Machine Learning Model for Textualize

This repository contains the machine learning model used for Textualize, a document summarization application. The model is fine-tuned using the t5-small pre-trained model and utilizes the "alexfabbri/multi_news" dataset from Hugging Face for training.

Model Information

• Model Type: t5-small
• Fine-tuning Dataset: alexfabbri/multi_news
• File Format: The model is saved as an .h5 file in this repository.

Performance

The model was trained and evaluated with the following performance metrics:

• Training Accuracy: 0.9048
• Validation Accuracy: 0.9227

Note: These accuracy values are based on sequence-level metrics and may vary depending on evaluation setup.

Fine-Tuning Documentation

For more detailed documentation on how the model was fine-tuned, you can refer to the following resources:

• Fine-tuning Model Notebook
• Fine_tuning_model_t5_small_with_multi_news_dataset.ipynb file in this repository

Files

• Important Model File: .h5 file located in the repository.