pocokhc/BitNet

This is an unofficial implementation of BitLinear.

BitNet

This is an unofficial implementation of BitNet.
It is implemented only for BitNetLinear.

Paper

• BitNet: Scaling 1-bit Transformers for Large Language Models
• The Era of 1-bit LLMs: All Large Language Models are in 1.58 Bits

Explanatory article

Although it is in Japanese, I am posting articles on Qiita.

BitNet b1.58(BitLinear)を実装してMNISTで検証してみた(Tensorflow/Torch)

Usage

Execution examples

# Tensorflow
> python examples/mnist_tf.py

# Torch
> python examples/mnist_torch.py

Development Version

Python : 3.12.2
Tensorflow : 2.16.1
Torch : 2.2.1