mT5-base fine-tuned on TyDiQA for multilingual QA ๐Ÿ—บ๐Ÿ“–โ“

Google's mT5-base fine-tuned on TyDi QA (secondary task) for multingual Q&A downstream task.

Details of mT5

Google's mT5

mT5 is pretrained on the mC4 corpus, covering 101 languages:

Afrikaans, Albanian, Amharic, Arabic, Armenian, Azerbaijani, Basque, Belarusian, Bengali, Bulgarian, Burmese, Catalan, Cebuano, Chichewa, Chinese, Corsican, Czech, Danish, Dutch, English, Esperanto, Estonian, Filipino, Finnish, French, Galician, Georgian, German, Greek, Gujarati, Haitian Creole, Hausa, Hawaiian, Hebrew, Hindi, Hmong, Hungarian, Icelandic, Igbo, Indonesian, Irish, Italian, Japanese, Javanese, Kannada, Kazakh, Khmer, Korean, Kurdish, Kyrgyz, Lao, Latin, Latvian, Lithuanian, Luxembourgish, Macedonian, Malagasy, Malay, Malayalam, Maltese, Maori, Marathi, Mongolian, Nepali, Norwegian, Pashto, Persian, Polish, Portuguese, Punjabi, Romanian, Russian, Samoan, Scottish Gaelic, Serbian, Shona, Sindhi, Sinhala, Slovak, Slovenian, Somali, Sotho, Spanish, Sundanese, Swahili, Swedish, Tajik, Tamil, Telugu, Thai, Turkish, Ukrainian, Urdu, Uzbek, Vietnamese, Welsh, West Frisian, Xhosa, Yiddish, Yoruba, Zulu.

Note: mT5 was only pre-trained on mC4 excluding any supervised training. Therefore, this model has to be fine-tuned before it is useable on a downstream task.

Pretraining Dataset: mC4

Other Community Checkpoints: here

Paper: mT5: A massively multilingual pre-trained text-to-text transformer

Authors: Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel

Details of the dataset ๐Ÿ“š

TyDi QA is a question answering dataset covering 11 typologically diverse languages with 204K question-answer pairs. The languages of TyDi QA are diverse with regard to their typology -- the set of linguistic features that each language expresses -- such that we expect models performing well on this set to generalize across a large number of the languages in the world. It contains language phenomena that would not be found in English-only corpora. To provide a realistic information-seeking task and avoid priming effects, questions are written by people who want to know the answer, but donโ€™t know the answer yet, (unlike SQuAD and its descendents) and the data is collected directly in each language without the use of translation (unlike MLQA and XQuAD).

Dataset Task Split # samples
TyDi QA GoldP train 49881
TyDi QA GoldP valid 5077

Results on validation dataset ๐Ÿ“

Metric # Value
EM 60.88

Model in Action ๐Ÿš€

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
tokenizer = AutoTokenizer.from_pretrained("Narrativa/mT5-base-finetuned-tydiQA-xqa")
model = AutoModelForCausalLM.from_pretrained("Narrativa/mT5-base-finetuned-tydiQA-xqa").to(device)

def get_response(question, context, max_length=32):
  input_text = 'question: %s  context: %s' % (question, context)
  features = tokenizer([input_text], return_tensors='pt')

  output = model.generate(input_ids=features['input_ids'].to(device), 
               attention_mask=features['attention_mask'].to(device),
               max_length=max_length)

  return tokenizer.decode(output[0])
  
# Some examples in different languages

context = 'HuggingFace won the best Demo paper at EMNLP2020.'
question = 'What won HuggingFace?'
get_response(question, context)

context = 'HuggingFace ganรณ la mejor demostraciรณn con su paper en la EMNLP2020.'
question = 'Quรฉ ganรณ HuggingFace?'
get_response(question, context)

context = 'HuggingFace ะฒั‹ะธะณั€ะฐะป ะปัƒั‡ัˆัƒัŽ ะดะตะผะพะฝัั‚ั€ะฐั†ะธะพะฝะฝัƒัŽ ั€ะฐะฑะพั‚ัƒ ะฝะฐ EMNLP2020.'
question = 'ะงั‚ะพ ะฟะพะฑะตะดะธะปะพ ะฒ HuggingFace?'
get_response(question, context)

Created by: Narrativa

About Narrativa: Natural Language Generation (NLG) | Gabriele, our machine learning-based platform, builds and deploys natural language solutions. #NLG #AI

Downloads last month
119
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Dataset used to train Narrativa/mT5-base-finetuned-tydiQA-xqa