Skip to content

Latest commit

 

History

History
233 lines (169 loc) · 10.3 KB

README.md

File metadata and controls

233 lines (169 loc) · 10.3 KB
Raw


Marco Lodola, Monument to Umberto Eco, Alessandria 2019

UmBERTo: an Italian Language Model trained with Whole Word Masking

UmBERTo is a Roberta-based Language Model trained on large Italian Corpora. This implementation is based on Facebook Research AI code (https://github.com/pytorch/fairseq)

Description

UmBERTo inherits from RoBERTa base model architecture which improves the initial BERT by identifying key hyperparameters for better results. Umberto extends Roberta and uses two innovative approaches: SentencePiece and Whole Word Masking. SentencePiece Model (SPM) is a language-independent subword tokenizer and detokenizer designed for Neural-based text processing and creates sub-word units specifically to the size of the chosen vocabulary and the language of the corpus. Whole Word Masking (WWM) applies mask to an entire word, if at least one of all tokens created by SentencePiece Tokenizer was originally chosen as mask. So only entire word are masked, not subwords.

Two models are released:

  • umberto-wikipedia-uncased-v1, an uncased model trained on a relative small corpus (~7GB) extracted from Wikipedia-ITA.
  • umberto-commoncrawl-cased-v1, a cased model trained on Commoncrawl ITA exploiting OSCAR (Open Super-large Crawled ALMAnaCH coRpus) Italian large corpus ( ~69GB)

Both models have 12-layer, 768-hidden, 12-heads, 110M parameters (BASE).

Model WWM CASED TOKENIZER VOCAB SIZE TRAIN STEPS FAIRSEQ TRANSFORMERS
umberto-wikipedia-uncased-v1 YES NO SPM 32K 100k Link Link
umberto-commoncrawl-cased-v1 YES YES SPM 32K 125k Link Link

We trained both the models on 8 Nvidia V100 GPUs (p2.8xlarge P2 EC2 instance) during 4 days on AWS Sagemaker.

Installation

Dependencies:

torch >= 1.3.1
sentencepiece
transformers
fairseq

Transformers

pip install transformers

To install transformers from original repo (TESTED):

git clone https://github.com/huggingface/transformers.git
cd transformers
pip install .

Fairseq

To use a version of fairseq with UmBERTo support, build from source doing these steps:

git clone https://github.com/musixmatchresearch/fairseq
cd fairseq
pip install .

Examples

Transformers

From official HuggingFace code.

with transformer pipeline

from transformers import pipeline

fill_mask = pipeline(
	"fill-mask",
	model="Musixmatch/umberto-commoncrawl-cased-v1",
	tokenizer="Musixmatch/umberto-commoncrawl-cased-v1"
)

result = fill_mask("Umberto Eco è <mask> un grande scrittore")

#[{'sequence': '<s> Umberto Eco è considerato un grande scrittore</s>', 'score': 0.1859988570213318, 'token': 5032}, 
#{'sequence': '<s> Umberto Eco è stato un grande scrittore</s>', 'score': 0.1781671643257141, 'token': 471}, 
#{'sequence': '<s> Umberto Eco è sicuramente un grande scrittore</s>', 'score': 0.16565577685832977, 'token': 2654}, 
#{'sequence': '<s> UmbertoEco è indubbiamente un grande scrittore</s>', 'score': 0.09328985959291458, 'token': 17908}, 
#{'sequence': '<s> Umberto Eco è certamente un grande scrittore</s>', 'score': 0.05470150709152222, 'token': 5269}]

with transformer AutoTokenizer,AutoModel

import torch
from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("Musixmatch/umberto-commoncrawl-cased-v1")
umberto = AutoModel.from_pretrained("Musixmatch/umberto-commoncrawl-cased-v1")

encoded_input = tokenizer.encode("Umberto Eco è stato un grande scrittore")
input_ids = torch.tensor(encoded_input).unsqueeze(0)  # Batch size 1
outputs = umberto(input_ids)
last_hidden_states = outputs[0]  # The last hidden-state is the first element of the output

Fairseq

import torch

umberto = torch.hub.load('musixmatchresearch/umberto', 'umberto_commoncrawl_cased')

assert isinstance(umberto.model, torch.nn.Module)
umberto.eval()  # disable dropout (or leave in train mode to finetune)

# Masked LM Inference
masked_line = 'Umberto Eco è <mask> un grande scrittore'
result = umberto.fill_mask(masked_line, topk=20)
# Output:
#('Umberto Eco è considerato un grande scrittore', 0.19939924776554108, ' considerato'), 
#('Umberto Eco è sicuramente un grande scrittore', 0.1669664829969406, ' sicuramente'), 
#('Umberto Eco è stato un grande scrittore', 0.16225320100784302, ' stato'), 
#('Umberto Eco è indubbiamente un grande scrittore', 0.09528309106826782, ' indubbiamente')
...

Results

We obtained state-of-the-art results for POS tagging, confirming that cased models trained with WWM perform better than uncased ones. Our model Umberto-Wikipedia-Uncased trained with WWM on a smaller dataset and uncased, produces important results comparable to the cased results.

Umberto-Wikipedia-Uncased

These results refers to umberto-wikipedia-uncased model.

Part of Speech (POS)

Dataset F1 Precision Recall Accuracy
UD_Italian-ISDT 98.563 98.508 98.618 98.717
UD_Italian-ParTUT 97.810 97.835 97.784 98.060

Named Entity Recognition (NER)

Dataset F1 Precision Recall Accuracy
ICAB-EvalITA07 86.240 85.939 86.544 98.534
WikiNER-ITA 90.483 90.328 90.638 98.661

Umberto-Commoncrawl-Cased

These results refers to umberto-commoncrawl-cased model.

Part of Speech (POS)

Dataset F1 Precision Recall Accuracy
UD_Italian-ISDT 98.870 98.861 98.879 98.977
UD_Italian-ParTUT 98.786 98.812 98.760 98.903

Named Entity Recognition (NER)

Dataset F1 Precision Recall Accuracy
ICAB-EvalITA07 87.565 86.596 88.556 98.690
WikiNER-ITA 92.531 92.509 92.553 99.136

References:

  • BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding Paper, Github
  • CamemBERT: a Tasty French Language Model Paper, Page
  • GilBERTo: An Italian pretrained language model based on RoBERTa Github
  • RoBERTa: A Robustly Optimized BERT Pretraining Approach Paper, Github
  • Sentencepiece: A simple and language independent subword tokenizer and detokenizer for neural text processing Paper, Github
  • Asynchronous Pipeline for Processing Huge Corpora on Medium to Low Resource Infrastructures Paper, Page
  • Italy goes to Stanford: a collection of CoreNLP modules for Italian (TINT) Paper, Github, Page

Credits

All of the original datasets are publicly available or were released with the owners' grant. The datasets are all released under a CC0 or CCBY license.

  • UD Italian-ISDT Dataset Github
  • UD Italian-ParTUT Dataset Github
  • WIKINER Page , Paper
  • I-CAB (Italian Content Annotation Bank), EvalITA Page
@inproceedings {magnini2006annotazione,
	title = {Annotazione di contenuti concettuali in un corpus italiano: I - CAB},
	author = {Magnini,Bernardo and Cappelli,Amedeo and Pianta,Emanuele and Speranza,Manuela and Bartalesi Lenzi,V and Sprugnoli,Rachele and Romano,Lorenza and Girardi,Christian and Negri,Matteo},
	booktitle = {Proc.of SILFI 2006},
	year = {2006}
}
@inproceedings {magnini2006cab,
	title = {I - CAB: the Italian Content Annotation Bank.},
	author = {Magnini,Bernardo and Pianta,Emanuele and Girardi,Christian and Negri,Matteo and Romano,Lorenza and Speranza,Manuela and Lenzi,Valentina Bartalesi and Sprugnoli,Rachele},
	booktitle = {LREC},
	pages = {963--968},
	year = {2006},
	organization = {Citeseer}
}

Acknowledgments

Special thanks to I-CAB (Italian Content Annotation Bank) and EvalITA authors to provide the datasets as part of Master Thesis Research project with School of Engineering, University of Bologna.

Authors

Loreto Parisi: loreto at musixmatch dot com, loretoparisi
Simone Francia: simone.francia at musixmatch dot com, simonefrancia
Paolo Magnani: paul.magnani95 at gmail dot com, paulthemagno

Citation

Cite this work with:

@misc{musixmatch-2020-umberto,
  author = {Loreto Parisi and Simone Francia and Paolo Magnani},
  title = {UmBERTo: an Italian Language Model trained with Whole Word Masking},
  year = {2020},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/musixmatchresearch/umberto}}
}

About Musixmatch AI

Musxmatch Ai mac app icon-128
We do Machine Learning and Artificial Intelligence @musixmatch
Follow us on Twitter Github