SONIQ: System-Optimized Noise-Injected Ultra-Low-Precision Quantization with Full-Precision Parity

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhou, Cyrus, Savarese, Pedro, Hassman, Zack, Richard, Vaughn, DiBrino, Michael, Maire, Michael, Li, Yanjing
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917067846844416
author Zhou, Cyrus
Savarese, Pedro
Hassman, Zack
Richard, Vaughn
DiBrino, Michael
Maire, Michael
Li, Yanjing
author_facet Zhou, Cyrus
Savarese, Pedro
Hassman, Zack
Richard, Vaughn
DiBrino, Michael
Maire, Michael
Li, Yanjing
contents Ultra-low-precision inference can sharply reduce memory and latency but often degrades accuracy and relies on specialized hardware. We present SONIQ, a system-optimized, noise-injected quantization framework that learns per-channel mixed precision for both weights and activations while training under the same rules used at inference. By injecting hardware-calibrated quantization noise during training, SONIQ steers models toward the discrete arithmetic used at deployment -- without bespoke runtimes. Across CNNs and Transformers, SONIQ achieves up to 16x and 7x compression, respectively, while matching or exceeding full-precision accuracy. Measured end-to-end, SONIQ delivers up to 7.3x CPU speedup over strong INT8 baselines and up to 6.3x (vector units) / 2.8x (tensor cores) GPU speedup relative to FP16. A practical outcome is that two per-channel precision levels -- one in the 1--4-bit range and one in the 4--8-bit range -- suffice in practice; at inference, each channel selects one of the two, keeping kernels simple and fast. To our knowledge, SONIQ is the first framework to reach or surpass full-precision accuracy under ultra-low (1--4 bits per parameter) regimes while remaining deployable on commodity hardware, narrowing the gap between quantization theory and practical, high-throughput inference.
format Preprint
id arxiv_https___arxiv_org_abs_2311_14114
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle SONIQ: System-Optimized Noise-Injected Ultra-Low-Precision Quantization with Full-Precision Parity
Zhou, Cyrus
Savarese, Pedro
Hassman, Zack
Richard, Vaughn
DiBrino, Michael
Maire, Michael
Li, Yanjing
Hardware Architecture
Machine Learning
Performance
Ultra-low-precision inference can sharply reduce memory and latency but often degrades accuracy and relies on specialized hardware. We present SONIQ, a system-optimized, noise-injected quantization framework that learns per-channel mixed precision for both weights and activations while training under the same rules used at inference. By injecting hardware-calibrated quantization noise during training, SONIQ steers models toward the discrete arithmetic used at deployment -- without bespoke runtimes. Across CNNs and Transformers, SONIQ achieves up to 16x and 7x compression, respectively, while matching or exceeding full-precision accuracy. Measured end-to-end, SONIQ delivers up to 7.3x CPU speedup over strong INT8 baselines and up to 6.3x (vector units) / 2.8x (tensor cores) GPU speedup relative to FP16. A practical outcome is that two per-channel precision levels -- one in the 1--4-bit range and one in the 4--8-bit range -- suffice in practice; at inference, each channel selects one of the two, keeping kernels simple and fast. To our knowledge, SONIQ is the first framework to reach or surpass full-precision accuracy under ultra-low (1--4 bits per parameter) regimes while remaining deployable on commodity hardware, narrowing the gap between quantization theory and practical, high-throughput inference.
title SONIQ: System-Optimized Noise-Injected Ultra-Low-Precision Quantization with Full-Precision Parity
topic Hardware Architecture
Machine Learning
Performance
url https://arxiv.org/abs/2311.14114