cnn.prepare_train_data(get_image_from='directory', data_path='101_ObjectCategories')

After preparing data, you can train your application(model):

cnn.train(epochs = 30, batch_size = 16)

Now you can use the application to classify images:

ant_img = cv2.imread('101_ObjectCategories/ant/image_0013.jpg')
test_data = numpy.expand_dims(ant_img, axis=0)
reshaped_test_data = cnn.reshape_data(test_data)
prediction_result = cnn.predict(reshaped_test_data)

print('predict labels')
print(prediction_result)

For a more in-depth tutorial about Omnis, you can check out:

In the examples folder of the repository, you will find more applications.


Installation

Before installing Omnis, please install Keras. We recommend the TensorFlow backend for Keras.

Then, you can install Omnis itself. There are two ways to install Omnis:

  • Install Omnis from PyPI (recommended):

If you don't use a virtual environment, you can run the command below (not recommended):

sudo pip install omnis

If you are using a virtual environment, you may want to avoid using sudo (conda virtual environment recommended):

pip install omnis
  • Alternatively: install Omnis from the GitHub source:

First, clone Omnis using git:

git clone https://github.com/mkh48v/omnis.git

Then, cd to the Omnis folder and run the install command:

cd omnis
sudo python setup.py install

Guiding principles

  • Simplicity. Omnis pursues a simple architecture. Designing a software with simple architecture not only helps you to understand the code easily but also helps your painful debugging.

  • Easiness. Don't worry about complicated algorithms or theories or mathematics. Omnis will handle difficult stuffs for you. Just learn how to use deep neural networks and USE THIS!

  • Modularity. No spaghetti code!


Why this name, Omnis?

Omnis means EVERY in Latin. The goal of Omnis is to make deep learning technologies easier so that EVERY one can use deep learning technologies without headache.


Project details


Release history Release notifications | RSS feed

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

Omnis-0.0.6.2.tar.gz (15.5 kB view hashes)

Uploaded Source

Built Distribution

Omnis-0.0.6.2-py3-none-any.whl (23.7 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page