Model Overview

BERT is a method of pre-training language representations which obtains state-of-the-art results on a wide array of NLP tasks.

Model Architecture

The BERT model uses the same architecture as the encoder of the Transformer. Input sequences are projected into an embedding space before being fed into the encoder structure. Additionally, positional and segment encodings are added to the embeddings to preserve positional information. The encoder structure is simply a stack of Transformer blocks, which consist of a multi-head attention layer followed by successive stages of feed-forward networks and layer normalization. The multi-head attention layer accomplishes self-attention on multiple input representations.

An illustration of the architecture taken from the Transformer paper is shown below.

Training

This model was trained using script available in GitHub repo.

Dataset

The following datasets were used to train this model:

• Wikipedia - Dataset containing a 170GB+ Wikipedia dump.
• Bookcorpus - Large-scale text corpus for unsupervised learning of sentence encoders/decoders.
• Stanford Sentiment Treebank v2 - Dataset for predicting Sentiment from longer Movie Reviews.

Performance

Performance numbers for this model are available in GitHub readme performance section.

References

• Original paper
• NVIDIA model implementation on GitHub

License

This model was trained using open-source software available in Deep Learning Examples repository. For terms of use, please refer to the license of the script and the datasets the model was derived from.