README.md

---
language:
- id
- en
- multilingual
license: apache-2.0
tags:
- idt5
---

# Indonesian Version of Multilingual T5 Transformer

Smaller version of the [Google's Multilingual T5-base](https://huggingface.co/google/mt5-base) model with only Indonesian and some English embeddings.

This model has to be fine-tuned before it is useable on a downstream task.\
Fine-tuned idT5 for the Question Generation and Question Answering tasks, available at [idT5-qa-qg](https://huggingface.co/muchad/idt5-qa-qg).

Paper: [idT5: Indonesian Version of Multilingual T5 Transformer](https://arxiv.org/abs/2302.00856)

Authors: *Mukhlish Fuadi, Adhi Dharma Wibawa, Surya Sumpeno*

## Citation
```
@misc{https://doi.org/10.48550/arxiv.2302.00856,
  doi = {10.48550/ARXIV.2302.00856},  
  url = {https://arxiv.org/abs/2302.00856},  
  author = {Fuadi, Mukhlish and Wibawa, Adhi Dharma and Sumpeno, Surya},  
  keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences, I.2.7},  
  title = {idT5: Indonesian Version of Multilingual T5 Transformer},  
  publisher = {arXiv},  
  year = {2023}
}
```

## Abstract
Indonesian language is spoken by almost 200 million people and is the 10th most spoken language in the world, but it is under-represented in NLP (Natural Language Processing) research. A sparsity of language resources has hampered previous work on Indonesian. The Transformer is a new architecture rapidly becoming dominant for NLP, surpassing alternatives like convolutional and recurrent neural networks. T5 (Text-to-Text Transfer Transformer) is a Transformer model that converts all text-based language problems to text-to-text format for English. The multilingual variant is mT5 (multilingual T5) which has shown promising results on many NLP tasks across languages. However, the size of this multilingual model is a drawback for its application in real production applications, which sometimes require only one language. In this study, the mT5 model was adapted for only one language, Indonesian, resulting in a pre-trained T5 model that was specific only for Indonesian with a smaller size. For performance comparison, we fine-tuned this model and the mT5 model to the Sentiment Analysis (SA), Question Generation (QG), and Question Answering (QA) tasks with the exact mechanism and dataset. Fine-tuned model based on our model achieved 77.18% accuracy on SA, 8% higher than the mT5-based model, and obtained nearly the same score as the mT5-based model on QG and QA. The results confirm that it is possible to produce a smaller pre-trained model that maintains comparable yields while reducing the model size by up to 58%. In addition, the resulting model requires less memory, loads faster, and inference times faster.