Pytorch implementation of Google AI's 2018 BERT, with simple annotation

BERT 2018 BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding Paper URL : https://arxiv.org/abs/1810.04805

Google AI's BERT paper shows the amazing result on various NLP task (new 17 NLP tasks SOTA), including outperform the human F1 score on SQuAD v1.1 QA task. This paper proved that Transformer(self-attention) based encoder can be powerfully used as alternative of previous language model with proper language model training method. And more importantly, they showed us that this pre-trained language model can be transfer into any NLP task without making task specific model architecture.

This amazing result would be record in NLP history, and I expect many further papers about BERT will be published very soon.

This repo is implementation of BERT. Code is very simple and easy to understand fastly. Some of these codes are based on The Annotated Transformer

Currently this project is working on progress. And the code is not verified yet.

In the paper, authors shows the new language model training methods, which are "masked language model" and "predict next sentence".

Original Paper : 3.3.1 Task #1: Masked LM

Input Sequence  : The man went to [MASK] store with [MASK] dog
Target Sequence :                  the                his

Randomly 15% of input token will be changed into something, based on under sub-rules

1. Randomly 80% of tokens, gonna be a [MASK] token
2. Randomly 10% of tokens, gonna be a [RANDOM] token(another word)
3. Randomly 10% of tokens, will be remain as same. But need to be predicted.

Original Paper : 3.3.2 Task #2: Next Sentence Prediction

Input : [CLS] the man went to the store [SEP] he bought a gallon of milk [SEP]
Label : Is Next

Input = [CLS] the man heading to the store [SEP] penguin [MASK] are flight ##less birds [SEP]
Label = NotNext

"Is this sentence can be continuously connected?"

understanding the relationship, between two text sentences, which is not directly captured by language modeling

1. Randomly 50% of next sentence, gonna be continuous sentence.
2. Randomly 50% of next sentence, gonna be unrelated sentence.

NOTICE : Your corpus should be prepared with two sentences in one line with tab(\t) separator

Welcome to the \t the jungle \n
I can stay \t here all night \n

python build_vocab.py -c data/corpus.small -o data/corpus.small.vocab

usage: build_vocab.py [-h] -c CORPUS_PATH -o OUTPUT_PATH [-s VOCAB_SIZE]
                      [-e ENCODING] [-m MIN_FREQ]

optional arguments:
  -h, --help            show this help message and exit
  -c CORPUS_PATH, --corpus_path CORPUS_PATH
  -o OUTPUT_PATH, --output_path OUTPUT_PATH
  -s VOCAB_SIZE, --vocab_size VOCAB_SIZE
  -e ENCODING, --encoding ENCODING
  -m MIN_FREQ, --min_freq MIN_FREQ

python build_dataset.py -d data/corpus.small -v data/corpus.small.vocab -o data/dataset.small

usage: build_dataset.py [-h] -v VOCAB_PATH -c CORPUS_PATH [-e ENCODING] -o
                        OUTPUT_PATH

optional arguments:
  -h, --help            show this help message and exit
  -v VOCAB_PATH, --vocab_path VOCAB_PATH
  -c CORPUS_PATH, --corpus_path CORPUS_PATH
  -e ENCODING, --encoding ENCODING
  -o OUTPUT_PATH, --output_path OUTPUT_PATH

python train.py -d data/dataset.small -v data/corpus.small.vocab -o output/

usage: train.py [-h] -d TRAIN_DATASET [-t TEST_DATASET] -v VOCAB_PATH -o
                OUTPUT_DIR [-hs HIDDEN] [-n LAYERS] [-a ATTN_HEADS]
                [-s SEQ_LEN] [-b BATCH_SIZE] [-e EPOCHS]

optional arguments:
  -h, --help            show this help message and exit
  -d TRAIN_DATASET, --train_dataset TRAIN_DATASET
  -t TEST_DATASET, --test_dataset TEST_DATASET
  -v VOCAB_PATH, --vocab_path VOCAB_PATH
  -o OUTPUT_DIR, --output_dir OUTPUT_DIR
  -hs HIDDEN, --hidden HIDDEN
  -n LAYERS, --layers LAYERS
  -a ATTN_HEADS, --attn_heads ATTN_HEADS
  -s SEQ_LEN, --seq_len SEQ_LEN
  -b BATCH_SIZE, --batch_size BATCH_SIZE
  -e EPOCHS, --epochs EPOCHS

Junseong Kim, Scatter Lab ([email protected] / [email protected])

This project following Apache 2.0 License as written in LICENSE file

Copyright 2018 Junseong Kim, Scatter Lab, respective BERT contributors

Copyright (c) 2018 Alexander Rush : The Annotated Trasnformer