Transformer models from BERT to GPT-4, environments from Hugging Face to OpenAI. Fine-tuning, training, and prompt engineering examples. A bonus section with ChatGPT, GPT-3.5-turbo, GPT-4, and DALL-E including jump starting GPT-4, speech-to-text, text-to-speech, text to image generation with DALL-E, Google Cloud AI,HuggingGPT, and more

[NAACL 2021] This is the code for our paper `Fine-Tuning Pre-trained Language Model with Weak Supervision: A Contrastive-Regularized Self-Training Approach'.

Transformers 3rd Edition

The purpose of this repository is to introduce new dialogue-level commonsense inference datasets and tasks. We chose dialogues as the data source because dialogues are known to be complex and rich in commonsense.

A quantitative study on over 1.25 million tweets about ChatGPT, employed data scrapping, data cleaning, EDA, topic modeling, and sentiment analysis.

NLP model zoo for Russian

Topic clustering library built on Transformer embeddings and cosine similarity metrics.Compatible with all BERT base transformers from huggingface.

PyTorch implementation of Sentiment Analysis of the long texts written in Serbian language (which is underused language) using pretrained Multilingual RoBERTa based model (XLM-R) on the small dataset.

Sentiment Analysis of tweets written in underused Slavic languages (Serbian, Bosnian and Croatian) using pretrained multilingual RoBERTa based model XLM-R on 2 different datasets.

This repository contains the code of our winning solution for the Shared Task on Detecting Signs of Depression from Social Media Text at LT-EDI-ACL2022.

This app Classifies the text generated by AI tools like chatGPT. Roberta-base-openai-detector Model has been used from hugging face to detect ai generated texts.

Source code for CoNLL 2021 paper by Huebner et al. 2021

This is a project of twitter sentiment analysis using machine learning(Support Vector Machines,Naive Bayes), deep learning(LSTM), Transformer(BERT,ROBERTA).

CatIss is an intelligent tool for automatic categorization of issue reports based on the RoBERTa model.

Convert pretrained RoBerta models to various long-document transformer models

This repository contains the solutions to three problem statements completed during the hackathon. Each problem statement is categorized based on its difficulty level: Easy, Moderate, and Hard.

It is the nlp task to classify empathetic dialogues datasets using RoBERTa, ERNIE-2.0 and XLNet with different preprocessing method. You can get some detailed introduction and experimental results in the link below.

NLP (Natural Language Processing)

Train a T5 model to generate simple Fake News and use a RoBERTa model to classify what's fake and what's real.

Code repository for the paper: How Much Hate with #china? A Preliminary Analysis on China-related Hateful Tweets Two Years After the Covid Pandemic Began