CN111047007A - 一种针对量化后的lstm的激活函数计算单元 - Google Patents
一种针对量化后的lstm的激活函数计算单元 Download PDFInfo
- Publication number
- CN111047007A CN111047007A CN201911076853.2A CN201911076853A CN111047007A CN 111047007 A CN111047007 A CN 111047007A CN 201911076853 A CN201911076853 A CN 201911076853A CN 111047007 A CN111047007 A CN 111047007A
- Authority
- CN
- China
- Prior art keywords
- unit
- input data
- calculation unit
- tanh
- function calculation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
Description
Claims (2)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911076853.2A CN111047007B (zh) | 2019-11-06 | 2019-11-06 | 一种针对量化后的lstm的激活函数计算单元 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911076853.2A CN111047007B (zh) | 2019-11-06 | 2019-11-06 | 一种针对量化后的lstm的激活函数计算单元 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111047007A true CN111047007A (zh) | 2020-04-21 |
| CN111047007B CN111047007B (zh) | 2021-07-30 |
Family
ID=70233048
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911076853.2A Active CN111047007B (zh) | 2019-11-06 | 2019-11-06 | 一种针对量化后的lstm的激活函数计算单元 |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111047007B (zh) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111652361A (zh) * | 2020-06-04 | 2020-09-11 | 南京博芯电子技术有限公司 | 长短时记忆网络的复合粒度近存储近似加速结构和方法 |
| WO2023280113A1 (zh) * | 2021-07-08 | 2023-01-12 | 华为技术有限公司 | 数据处理方法、神经网络模型的训练方法及装置 |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106775599A (zh) * | 2017-01-09 | 2017-05-31 | 南京工业大学 | 递归神经网络的多计算单元粗粒度可重构系统及方法 |
| CN107688849A (zh) * | 2017-07-28 | 2018-02-13 | 北京深鉴科技有限公司 | 一种动态策略定点化训练方法及装置 |
| CN108090560A (zh) * | 2018-01-05 | 2018-05-29 | 中国科学技术大学苏州研究院 | 基于fpga的lstm递归神经网络硬件加速器的设计方法 |
| CN109583561A (zh) * | 2017-09-28 | 2019-04-05 | 杭州海康威视数字技术股份有限公司 | 一种深度神经网络的激活量量化方法及装置 |
| CN109816105A (zh) * | 2019-01-16 | 2019-05-28 | 北京时代民芯科技有限公司 | 一种可配置的神经网络激活函数实现装置 |
-
2019
- 2019-11-06 CN CN201911076853.2A patent/CN111047007B/zh active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106775599A (zh) * | 2017-01-09 | 2017-05-31 | 南京工业大学 | 递归神经网络的多计算单元粗粒度可重构系统及方法 |
| CN107688849A (zh) * | 2017-07-28 | 2018-02-13 | 北京深鉴科技有限公司 | 一种动态策略定点化训练方法及装置 |
| CN109583561A (zh) * | 2017-09-28 | 2019-04-05 | 杭州海康威视数字技术股份有限公司 | 一种深度神经网络的激活量量化方法及装置 |
| CN108090560A (zh) * | 2018-01-05 | 2018-05-29 | 中国科学技术大学苏州研究院 | 基于fpga的lstm递归神经网络硬件加速器的设计方法 |
| CN109816105A (zh) * | 2019-01-16 | 2019-05-28 | 北京时代民芯科技有限公司 | 一种可配置的神经网络激活函数实现装置 |
Non-Patent Citations (3)
| Title |
|---|
| SIRISH KUMAR PASUPULETI ET AL: "Low Complex & High Accuracy Computation Approximations to Enable On-Device RNN Applications", 《2019 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS (ISCAS)》 * |
| VLADIMIR RYBALKIN ET AL: "FINN-L: Library Extensions and Design Trade-Off Analysis for Variable Precision LSTM Networks on FPGAs", 《2018 28TH INTERNATIONAL CONFERENCE ON FIELD PROGRAMMABLE LOGIC AND APPLICATIONS (FPL)》 * |
| 马秋然 等: "一种基于折线逼近操作的极化码译码算法", 《HTTP://KNS.CNKI.NET/KCMS/DETAIL/51.1196.TP.20190618.0905.001.HTML》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111652361A (zh) * | 2020-06-04 | 2020-09-11 | 南京博芯电子技术有限公司 | 长短时记忆网络的复合粒度近存储近似加速结构和方法 |
| CN111652361B (zh) * | 2020-06-04 | 2023-09-26 | 南京博芯电子技术有限公司 | 长短时记忆网络的复合粒度近存储近似加速结构系统和方法 |
| WO2023280113A1 (zh) * | 2021-07-08 | 2023-01-12 | 华为技术有限公司 | 数据处理方法、神经网络模型的训练方法及装置 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111047007B (zh) | 2021-07-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112230881B (zh) | 浮点数处理器 | |
| CN107305485B (zh) | 一种用于执行多个浮点数相加的装置及方法 | |
| CN105354006B (zh) | 一种非线性函数的快速运算装置及其方法 | |
| Tambe et al. | Adaptivfloat: A floating-point based data type for resilient deep learning inference | |
| Spagnolo et al. | Aggressive approximation of the softmax function for power-efficient hardware implementations | |
| CN110519058A (zh) | 一种对于基于格的公钥加密算法的加速方法 | |
| CN112119407A (zh) | 由补偿指令使能的低精度深度神经网络 | |
| CN108196822A (zh) | 一种双精度浮点开方运算的方法及系统 | |
| Wu et al. | An energy-efficient approximate divider based on logarithmic conversion and piecewise constant approximation | |
| CN111047007B (zh) | 一种针对量化后的lstm的激活函数计算单元 | |
| KR20230132795A (ko) | 정규화 기능을 위한 디지털 회로 | |
| Choi et al. | Retrain-less weight quantization for multiplier-less convolutional neural networks | |
| Christ et al. | Low-precision logarithmic arithmetic for neural network accelerators | |
| US11243743B2 (en) | Optimization of neural networks using hardware calculation efficiency and adjustment factors | |
| CN110110852B (zh) | 一种深度学习网络移植到fpag平台的方法 | |
| CN113419779B (zh) | 可扩展多精度数据流水线系统和方法 | |
| Kwak et al. | Torchaxf: Enabling rapid simulation of approximate dnn models using gpu-based floating-point computing framework | |
| Khan et al. | Comparative analysis of different algorithm for design of high-speed multiplier accumulator unit (MAC) | |
| Liu et al. | E-ERA: An energy-efficient reconfigurable architecture for RNNs using dynamically adaptive approximate computing | |
| Klinefelter et al. | Error-energy analysis of hardware logarithmic approximation methods for low power applications | |
| Vasudevan et al. | Image processing using approximate datapath units | |
| Duarte et al. | Deep Nibble: A 4-bit Number Format for Efficient DNN Training and Inference in FPGA | |
| CN115965048A (zh) | 数据处理装置、数据处理方法和电子设备 | |
| Aguilera-Galicia et al. | IEEE-754 half-precision floating-point low-latency reciprocal square root IP-core | |
| CN109416757B (zh) | 用于处理数值数据的方法、设备和计算机可读存储介质 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Address after: 100044 1-609, floor 6, building 1, courtyard 42, Xiejie street, gaoliangqiao, Haidian District, Beijing Applicant after: Beijing Zhongke Shengxin Technology Co.,Ltd. Address before: Room 403-14, 4th floor, building 6, yard 54, Shijingshan Road, Shijingshan District, Beijing 100043 Applicant before: Beijing Zhongke Shengxin Technology Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20220324 Address after: 225200 No. 58, Sanyuan South Road, Xiannv Town, Jiangdu District, Yangzhou City, Jiangsu Province Patentee after: Yida Ind Co.,Ltd. Yangzhou Address before: 100044 1-609, floor 6, building 1, courtyard 42, Xiejie street, gaoliangqiao, Haidian District, Beijing Patentee before: Beijing Zhongke Shengxin Technology Co.,Ltd. |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20241202 Address after: Building 1, 6th Floor, No. 42 Gaoliangqiao Xie Street, Haidian District, Beijing 100081, China Patentee after: Beijing Zhongke Shengxin Technology Co.,Ltd. Country or region after: China Address before: 225200 No. 58, Sanyuan South Road, Xiannv Town, Jiangdu District, Yangzhou City, Jiangsu Province Patentee before: Yida Ind Co.,Ltd. Yangzhou Country or region before: China |
|
| TR01 | Transfer of patent right |