UNIQ: Uniform Noise Injection for Non-Uniform Quantization of Neural Networks

Abstract

We present a novel method for neural network quantization that emulates a non-uniform $k$-quantile quantizer, which adapts to the distribution of the quantized parameters. Our approach provides a novel alternative to the existing uniform quantization techniques for neural networks. We suggest to compare the results as a function of the bit-operations (BOPS) performed, assuming a look-up table availability for the non-uniform case. In this setup, we show the advantages of our strategy in the low computational budget regime. While the proposed solution is harder to implement in hardware, we believe it sets a basis for new alternatives to neural networks quantization.

References in corpus (1)

• A Quantization-Friendly Separable Convolution for MobileNets

Cited by in corpus (10)

• Artificial Neural Networks for Photonic Applications: From Algorithms to Implementation
• Reducing Computational Complexity of Neural Networks in Optical Channel Equalization: From Concepts to Implementation
• Ps and Qs: Quantization-aware pruning for efficient low latency neural network inference
• NICE: Noise Injection and Clamping Estimation for Neural Network Quantization
• MARS: Multi-macro Architecture SRAM CIM-Based Accelerator with Co-designed Compressed Neural Networks
• HCM: Hardware-Aware Complexity Metric for Neural Network Architectures
• NAWQ-SR: A Hybrid-Precision NPU Engine for Efficient On-Device Super-Resolution
• CStream: Parallel Data Stream Compression on Multicore Edge Devices
• Adaptive Precision Training (AdaPT): A dynamic fixed point quantized training approach for DNNs
• Training Multi-Layer Binary Neural Networks With Local Binary Error Signals