A Configurable Approximate Multiplier for Quantized Convolutional Neural Networks and Its Implementation
A convolutional neural network and multiplier technology, applied in the field of configurable approximate multipliers, can solve problems such as poor efficiency, small bit width, resource waste, etc., and achieve the effect of reducing area overhead and improving computing efficiency
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment Construction
[0034] The present invention will be described in further detail below in conjunction with the accompanying drawings and specific embodiments.
[0035] Such as figure 2 As shown, the present invention proposes a configurable approximate multiplier for quantizing convolutional neural networks, including the following modules:
[0036] (1) Sign extension module: will represent the range in -2 n-2 to 2 n-2 The n-bit signed fixed-point number of -1 is expressed as two n / 2-bit signed fixed-point numbers. When the n-bit signed fixed-point number is non-negative, the n / 2-1 bits from the lowest bit to the top are truncated, and in it 0 is added before the highest bit, and the whole is used as the input of the low bit multiplier, and the other n / 2 bits are used as the input of the high bit multiplier.
[0037] When n=8, the split method is:
[0038] 00XX_XXXX=0XXX_XXX→0XXX_0XXX
[0039] When the n-bit signed fixed-point number is negative, if the decimal value is less than -(2 n...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap