qubitai-dltk: A Python repository from Mainak-C

About

Our philosophy is to create a Deep Technologies platform with ethical AI for enterprises that offers meaningful insights and actions.

DLTK Unified Deep Learning platform can be leveraged to build solutions that are Application-Specific and Industry-Specific where AI opportunity found by using DLTK SDKs, APIs and Microservices. With best of the breed AI Services from platform pioneers like H2O, Google's TensorFlow, WEKA and a few trusted open-sources models and libraries, we offer custom AI algorithms with co-innovation support.

Getting Started

Pre-requisite

OpenDLTK : OpenDLTK is collection of open-source docker images, where processing of images, text or structured tabular data is done using state-of-the-art AI models.

Please follow the below link for instructions on OpenDLTK Installation

Installation

Installing through pip

pip install qubitai-dltk

Installing from Source

Clone the repo

git clone https://github.com/dltk-ai/qubitai-dltk.git

Set working directory to qubitai-dltk folder

cd qubitai-dltk

Install requirements from requirements.txt file

pip install -r requirements.txt

Usage

A detailed documentation is present here, on how to use various services supported by DLTK, to verify whether all setup are done properly, we will be using a sample NLP code to analyze sentiment of the input text.

Example

import dltk_ai
client = dltk_ai.DltkAiClient(base_url='http://localhost:8000')

text = "The product is very easy to use and has got a really good life expectancy."

sentiment_analysis_response = client.sentiment_analysis(text)

print(sentiment_analysis_response)

Important Parameters:

APIkey : a valid API key generated by following steps as shown here

base_url : The base_url is the url for the machine where base service is installed. (default: http://localhost:8000)

Expected Output

{
  "nltk_vader": {"emotion": "POSITIVE", "scores": {"negative": 0.0, "neutral": 0.653, "positive": 0.347, "compound": 0.7496}}
}

Services

Machine Learning

ML Scikit - This Microservice uses widely used Scikit package for training and evaluating classification, regression, clustering models and other ML related tasks on dataset provided by user.

ML H2O - This Microservice uses H2O.ai python SDK for training and evaluating classification, regression, clustering models and other ML related tasks on dataset provided by user.

ML Weka - This Microservice uses WEKA for training and evaluating classification, regression, clustering models and other ML related tasks on dataset provided by user.

Example Notebooks

ML Classification Colab Notebook

ML Regrression Colab Notebook

ML Clustering Colab Notebook

Natural Language Processing (NLP)

This microservice provides features like Sentiment analysis, Name Entity Recognition, Tag Extraction using widely used Spacy and NLTK package. It also provide support for various AI engines like Azure & IBM.

Example Notebook

NLP Colab Notebook

Computer Vision

Image Classification - This microservice classify images into various classes using pretrained model and also using supported AI Engines.

Object Detection - This microservice detect objects in Images provided by user using pretrained model and using supported AI Engines.

Example Notebooks

Image Classification Colab Notebook

Object Detection Colab Notebook

Face Analytics Colab Notebook

Note

To use third party AI engines like Microsoft Azure & IBM watson, please ensure that its credentials were configured while setting up openDLTK.