Multilabel Text Sequence Classification

This is a repository with the assignments of IE678 Deep Learning course at University of Mannheim.

A visual debugging tool for fine-tuned BERT models for (multilabel) sequence classification tasks

A useful repository for calculating classification baselines using Bert

NLP Project - Sentence Classification - Toxicity- Approx 20,000 comments - ranging from 2 to 30 words. Balanced Data Set. 1. Traditional, pre-2010 NLP and ML techniques used. 2. Dense Word Vectors - w2v & Glove, sentence vector created from averaged word vectors, ANN. 3. Glove combined with bi-LSTMs and 2D Convs.

Second place solution for McKinsey Analytics Hackathon

This web based application takes a user input and shows the sentiment and polarity scores of the user's input.

Predict intent from user queries

In this project we are going to create a network to predict a category for the sequence.

Developed system for the "Climate Activism Stance and Hate Event Detection" Shared Task at CASE 2024

Implementation of BERT for sequence classification with Hugging face's transformers.

A workflow for processing metagenomics data sets using Centrifuge. Category:Single-Sample

music genre classification using 2D CNN, 1D CNN - LSTM and Librosa

Fine-tuning CamemBERT for French keywords extraction on custom dataset.

This is the repository to my bachelor's thesis. The aim was to classify antisemitic comments. A custom dataset was created with the 4chan API and labeled. Different DistilBERT models for sequence-classification were trained and compared. A case study was then conducted.

Saccharomyces cerevisiae information, gene calling and ORF identification assistant

VOGUE: Variable Order HMM with Duration

Encode and decode hand movements using C# and Kinect. Based on K-means and Hidden Markov Models.