Deep TabNine is a new generation of auto-completion tool which is based on AI. This tool predicts and suggests the ending lines or leftover part of the codes based on the codes written so far. It is considered to be a deep learning-based tool which supports multiple programming languages.
This tool has been developed by Jacob Jackson. He is a computer science student at the University of Waterloo in Canada. Deep TabNine is built on the model called GPT-2 which is trained to predict the next word in 46GB of Internet text. This tool can really help the developers to code fast and will save a lot of time.
More About Deep TabNine & Its model
Deep TabNine’s GPT-2 model which is a large transformer-based language model with 1.5 billion parameters and trained on a dataset of million web pages. GPT-2 model was trained to predict natural code but instead, it predicts the building blocks of code.
How Deep TabNibe Works For:
There are also other similar tools like Deep TabNine is available in the market which includes Microsoft IntelliSense for Visual Studio. But the Deep TabNine autocomplete tool has a major advantage as it has the ability to suggest multiple tokens instead of a single token.
The company is currently providing a normal version of Deep TabNine instead of releasing the model based on deep-learning. It happened because the company is concerned regarding the malicious apps of technology.
Deep TabNine comes with one tradeoff making it not to run efficiently on laptops and the suggestions provided by it will take more time as compared with the standard version. The creator is currently working to create a better model of it so that it runs efficiently on laptops. But for now, he is providing TabNine Cloud beta service which uses GPUs to speed up the process of showing suggestions.