About Logo

Research Papers in AI

Most research papers nowadays publish the source code to allow reproducibility. This is great, yet often, source code does not have an API for inference. This means you cannot run the model with a custom string test input or make it part of your production system. Consequently, anyone who would like to use the proposed method must go through the source code and even design an API for inference. This cumbersome process needs to be repeated for any further papers. Moreover, there was no single, complete solution for Turkish NLP in Python.

The Solution: VNLP Library

VNLP comes to the stage here. It implements the state of the art papers, offering lots of functionality to be the single complete Turkish NLP solution. VNLP provides Python and CLI APIs for inference. It is very easy to install and use, taking just a few lines of code.


You can install VNLP via pip.

 > pip install vngrs-nlp      


Implemented classes, facilitated papers, methods, datasets, training procedure, and evaluation metrics are described in detail on the documentation page.

How it works?

Engineering Perspective

VNLP implements state-of-the-art AI papers in an efficient way to solve fundamental NLP problems. It leverages TensorFlow and NumPy to implement deep neural networks and functions. Its compact models are trained on large datasets and provide lightweight solutions. VNLP is developed in Python and supports Python and CLI APIs.



Named Entity Recognition

Stemmer: Morphological Disambiguator

Dependency & Part of Speech

Text Summarization

Sentiment Analysis

Sentence Splitting

Spelling & Typo Correction

Pre-trained Word Embeddings

Convert Numbers to Words


Stop Word Removal


Try the Demo

As VNGRS, we created a Demo page to demonstrate the capabilities of VNLP. Click the button below to try it yourself.

Try our demo