FlauBERT: Unsupervised Language Model Pre-training for French

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 2479–2490 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 2479 FlauBERT: Unsupervised Language Model Pre-training for French Hang Le1 Lo¨ıc Vial1 Jibril Frej1 Vincent Segonne2 Maximin Coavoux1 Benjamin Lecouteux1 Alexandre Allauzen3 Benoˆıt Crabb´e2 Laurent Besacier1 Didier Schwab1 1Univ. Grenoble Alpes, CNRS, LIG 2Universit´e Paris Diderot 3E.S.P.C.I, CNRS LAMSADE, PSL Research University {thi-phuong-hang.le, loic.vial, jibril.frej}@univ-grenoble-alpes.fr {maximin.coavoux, didier.schwab, benjamin.lecouteux, laurent.besacier}@univ-grenoble-alpes.fr {vincent.segonne@etu, bcrabbe@linguist}.univ-paris-diderot.fr, alexandre.allauzen@espci.fr Abstract Language models have become a key step to achieve state-of-the art results in many different Natural Language Processing (NLP) tasks. Leveraging the huge amount of unlabeled texts nowadays available, they provide an efficient way to pre-train continuous word representations that can be fine-tuned for a downstream task, along with their contextualization at the sentence level. This has been widely demonstrated for English using contextualized representations (Dai and Le, 2015; Peters et al., 2018; Howard and Ruder, 2018; Radford et al., 2018; Devlin et al., 2019; Yang et al., 2019b). In this paper, we introduce and share FlauBERT, a model learned on a very large and heterogeneous French corpus. Models of different sizes are trained using the new CNRS (French National Centre for Scientific Research) Jean Zay supercomputer. We apply our French language models to diverse NLP tasks (text classification, paraphrasing, natural language inference, parsing, word sense disambiguation) and show that most of the time they outperform other pre-training approaches. Different versions of FlauBERT as well as a unified evaluation protocol for the downstream tasks, called FLUE (French Language Understanding Evaluation), are shared to the research community for further reproducible experiments in French NLP. Keywords: FlauBERT, FLUE, BERT, Transformer, French, language model, pre-training, NLP benchmark, text classification, parsing, word sense disambiguation, natural language inference, paraphrase. 1. Introduction A recent game-changing contribution in Natural Language Processing (NLP) was the introduction of deep unsuper- vised language representations pre-trained using only plain text corpora. Previous word embedding pre-training ap- proaches, such as word2vec (Mikolov et al., 2013) or GloVe (Pennington et al., 2014), learn a single vector for each wordform. By contrast, these new models are trained to produce contextual embeddings: the output representation depends on the entire input sequence (e.g. each token in- stance has a vector representation that depends on its left and right context). Initially based on recurrent neural net- works (Dai and Le, 2015; Ramachandran et al., 2017; Howard and Ruder, 2018; Peters et al., 2018), these mod- els quickly converged towards the use of the Transformer (Vaswani et al., 2017), such as GPT (Radford et al., 2018), BERT (Devlin et al., 2019), XLNet (Yang et al., 2019b), RoBERTa (Liu et al., 2019). Using these pre-trained mod- els in a transfer learning fashion has shown to yield strik- ing improvements across a wide range of NLP tasks. One can easily build state-of-the-art NLP systems thanks to the publicly available pre-trained weights, saving time, energy, and resources. As a consequence, unsupervised language model pre-training has become a de facto standard in …

Content truncated to 500 words