CN112955909A - 神经网络的分布式训练方法及装置 - Google Patents
神经网络的分布式训练方法及装置 Download PDFInfo
- Publication number
- CN112955909A CN112955909A CN201980069560.3A CN201980069560A CN112955909A CN 112955909 A CN112955909 A CN 112955909A CN 201980069560 A CN201980069560 A CN 201980069560A CN 112955909 A CN112955909 A CN 112955909A
- Authority
- CN
- China
- Prior art keywords
- gradient
- aggregation
- neural network
- calculation
- data
- 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.)
- Pending
Links
- 238000012549 training Methods 0.000 title claims abstract description 183
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 148
- 238000000034 method Methods 0.000 title claims abstract description 79
- 230000002776 aggregation Effects 0.000 claims abstract description 287
- 238000004220 aggregation Methods 0.000 claims abstract description 287
- 238000004364 calculation method Methods 0.000 claims abstract description 209
- 230000015654 memory Effects 0.000 claims description 38
- 238000004590 computer program Methods 0.000 claims description 17
- 230000004931 aggregating effect Effects 0.000 claims description 6
- 230000008569 process Effects 0.000 description 19
- 238000004891 communication Methods 0.000 description 11
- 238000013461 design Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 11
- 238000011144 upstream manufacturing Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 10
- 238000009825 accumulation Methods 0.000 description 9
- 238000012545 processing Methods 0.000 description 7
- 230000005540 biological transmission Effects 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 239000004065 semiconductor Substances 0.000 description 4
- 230000001133 acceleration Effects 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000013515 script Methods 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011981 development test Methods 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 208000016339 iris pattern Diseases 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000006116 polymerization reaction Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/10—Interfaces, programming languages or software development kits, e.g. for simulating neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Complex Calculations (AREA)
Abstract
本申请提供一种神经网络的分布式训练方法及装置，能够有效减少确定梯度聚合方案的次数和耗时，从而提高训练效率。该方法包括：在启动训练任务之前，一次性地确定梯度聚合阈值，然后根据该梯度聚合阈值，将神经网络中的多个梯度计算算子划分为多个各包括至少一个梯度计算算子的梯度聚合集合，且当任意一个梯度聚合集合中的所有梯度计算算子均完成一个梯度计算时，则将该任意一个梯度聚合集合作为一个整体，执行一次梯度聚合操作。
Description
PCT国内申请，说明书已公开。
Claims (13)
- PCT国内申请，权利要求书已公开。
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2019/074365 WO2020155083A1 (zh) | 2019-02-01 | 2019-02-01 | 神经网络的分布式训练方法及装置 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112955909A true CN112955909A (zh) | 2021-06-11 |
Family
ID=71840223
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201980069560.3A Pending CN112955909A (zh) | 2019-02-01 | 2019-02-01 | 神经网络的分布式训练方法及装置 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN112955909A (zh) |
| WO (1) | WO2020155083A1 (zh) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210081770A1 (en) * | 2019-09-17 | 2021-03-18 | GOWN Semiconductor Corporation | System architecture based on soc fpga for edge artificial intelligence computing |
| CN114580664A (zh) * | 2022-03-03 | 2022-06-03 | 字节跳动(香港)有限公司 | 训练分析方法、装置、存储介质及电子设备 |
| CN114900482A (zh) * | 2022-03-28 | 2022-08-12 | 中国科学技术大学苏州高等研究院 | Ps架构下基于可编程交换机的梯度调度方法和装置 |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140279098A1 (en) * | 2013-03-15 | 2014-09-18 | Brandon Ham | Bill Splitting and Payment System and Method |
| US20180075347A1 (en) * | 2016-09-15 | 2018-03-15 | Microsoft Technology Licensing, Llc | Efficient training of neural networks |
| US20180204111A1 (en) * | 2013-02-28 | 2018-07-19 | Z Advanced Computing, Inc. | System and Method for Extremely Efficient Image and Pattern Recognition and Artificial Intelligence Platform |
| CN108960410A (zh) * | 2018-06-13 | 2018-12-07 | 华为技术有限公司 | 基于神经网络的参数更新方法、相关平台及计算机存储介质 |
| US10152676B1 (en) * | 2013-11-22 | 2018-12-11 | Amazon Technologies, Inc. | Distributed training of models using stochastic gradient descent |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107229518B (zh) * | 2016-03-26 | 2020-06-30 | 阿里巴巴集团控股有限公司 | 一种分布式集群训练方法和装置 |
| TW201812646A (zh) * | 2016-07-18 | 2018-04-01 | 美商南坦奧美克公司 | 分散式機器學習系統、分散式機器學習方法、以及產生代用資料之方法 |
| CN108122032B (zh) * | 2016-11-29 | 2020-02-14 | 华为技术有限公司 | 一种神经网络模型训练方法、装置、芯片和系统 |
-
2019
- 2019-02-01 WO PCT/CN2019/074365 patent/WO2020155083A1/zh active Application Filing
- 2019-02-01 CN CN201980069560.3A patent/CN112955909A/zh active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180204111A1 (en) * | 2013-02-28 | 2018-07-19 | Z Advanced Computing, Inc. | System and Method for Extremely Efficient Image and Pattern Recognition and Artificial Intelligence Platform |
| US20140279098A1 (en) * | 2013-03-15 | 2014-09-18 | Brandon Ham | Bill Splitting and Payment System and Method |
| US10152676B1 (en) * | 2013-11-22 | 2018-12-11 | Amazon Technologies, Inc. | Distributed training of models using stochastic gradient descent |
| US20180075347A1 (en) * | 2016-09-15 | 2018-03-15 | Microsoft Technology Licensing, Llc | Efficient training of neural networks |
| CN108960410A (zh) * | 2018-06-13 | 2018-12-07 | 华为技术有限公司 | 基于神经网络的参数更新方法、相关平台及计算机存储介质 |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210081770A1 (en) * | 2019-09-17 | 2021-03-18 | GOWN Semiconductor Corporation | System architecture based on soc fpga for edge artificial intelligence computing |
| US11544544B2 (en) * | 2019-09-17 | 2023-01-03 | Gowin Semiconductor Corporation | System architecture based on SoC FPGA for edge artificial intelligence computing |
| CN114580664A (zh) * | 2022-03-03 | 2022-06-03 | 字节跳动(香港)有限公司 | 训练分析方法、装置、存储介质及电子设备 |
| CN114900482A (zh) * | 2022-03-28 | 2022-08-12 | 中国科学技术大学苏州高等研究院 | Ps架构下基于可编程交换机的梯度调度方法和装置 |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2020155083A1 (zh) | 2020-08-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11521066B2 (en) | Method and apparatus for partitioning deep neural networks | |
| WO2019018375A1 (en) | NEURONAL ARCHITECTURE RESEARCH FOR CONVOLUTION NEURAL NETWORKS | |
| US9471470B2 (en) | Automatically recommending test suite from historical data based on randomized evolutionary techniques | |
| CN112955909A (zh) | 神经网络的分布式训练方法及装置 | |
| US11455523B2 (en) | Risk evaluation method, computer-readable recording medium, and information processing apparatus | |
| US11580458B2 (en) | Method and system for performance tuning and performance tuning device | |
| US20230196202A1 (en) | System and method for automatic building of learning machines using learning machines | |
| US20160187861A1 (en) | Systems and methods to adaptively select execution modes | |
| CN111008152B (zh) | 一种基于函数依赖图的内核模块兼容影响域分析方法、系统和介质 | |
| CN112700006B (zh) | 网络架构搜索方法、装置、电子设备及介质 | |
| US20190220924A1 (en) | Method and device for determining key variable in model | |
| US11972238B2 (en) | Propagating reduced-precision on computation graphs | |
| EP4058943A1 (en) | Threshold triggered back propagation of an artificial neural network | |
| CN112001491A (zh) | 针对处理器确定神经网络架构的搜索方法和装置 | |
| Audrito et al. | Resilient blocks for summarising distributed data | |
| US11475311B2 (en) | Neural network instruction streaming | |
| CN113312169B (zh) | 一种计算资源的分配方法及装置 | |
| CN111860758B (zh) | 一种深度学习模型的运行方法、装置、电子设备及介质 | |
| CN114841664A (zh) | 一种多任务处理顺序确定方法及装置 | |
| CN111522635B (zh) | 计算任务处理方法、装置、服务器及存储介质 | |
| CN115564055A (zh) | 异步联合学习训练方法、装置、计算机设备及存储介质 | |
| KR20220036493A (ko) | 뉴럴 네트워크 추론을 위한 최적화 방법 및 이를 수행하는 컴퓨팅 장치 | |
| US20240152765A1 (en) | Training time and resource consumption prediction in deep learning | |
| CN114579202B (zh) | 任务处理方法、装置、计算机设备及计算机可读存储介质 | |
| US20230334370A1 (en) | Model gradient determining methods, apparatuses, devices, and media based on federated learning |
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 |