German Text Sentiment Classification

Overview

This project focuses on classifying sentiment in German text using two transformer-based models: BERT and RoBERTa. The models are trained and evaluated on the SB10K dataset, which is accessible from here. The goal is to classify text into three sentiment categories: positive, negative, and neutral.

Model Architecture

The project implements the following models:

• SimpleBERTModel: A BERT-based model for sentiment classification.
• SimpleRoBERTaModel: A RoBERTa-based model for sentiment classification.

Both models are fine-tuned on the SB10K dataset using PyTorch and the transformers library.

Dataset

The dataset used for this project is the SB10K dataset. It can be downloaded from the following link: SB10K Dataset.

Project Structure

• SentimentDataset: A custom dataset class for loading and processing the SB10K data.
• SimpleBERTModel: A custom BERT model for sentiment classification.
• SimpleRoBERTaModel: A custom RoBERTa model for sentiment classification.

Requirements

• torch
• transformers
• pandas
• sklearn

You can install the required packages using:

pip install torch transformers pandas scikit-learn

2. Run the Training: Execute the train.py script to start the training process. This script trains both the BERT and RoBERTa models and evaluates their performance on the test set.

python train.py

Model Diagram