inltk 0.0.8

Natural Language Toolkit for Indian Languages (iNLTK)

Natural Language Toolkit for Indian Languages (iNLTK)

Installation

pip install http://download.pytorch.org/whl/cpu/torch-1.0.0-cp36-cp36m-linux_x86_64.whl
pip install inltk

iNLTK runs on CPU and NOT on GPU, as is the desired behaviour for most of the Deep Learning models in production.

The first command above will install pytorch-cpu, which, as the name suggests, does not have cuda support.

Supported languages

Language	Code
Hindi	hi
Punjabi	pa
Sanskrit	sa
Gujarati	gu
Kannada	kn
Malyalam	ml
Nepali	ne
Odia	or
Marathi	mr
Bengali	bn

Usage

Setup the language

from inltk.inltk import setup

setup('<code-of-language>') // if you wanted to use hindi, then setup('hi')

Note: You need to run setup('<code-of-language>') when you use a language for the FIRST TIME ONLY. This will download all the necessary models required to do inference for that language.

Tokenize

from inltk.inltk import tokenize

tokenize(text ,'<code-of-language>') // where text is string in <code-of-language>

Predict Next 'n' words

from inltk.inltk import predict_next_words

predict_next_words(text , n, '<code-of-language>') 

// text --> string in <code-of-language>
// n --> number of words you want to predict (integer)

Note: You can also pass a fourth parameter, randomness, to predict_next_words. It has a default value of 0.8

Repositories containing models used in iNLTK

Language	Repository	Perplexity of Language model	Wikipedia Articles Dataset	Classification accuracy	Classification Kappa score
Hindi	NLP for Hindi	~36	55,000 articles	~79 (News Classification)	~30 (Movie Review Classification)
Punjabi	NLP for Punjabi	~13	44,000 articles	~89 (News Classification)	~60 (News Classification)
Sanskrit	NLP for Sanskrit	~6	22,273 articles	~70 (Shloka Classification)	~56 (Shloka Classification)
Gujarati	NLP for Gujarati	~34	31,913 articles	~91 (News Classification)	~85 (News Classification)
Kannada	NLP for Kannada	~70	32,997 articles	~94 (News Classification)	~90 (News Classification)
Malyalam	NLP for Malyalam	~26	12,388 articles	~94 (News Classification)	~91 (News Classification)
Nepali	NLP for Nepali	~32	38,757 articles	~97 (News Classification)	~96 (News Classification)
Odia	NLP for Odia	~27	17,781 articles	~95 (News Classification)	~92 (News Classification)
Marathi	NLP for Marathi	~18	85,537 articles	~91 (News Classification)	~84 (News Classification)
Bengali	NLP for Bengali	~41	72,374 articles	~94 (News Classification)	~92 (News Classification)