README.md

---
language: en
license: apache-2.0
datasets:
- bookcorpus
- wikipedia
---

# BERT large model (cased) whole word masking finetuned on SQuAD

Pretrained model on English language using a masked language modeling (MLM) objective. It was introduced in
[this paper](https://arxiv.org/abs/1810.04805) and first released in
[this repository](https://github.com/google-research/bert). This model is cased: it makes a difference between english and English.

Differently to other BERT models, this model was trained with a new technique: Whole Word Masking. In this case, all of the tokens corresponding to a word are masked at once. The overall masking rate remains the same.

The training is identical -- each masked WordPiece token is predicted independently. 

After pre-training, this model was fine-tuned on the SQuAD dataset with one of our fine-tuning scripts. See below for more information regarding this fine-tuning.

Disclaimer: The team releasing BERT did not write a model card for this model so this model card has been written by
the Hugging Face team.

## Model description

BERT is a transformers model pretrained on a large corpus of English data in a self-supervised fashion. This means it
was pretrained on the raw texts only, with no humans labelling them in any way (which is why it can use lots of
publicly available data) with an automatic process to generate inputs and labels from those texts. More precisely, it
was pretrained with two objectives:

- Masked language modeling (MLM): taking a sentence, the model randomly masks 15% of the words in the input then run
  the entire masked sentence through the model and has to predict the masked words. This is different from traditional
  recurrent neural networks (RNNs) that usually see the words one after the other, or from autoregressive models like
  GPT which internally mask the future tokens. It allows the model to learn a bidirectional representation of the
  sentence.
- Next sentence prediction (NSP): the models concatenates two masked sentences as inputs during pretraining. Sometimes
  they correspond to sentences that were next to each other in the original text, sometimes not. The model then has to
  predict if the two sentences were following each other or not.

This way, the model learns an inner representation of the English language that can then be used to extract features
useful for downstream tasks: if you have a dataset of labeled sentences for instance, you can train a standard
classifier using the features produced by the BERT model as inputs.

This model has the following configuration:

- 24-layer
- 1024 hidden dimension
- 16 attention heads
- 336M parameters.

## Intended uses & limitations
This model should be used as a question-answering model. You may use it in a question answering pipeline, or use it to output raw results given a query and a context. You may see other use cases in the [task summary](https://huggingface.co/transformers/task_summary.html#extractive-question-answering) of the transformers documentation.## Training data

The BERT model was pretrained on [BookCorpus](https://yknzhu.wixsite.com/mbweb), a dataset consisting of 11,038
unpublished books and [English Wikipedia](https://en.wikipedia.org/wiki/English_Wikipedia) (excluding lists, tables and
headers).

## Training procedure

### Preprocessing

The texts are lowercased and tokenized using WordPiece and a vocabulary size of 30,000. The inputs of the model are
then of the form:

```
[CLS] Sentence A [SEP] Sentence B [SEP]
```

With probability 0.5, sentence A and sentence B correspond to two consecutive sentences in the original corpus and in
the other cases, it's another random sentence in the corpus. Note that what is considered a sentence here is a
consecutive span of text usually longer than a single sentence. The only constrain is that the result with the two
"sentences" has a combined length of less than 512 tokens.

The details of the masking procedure for each sentence are the following:
- 15% of the tokens are masked.
- In 80% of the cases, the masked tokens are replaced by `[MASK]`.
- In 10% of the cases, the masked tokens are replaced by a random token (different) from the one they replace.
- In the 10% remaining cases, the masked tokens are left as is.

### Pretraining

The model was trained on 4 cloud TPUs in Pod configuration (16 TPU chips total) for one million steps with a batch size
of 256. The sequence length was limited to 128 tokens for 90% of the steps and 512 for the remaining 10%. The optimizer
used is Adam with a learning rate of 1e-4, \\(\beta_{1} = 0.9\\) and \\(\beta_{2} = 0.999\\), a weight decay of 0.01,
learning rate warmup for 10,000 steps and linear decay of the learning rate after.

### Fine-tuning

After pre-training, this model was fine-tuned on the SQuAD dataset with one of our fine-tuning scripts. In order to reproduce the training, you may use the following command:
```
python -m torch.distributed.launch --nproc_per_node=8 ./examples/question-answering/run_qa.py \
    --model_name_or_path bert-large-cased-whole-word-masking \
    --dataset_name squad \
    --do_train \
    --do_eval \
    --learning_rate 3e-5 \
    --num_train_epochs 2 \
    --max_seq_length 384 \
    --doc_stride 128 \
    --output_dir ./examples/models/wwm_cased_finetuned_squad/ \
    --per_device_eval_batch_size=3   \
    --per_device_train_batch_size=3   \
```

### BibTeX entry and citation info

```bibtex
@article{DBLP:journals/corr/abs-1810-04805,
  author    = {Jacob Devlin and
               Ming{-}Wei Chang and
               Kenton Lee and
               Kristina Toutanova},
  title     = {{BERT:} Pre-training of Deep Bidirectional Transformers for Language
               Understanding},
  journal   = {CoRR},
  volume    = {abs/1810.04805},
  year      = {2018},
  url       = {http://arxiv.org/abs/1810.04805},
  archivePrefix = {arXiv},
  eprint    = {1810.04805},
  timestamp = {Tue, 30 Oct 2018 20:39:56 +0100},
  biburl    = {https://dblp.org/rec/journals/corr/abs-1810-04805.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}
```