This project used in my dissertation for detecting sarcasm in twitter dataset using Sate-of-the-Art Transformer

training sentence sentiment classification with Bert

revising reddit troll hunting with some new nlp techniques (DistilBERT, Multi-Sample Dropout, OOF, Group K-fold on subreddit and some fancy pre-processing)

NLP Service to perform text classification. This is the first part of Project Jarvis. This service integrates to the chat-bot service

Language Detection using DistilBERT

Pytorch-Named-Entity-Recognition-with-transformers

Transformers

BERT (Bidirectional Encoder Representations from Transformers) is a transformer-based method of learning language representations. It is a bidirectional transformer pre-trained model developed using a combination of two tasks namely: masked language modeling objective and next sentence prediction on a large corpus.

NLP deep learning model for multilingual toxicity detection in text 📚

Web Interface for Question Answering System

Kaggle's Tweet Sentiment Extraction challenge. Model had to extract phrases out of a tweet which maximise a given sentiment.

Fake jobs posting predictions task

NLP Workshop -ML India

Compares the DistilBERT and MobileBERT architectures for mobile deployments.

FoodBERT: Food Extraction with DistilBERT

Notebook (Demonstration) for training Distilbert on Glue, and uploading model to Huggingface.

Usage of BERT models for text clustering techniques using sentence embeddings

In this repository, I have collected different sources, visualizations, and code examples of BERT

Model training and prediction algorithm classifying texts per topic and sentiment, using DistilBERT and PyTorch. Uses Github Actions to continuously train and deploy the model to Algorithmia.