đ HattoFlanT5-Large: A Multilingual Model
HattoFlanT5-Large is a multilingual model that extends the vocabulary of Flan - T5 and conducts continual pretraining on a large - scale dataset. It supports multiple languages including Vietnamese, English, and Chinese, and can be applied to various NLP tasks such as summarization, translation, and question - answering.
đ Quick Start
Installation
This model is based on the transformers library. You can install it using the following command:
pip install transformers
Usage
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("Hatto/HattoFlanT5-Large")
model = AutoModelForSeq2SeqLM.from_pretrained("Hatto/HattoFlanT5-Large")
model.cuda()
⨠Features
- Extended Vocabulary: Utilized SentencePiece to retrain a tokenizer for Vietnamese, English, and Chinese, and merged it with Flan - T5's original vocabulary, resulting in a vocabulary of 106611 tokens.
- Continual Pretraining: Conducted single - epoch continual pretraining on a dataset over 100 GB, including news, Wikipedia, books, and legal documents from multiple languages.
- Multilingual Support: Supports Vietnamese, English, and Chinese, and can be used for various NLP tasks such as summarization, translation, and question - answering.
đĻ Installation
This model relies on the transformers library. You can install it via pip:
pip install transformers
đģ Usage Examples
Basic Usage
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
tokenizer = AutoTokenizer.from_pretrained("Hatto/HattoFlanT5-Large")
model = AutoModelForSeq2SeqLM.from_pretrained("Hatto/HattoFlanT5-Large")
model.cuda()
đ Documentation
Extend vocabulary and Pretrain
We used SentencePiece to retrain a tokenizer for Vietnamese, English, and Chinese. The vocabulary of this newly trained tokenizer was then merged with Flan - T5's original vocabulary, removing duplicate tokens. The final merged vocabulary has 106611 tokens.
For single - epoch continual pretraining (also known as incremental pretraining), we used the Flan - T5 - Large model. This pretraining was carried out on a diverse dataset larger than 100 GB, which includes:
- NewsCorpus
- Vietnamese Wikipedia
- Vietnamese books
- Vietnamese legal documents
- Vietnamese legal text
- English Wikipedia
- Chinese Text
Finetune and Benchmark
The model was fine - tuned and benchmarked on the following datasets:
- Wikilingua
- Vietnews
- Pho_NER
- .....
đ License
This project is licensed under the Apache - 2.0 license.
đ Citation
| Property |
Details |
| Model Type |
Flan - T5 - Large with extended vocabulary |
| Training Data |
NewsCorpus, Vietnamese Wikipedia, Vietnamese books, Vietnamese legal documents, Vietnamese legal text, English Wikipedia, Chinese Text |
| Supported Languages |
Vietnamese, English, Chinese |
| Library Name |
transformers |
| Tags |
t5, flant5, summarization, translation, question - answering |
| Pipeline Tag |
fill - mask |
%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
