CN119452368A - 用于分布专用混合专家机器学习模型的层的系统和方法 - Google Patents

用于分布专用混合专家机器学习模型的层的系统和方法 Download PDF

Info

Publication number
CN119452368A
CN119452368A CN202380044796.8A CN202380044796A CN119452368A CN 119452368 A CN119452368 A CN 119452368A CN 202380044796 A CN202380044796 A CN 202380044796A CN 119452368 A CN119452368 A CN 119452368A
Authority
CN
China
Prior art keywords
accelerators
layers
sparse
experts
machine learning
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
Application number
CN202380044796.8A
Other languages
English (en)
Chinese (zh)
Inventor
D·R·怕特尔
左薇
余原
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Microsoft Technology Licensing LLC
Original Assignee
Microsoft Technology Licensing LLC
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Microsoft Technology Licensing LLC filed Critical Microsoft Technology Licensing LLC
Publication of CN119452368A publication Critical patent/CN119452368A/zh
Pending legal-status Critical Current

Links

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
    • G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
    • G06N3/063—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using electronic means
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • 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
    • G06N3/042—Knowledge-based neural networks; Logical representations of neural 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
    • G06N3/045—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
    • G06N3/08—Learning methods
    • 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
    • G06N3/08—Learning methods
    • G06N3/082—Learning methods modifying the architecture, e.g. adding, deleting or silencing nodes or connections
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T1/00—General purpose image data processing
    • G06T1/60—Memory management
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Neurology (AREA)
  • Medical Informatics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Complex Calculations (AREA)
CN202380044796.8A 2022-06-24 2023-05-16 用于分布专用混合专家机器学习模型的层的系统和方法 Pending CN119452368A (zh)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US17/848,679 US12579470B2 (en) 2022-06-24 2022-06-24 Systems and methods for distributing layers of special mixture-of-experts machine learning models
US17/848,679 2022-06-24
PCT/US2023/022447 WO2023249754A1 (en) 2022-06-24 2023-05-16 Systems and methods for distributing layers of special mixture-of-experts machine learning models

Publications (1)

Publication Number Publication Date
CN119452368A true CN119452368A (zh) 2025-02-14

Family

ID=87036771

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202380044796.8A Pending CN119452368A (zh) 2022-06-24 2023-05-16 用于分布专用混合专家机器学习模型的层的系统和方法

Country Status (6)

Country Link
US (2) US12579470B2 (enExample)
EP (1) EP4544451A1 (enExample)
JP (1) JP2025522299A (enExample)
KR (1) KR20250029051A (enExample)
CN (1) CN119452368A (enExample)
WO (1) WO2023249754A1 (enExample)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20250258725A1 (en) * 2024-02-13 2025-08-14 Hewlett Packard Enterprise Development Lp Control path communication between non-host processing devices
CN117972293B (zh) * 2024-03-28 2024-06-07 北京思凌科半导体技术有限公司 基于混合专家模型的计算方法、装置、设备及存储介质

Family Cites Families (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018085643A1 (en) * 2016-11-04 2018-05-11 Google Llc Mixture of experts neural networks
US10509846B2 (en) * 2017-12-13 2019-12-17 Intel Corporation Accelerator for processing data
US11893502B2 (en) 2017-12-20 2024-02-06 Advanced Micro Devices, Inc. Dynamic hardware selection for experts in mixture-of-experts model
WO2020023483A1 (en) * 2018-07-23 2020-01-30 Google Llc Continuous parametrizations of neural network layer weights
US11003823B2 (en) 2018-08-09 2021-05-11 Palo Alto Research Center Incorporated Re-design of analog circuits
US20200117978A1 (en) 2018-10-12 2020-04-16 Alibaba Group Holding Limited Systems and methods for efficiently mapping neural networks to programmable logic devices
WO2020099583A1 (en) 2018-11-18 2020-05-22 Innatera Nanosystems B.V. Spiking neural network
US12124941B2 (en) 2020-03-27 2024-10-22 Intel Corporation Methods and apparatus for dynamic batching of data for neural network workloads
US11586894B2 (en) 2020-05-04 2023-02-21 SiMa Technologies, Inc. Ordering computations of a machine learning network in a machine learning accelerator for efficient memory usage
US12530571B2 (en) 2020-07-08 2026-01-20 Nvidia Corporation Image generation using one or more neural networks
US20220036186A1 (en) * 2020-07-30 2022-02-03 Waymo Llc Accelerated deep reinforcement learning of agent control policies
US20220059200A1 (en) 2020-08-21 2022-02-24 Washington University Deep-learning systems and methods for medical report generation and anomaly detection
US12518135B2 (en) * 2021-02-05 2026-01-06 Google Llc Sparse and differentiable mixture of experts neural networks
US20230281510A1 (en) * 2022-03-07 2023-09-07 Qualcomm Incorporated Machine learning model architecture combining mixture of experts and model ensembling

Also Published As

Publication number Publication date
JP2025522299A (ja) 2025-07-15
KR20250029051A (ko) 2025-03-04
US12579470B2 (en) 2026-03-17
WO2023249754A1 (en) 2023-12-28
US20230419166A1 (en) 2023-12-28
EP4544451A1 (en) 2025-04-30
US20260178987A1 (en) 2026-06-25

Similar Documents

Publication Publication Date Title
US20230076850A1 (en) Computation of neural network node with large input values
CN110262901A (zh) 一种数据处理方法及数据处理系统
CN117112145B (zh) 训练模型分配方法、装置、计算机设备和存储介质
CN110689121A (zh) 一种用多核处理器实现神经网络模型拆分方法及相关产品
CN111105023B (zh) 数据流重构方法及可重构数据流处理器
US20240273346A1 (en) Self-balancing mixture of experts
CN114356540A (zh) 一种参数更新方法、装置、电子设备和存储介质
CN119783812B (zh) 面向新一代异构超算大模型并行训练与推理适配优化方法
EP4544451A1 (en) Systems and methods for distributing layers of special mixture-of-experts machine learning models
CN119046015B (zh) 神经网络模型训练处理的电子设备、方法和介质
Meng et al. An efficient CNN accelerator achieving high PE utilization using a dense-/sparse-aware redundancy reduction method and data–index decoupling workflow
Jiang et al. Exploiting potential of deep neural networks by layer-wise fine-grained parallelism
Zhan et al. Field programmable gate array‐based all‐layer accelerator with quantization neural networks for sustainable cyber‐physical systems
CN117114055B (zh) 面向工业应用场景的fpga二值神经网络加速方法
CN119065847A (zh) 多核大模型计算方法、装置、设备及存储介质
Liu et al. Parallelization Techniques for Large Language Models: A Review from Training to Inference
JP2023024960A (ja) 効率的なニューラルネットワーク実行のためのメモリ使用の最適化
Xu et al. EdgeMesh: A hybrid distributed training mechanism for heterogeneous edge devices
CN121188009B (zh) 向分布式模型提供数据的方法、装置、设备和存储介质
Chiu et al. Design and implementation of the CNN accelator based on multi-streaming SIMD mechanisms
US12373261B2 (en) Just-in-time re-partitioning of feature maps for efficient balancing of compute core workloads
US20260044380A1 (en) Mixed parallelism for execution of artificial intelligence workloads
CN121387770B (zh) 张量处理方法及装置、电子设备和存储介质
Karypis High-Performance Static and Streaming Tensor Factorization Algorithms
Aishwarya et al. Parallel Implementation of Dutch Flag Sorting Algorithm Using MPI and CUDA

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