The performance of neural networks improves when more parameters are used. However, the model sizes are constrained by the available on-device memory during training and inference. Although applying techniques like quantization can alleviate the constraint, they suffer from performance degradation. In this work, we introduce NeuZip, a new weight compression scheme based on the entropy of floating-point numbers in neural networks. With NeuZip, we are able to achieve memory-efficient training and inference without sacrificing performance. Notably, we significantly reduce the memory footprint of training a Llama-3 8B model from 31GB to less than 16GB, while keeping the training dynamics fully unchanged. In inference, our method can reduce memory usage by more than half while maintaining near-lossless performance. Our code is publicly available.
@article{arxiv.2410.20650,
title = {NeuZip: Memory-Efficient Training and Inference with Dynamic Compression of Neural Networks},
author = {Yongchang Hao and Yanshuai Cao and Lili Mou},
journal= {arXiv preprint arXiv:2410.20650},
year = {2024}
}