BERT MULTI GPU

BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

REQUIREMENT

python 3

tensorflow 1.12.0

0, edit the input and output file name in create_pretraining_data.py and run_pretraining_gpu_v2.py

1, run create_pretraining_data.py

2, run run_pretraining_gpu_v2.py

Edit n_gpus in run_pretraining_gpu_v2.py

In sample_text.txt, sentence is end by \n, paragraph is splitted by empty line.

Quora question pairs English dataset,

Official BERT: ACC 91.2, AUC 96.9

This BERT with pretrain loss 2.05: ACC 90.1, AUC 96.3

For inference speed research.