Using an Autoencoder to encode features for k-Means Clustering on the AG News Dataset

The objective of this task is to build a text classification model using the Hugging Face library to classify a dataset of text into one of multiple categories.

Public validation of SRI (Structural Retention) and CI (Collapse Index) metrics on AG News dataset: 90.4% accuracy, 9.2% flip rate, revealing model brittleness beyond standard benchmarks with AUC 0.874!

An analysis of different LLMs' performance on balanced, imbalanced data with different bias ratio

A text classification project combining attention mechanisms with recurrent models (LSTM and GRU) on the AG News dataset using TensorFlow.

This repository contains the code, model configurations, and report for fine-tuning a roberta-base model using Low-Rank Adaptation (LoRA)- a Parameter-Efficient Fine-Tuning (PEFT) method - on the AG News classification task. The goal was to achieve high test accuracy while keeping trainable parameters under 1 million.

This project fine-tunes the DistilBERT model on the AG News dataset for text classification

A complete pipeline for fine-tuning BERT on the AG News dataset using PyTorch 🤖🔥 Includes latent space visualization with t-SNE 🧩📉 and model checkpointing integrated with Google Drive ☁️💾

Recurrent Neural Network (RNN) and Long-Short Term Memory networks (LSTMs) trained to classify text data into 4 categories from the AG news dataset

Predicting the topic of news articles

OpenAI's GPT-2 Fine Tunning Code