Oh! We Freeze: Improving Quantized Knowledge Distillation via Signal Propagation Analysis for Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bhardwaj, Kartikeya, Pandey, Nilesh Prasad, Priyadarshi, Sweta, Lee, Kyunggeun, Ma, Jun, Teague, Harris
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910388014022656
author Bhardwaj, Kartikeya
Pandey, Nilesh Prasad
Priyadarshi, Sweta
Lee, Kyunggeun
Ma, Jun
Teague, Harris
author_facet Bhardwaj, Kartikeya
Pandey, Nilesh Prasad
Priyadarshi, Sweta
Lee, Kyunggeun
Ma, Jun
Teague, Harris
contents Large generative models such as large language models (LLMs) and diffusion models have revolutionized the fields of NLP and computer vision respectively. However, their slow inference, high computation and memory requirement makes it challenging to deploy them on edge devices. In this study, we propose a light-weight quantization aware fine tuning technique using knowledge distillation (KD-QAT) to improve the performance of 4-bit weight quantized LLMs using commonly available datasets to realize a popular language use case, on device chat applications. To improve this paradigm of finetuning, as main contributions, we provide insights into stability of KD-QAT by empirically studying the gradient propagation during training to better understand the vulnerabilities of KD-QAT based approaches to low-bit quantization errors. Based on our insights, we propose ov-freeze, a simple technique to stabilize the KD-QAT process. Finally, we experiment with the popular 7B LLaMAv2-Chat model at 4-bit quantization level and demonstrate that ov-freeze results in near floating point precision performance, i.e., less than 0.7% loss of accuracy on Commonsense Reasoning benchmarks.
format Preprint
id arxiv_https___arxiv_org_abs_2403_18159
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Oh! We Freeze: Improving Quantized Knowledge Distillation via Signal Propagation Analysis for Large Language Models
Bhardwaj, Kartikeya
Pandey, Nilesh Prasad
Priyadarshi, Sweta
Lee, Kyunggeun
Ma, Jun
Teague, Harris
Machine Learning
Artificial Intelligence
Computation and Language
Large generative models such as large language models (LLMs) and diffusion models have revolutionized the fields of NLP and computer vision respectively. However, their slow inference, high computation and memory requirement makes it challenging to deploy them on edge devices. In this study, we propose a light-weight quantization aware fine tuning technique using knowledge distillation (KD-QAT) to improve the performance of 4-bit weight quantized LLMs using commonly available datasets to realize a popular language use case, on device chat applications. To improve this paradigm of finetuning, as main contributions, we provide insights into stability of KD-QAT by empirically studying the gradient propagation during training to better understand the vulnerabilities of KD-QAT based approaches to low-bit quantization errors. Based on our insights, we propose ov-freeze, a simple technique to stabilize the KD-QAT process. Finally, we experiment with the popular 7B LLaMAv2-Chat model at 4-bit quantization level and demonstrate that ov-freeze results in near floating point precision performance, i.e., less than 0.7% loss of accuracy on Commonsense Reasoning benchmarks.
title Oh! We Freeze: Improving Quantized Knowledge Distillation via Signal Propagation Analysis for Large Language Models
topic Machine Learning
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2403.18159