kpe/bert-for-tf2
A TensorFlow/Keras implementation of BERT, ALBERT, and adapter-BERT with pre-trained weight loading support.
Not currently ranked — collecting fresh signals.
star history
This repository provides a Keras implementation of BERT (Bidirectional Encoder Representations from Transformers) for TensorFlow 2.0, with support for loading original pre-trained weights from Google Research. It also supports ALBERT (A Lite BERT) and adapter-BERT through configuration parameters. The implementation builds from scratch using basic TensorFlow operations following the original modeling.py code.
Frequently asked
- What is kpe/bert-for-tf2?
- A TensorFlow/Keras implementation of BERT, ALBERT, and adapter-BERT with pre-trained weight loading support.
- Is bert-for-tf2 open source?
- Yes — kpe/bert-for-tf2 is open source, released under the MIT license.
- What language is bert-for-tf2 written in?
- kpe/bert-for-tf2 is primarily written in Python.
- How popular is bert-for-tf2?
- kpe/bert-for-tf2 has 807 stars on GitHub.
- Where can I find bert-for-tf2?
- kpe/bert-for-tf2 is on GitHub at https://github.com/kpe/bert-for-tf2.