CN112613597B - Comprehensive pipe rack risk automatic identification convolutional neural network model and construction method - Google Patents
Comprehensive pipe rack risk automatic identification convolutional neural network model and construction method Download PDFInfo
- Publication number
- CN112613597B CN112613597B CN202011374969.7A CN202011374969A CN112613597B CN 112613597 B CN112613597 B CN 112613597B CN 202011374969 A CN202011374969 A CN 202011374969A CN 112613597 B CN112613597 B CN 112613597B
- Authority
- CN
- China
- Prior art keywords
- neural network
- network system
- data
- deep learning
- convolution operation
- 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.)
- Active
Links
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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield 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
- G06N3/06—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; 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Neurology (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to an automatic comprehensive pipe gallery risk identification convolutional neural network model, which comprises a designer, wherein a nested architecture BP neural network system adopting a C/S structure and a B/S structure is arranged in the designer, and a plurality of CNN convolutional operation systems, an LSTM-based intelligent prediction system, a deep learning neural network system and a bottom operating system are additionally arranged in the designer, wherein the bottom operating system is respectively connected with the BP neural network system, the CNN convolutional operation systems, the LSTM-based intelligent prediction system and the deep learning neural network system through scheduling programs. The application method comprises two steps of system prefabrication, system training and the like. On one hand, the system has simple system constitution, strong data operation processing capability and good universality; on the other hand, the system has good autonomous operation capability and calculation result verification capability in data operation, so that the working efficiency of the operation of the design system and the precision of the design operation are greatly improved.
Description
Technical Field
The invention relates to a design and data processing system, in particular to a comprehensive pipe rack risk automatic identification convolutional neural network model and a detection method.
Background
2017108985024A convolutional neural network CNN hardware accelerator and acceleration method, although the current system can improve the data calculation processing efficiency in design work to a certain extent, the data processing efficiency is relatively low, the degree of autonomy, automation and intelligence of data operation is relatively low, meanwhile, the capability of effectively checking and correcting data calculation is also lacking, so that the running efficiency and accuracy bureau of the current data calculation processing system are relatively poor, the labor intensity of the data calculation operation is relatively high, and the actual use requirement is difficult to meet.
Thus, in response to this need, there is an urgent need to develop a new system and method related to the system and method to meet the needs of practical use.
Disclosure of Invention
Aiming at the defects existing in the prior art, the invention provides a comprehensive pipe rack risk automatic identification convolutional neural network model and a method thereof, so as to achieve the purpose of improving the image information acquisition operation efficiency, accuracy and flexibility.
In order to achieve the above object, the present invention is realized by the following technical scheme:
the utility model provides an automatic discernment convolutional neural network model of utility tunnel risk, including the designer, establish the nested framework BP neural network system that adopts C/S structure and B/S structure in the designer, establish a plurality of CNN convolution operation system in addition simultaneously, intelligent prediction system based on LSTM, deep learning neural network system and bottom operating system, wherein bottom operating system is connected with BP neural network system respectively through the scheduler, CNN convolution operation system, intelligent prediction system based on LSTM, deep learning neural network system, CNN convolution operation system are parallelly connected each other, and be connected with BP neural network system ' S input layer respectively, LSTM ' S intelligent prediction system, deep learning neural network system interconnect and be connected with BP neural network system input layer simultaneously, BP neural network system ' S output layer is connected with deep learning neural network system and a CNN convolution operation system in addition.
Furthermore, the deep learning neural network system is any one or a plurality of common deep confidence networks which are used for pre-training based on convolution operation, self-coding neural networks based on multi-layer neurons and multi-layer self-coding neural networks, and further optimizing the weights of the neural networks by combining identification information.
Further, the designer includes AI basic artificial intelligence logic processing module, based on FPGA's data processing unit, GPU image processing unit, data bus module, clock circuit module, drive circuit module, I/O data communication port and network communication port, data bus module respectively with AI basic artificial intelligence logic processing module, FPGA's data processing unit, GPU image processing unit, clock circuit module, drive circuit module, I/O data communication port and network communication port are connected, just drive circuit module is in addition connected with each GPU image processing unit, I/O data communication port and network communication port respectively.
Furthermore, the I/O data communication port is connected with at least one CCD camera, a laser scanning head, a control keyboard and a display respectively, and the network communication port is connected with an external communication network.
Furthermore, the number of the data processing units and the number of the GPU image processing units of the FPGA are not less than two, the data processing units of the FPGAs and the GPU image processing units are connected in parallel, the data processing units of the FPGAs are connected in series by adopting synchronous duplex circuits, and the GPU image processing units are connected in series by adopting asynchronous duplex circuits.
Further, the underlying operating system is a program system based on an SOA system.
A construction method for automatically identifying convolutional neural network model of comprehensive pipe rack risk includes the following steps;
s1, prefabricating a system, namely positioning a designer at a designated position, connecting the designer with external hardware equipment, modulating connection relations among a BP neural network system, a CNN convolution operation system, an LSTM-based intelligent prediction system and a deep learning neural network system through a bottom layer operation system, communicating one CNN convolution operation system in each CNN convolution operation system with an output layer of the BP neural network system, and connecting the CNN convolution operation system at the output end of the BP neural network system with each CNN convolution operation system connected with an input layer of the BP neural network system respectively, so that the prefabrication of the system is completed;
s2, training the system, namely, after the step S1 is completed, inputting the common algorithm participated in design operation and the data participated in calculation into the LSTM-based intelligent prediction system and the deep learning neural network system, training the BP neural network system through the LSTM-based intelligent prediction system and the deep learning neural network system, generating artificial intelligent design logic, processing and operating the common algorithm participated in design operation and the data participated in calculation through the CNN convolution operation system, transmitting the processed data into the BP neural network system, operating the received data according to the artificial intelligent design logic by the BP neural network system, outputting an operation result from an output layer, backing up an output data result, transmitting the backed up data result into the CNN convolution operation system connected with the deep learning neural network system and the BP neural network system after the current operation is completed by the BP neural network system, and updating the artificial intelligent design logic of the BP neural network system through the LSTM-based intelligent prediction system and the deep learning neural network system on one hand; on the other hand, the operation result is input into the BP neural network system after the CNN convolution operation system is operated for the second time, verification is carried out by utilizing the following artificial intelligent design logic, and final output data is obtained after the verification is passed.
On one hand, the system has simple system constitution, strong data operation processing capability and good universality; on the other hand, the system has good autonomous operation capability and calculation result verification capability in data operation, so that the working efficiency of the operation of a design system and the precision of the design operation are greatly improved, and the labor intensity and the cost of data calculation can be reduced.
Drawings
The invention is described in detail below with reference to the drawings and the detailed description.
FIG. 1 is a schematic diagram of a system architecture of the present invention;
fig. 2 is a schematic diagram of an image information acquisition terminal structure based on an FPGA;
FIG. 3 is a flow chart of the method of the present invention.
Detailed Description
The invention is further described in connection with the following detailed description, in order to make the technical means, the creation characteristics, the achievement of the purpose and the effect of the invention easy to understand.
As shown in fig. 1 and 2, an automatic comprehensive pipe rack risk identification convolutional neural network model comprises a designer, wherein a nested architecture BP neural network system adopting a C/S structure and a B/S structure is arranged in the designer, and meanwhile, a plurality of CNN convolutional operation systems, an intelligent prediction system based on LSTM, a deep learning neural network system and a bottom layer operation system are additionally arranged in the designer, wherein the bottom layer operation system is respectively connected with the BP neural network system, the CNN convolutional operation systems, the intelligent prediction system based on LSTM and the deep learning neural network system through scheduling programs, the CNN convolutional operation systems are mutually connected in parallel and are respectively connected with an input layer of the BP neural network system, the intelligent prediction system based on LSTM and the deep learning neural network system are mutually connected with an input layer of the BP neural network system, and an output layer of the BP neural network system is additionally connected with the deep learning neural network system and one CNN convolutional operation system.
In this embodiment, the deep learning neural network system is any one or several of a neural network system based on convolution operation, a self-coding neural network based on multi-layer neurons, and a deep confidence network which performs pre-training in a multi-layer self-coding neural network mode, and further optimizes the weight of the neural network by combining with identification information.
The designer comprises an AI-based artificial intelligence logic processing module, an FPGA-based data processing unit, a GPU image processing unit, a data bus module, a clock circuit module, a driving circuit module, an I/O data communication port and a network communication port, wherein the data bus module is respectively connected with the AI-based artificial intelligence logic processing module, the FPGA data processing unit, the GPU image processing unit, the clock circuit module, the driving circuit module, the I/O data communication port and the network communication port, and the driving circuit module is further respectively connected with each GPU image processing unit, each I/O data communication port and each network communication port.
Further preferably, the I/O data communication port is further connected to at least one CCD camera, a laser scanning head, a control keyboard, and a display, respectively, and the network communication port is connected to an external communication network.
Further preferably, the number of the data processing units and the number of the GPU image processing units of the FPGA are not less than two, the data processing units of the FPGAs and the GPU image processing units are all connected in parallel, wherein the data processing units of the FPGAs are connected in series by adopting synchronous duplex circuits, and the GPU image processing units are connected in series by adopting asynchronous duplex circuits.
Further, the underlying operating system is a program system based on an SOA system.
A construction method for automatically identifying convolutional neural network model of comprehensive pipe rack risk includes the following steps;
s1, prefabricating a system, namely positioning a designer at a designated position, connecting the designer with external hardware equipment, modulating connection relations among a BP neural network system, a CNN convolution operation system, an LSTM-based intelligent prediction system and a deep learning neural network system through a bottom layer operation system, communicating one CNN convolution operation system in each CNN convolution operation system with an output layer of the BP neural network system, and connecting the CNN convolution operation system at the output end of the BP neural network system with each CNN convolution operation system connected with an input layer of the BP neural network system respectively, so that the prefabrication of the system is completed;
s2, training the system, namely, after the step S1 is completed, inputting the common algorithm participated in design operation and the data participated in calculation into the LSTM-based intelligent prediction system and the deep learning neural network system, training the BP neural network system through the LSTM-based intelligent prediction system and the deep learning neural network system, generating artificial intelligent design logic, processing and operating the common algorithm participated in design operation and the data participated in calculation through the CNN convolution operation system, transmitting the processed data into the BP neural network system, operating the received data according to the artificial intelligent design logic by the BP neural network system, outputting an operation result from an output layer, backing up an output data result, transmitting the backed up data result into the CNN convolution operation system connected with the deep learning neural network system and the BP neural network system after the current operation is completed by the BP neural network system, and updating the artificial intelligent design logic of the BP neural network system through the LSTM-based intelligent prediction system and the deep learning neural network system on one hand; on the other hand, the operation result is input into the BP neural network system after the CNN convolution operation system is operated for the second time, verification is carried out by utilizing the following artificial intelligent design logic, and final output data is obtained after the verification is passed.
On one hand, the system has simple system constitution, strong data operation processing capability and good universality; on the other hand, the system has good autonomous operation capability and calculation result verification capability in data operation, so that the working efficiency of the operation of a design system and the precision of the design operation are greatly improved, and the labor intensity and the cost of data calculation can be reduced.
It will be appreciated by those skilled in the art that the invention is not limited to the embodiments described above. The foregoing embodiments and description have been presented only to illustrate the principles of the invention. The present invention is capable of various changes and modifications without departing from its spirit and scope. Such variations and modifications are intended to fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (4)
1. A construction method for automatically identifying convolutional neural network model of comprehensive pipe rack risk is characterized by comprising the following steps: the construction method of the multichannel parallel CNN accelerator design system comprises the following steps of;
s1, prefabricating a system, namely positioning a designer at a designated position, connecting the designer with external hardware equipment, modulating connection relations among a BP neural network system, a CNN convolution operation system, an LSTM-based intelligent prediction system and a deep learning neural network system through a bottom layer operation system, communicating one CNN convolution operation system in each CNN convolution operation system with an output layer of the BP neural network system, and connecting the CNN convolution operation system at the output end of the BP neural network system with each CNN convolution operation system connected with an input layer of the BP neural network system respectively, so that the prefabrication of the system is completed;
s2, training the system, namely, after the step S1 is completed, inputting the common algorithm participated in design operation and the data participated in calculation into the LSTM-based intelligent prediction system and the deep learning neural network system, training the BP neural network system through the LSTM-based intelligent prediction system and the deep learning neural network system, generating artificial intelligent design logic, processing and operating the common algorithm participated in design operation and the data participated in calculation through the CNN convolution operation system, transmitting the processed data into the BP neural network system, operating the received data according to the artificial intelligent design logic by the BP neural network system, outputting an operation result from an output layer, backing up an output data result, transmitting the backed up data result into the CNN convolution operation system connected with the deep learning neural network system and the BP neural network system after the current operation is completed by the BP neural network system, and updating the artificial intelligent design logic of the BP neural network system through the LSTM-based intelligent prediction system and the deep learning neural network system on one hand; on the other hand, the operation result is input into the BP neural network system after the CNN convolution operation system is operated for the second time, verification is carried out by utilizing the artificial intelligence design logic after the follow-up, and final output data is obtained after the verification is passed;
the system comprises a designer, wherein an intelligent prediction system based on LSTM, a deep learning neural network system, a bottom operating system and a plurality of CNN convolution operation systems are arranged in the designer, wherein the bottom operating system is respectively connected with the BP neural network system, the CNN convolution operation system, the intelligent prediction system based on LSTM and the deep learning neural network system through scheduling programs, the CNN convolution operation systems are mutually connected in parallel and are respectively connected with an input layer of the BP neural network system, the intelligent prediction system based on LSTM and the deep learning neural network system are mutually connected and are simultaneously connected with an input layer of the BP neural network system, and an output layer of the BP neural network system is additionally connected with the deep learning neural network system and one CNN convolution operation system; the deep learning neural network system is any one or a plurality of common deep confidence networks which are used for further optimizing the weight of the neural network by combining the identification information, wherein the deep learning neural network system is a convolutional operation-based neural network system, a multi-layer neuron-based self-coding neural network and a multi-layer self-coding neural network; the designer comprises an AI-based artificial intelligence logic processing module, a data processing unit based on an FPGA, a GPU image processing unit, a data bus module, a clock circuit module, a driving circuit module, an I/O data communication port and a network communication port, wherein the data bus module is respectively connected with the AI-based artificial intelligence logic processing module, the FPGA data processing unit, the GPU image processing unit, the clock circuit module, the driving circuit module, the I/O data communication port and the network communication port, and the driving circuit module is additionally connected with each GPU image processing unit, the I/O data communication port and the network communication port.
2. The method for constructing the utility tunnel risk automatic identification convolutional neural network model according to claim 1, which is characterized in that: the I/O data communication port is connected with at least one CCD camera, a laser scanning head, a control keyboard and a display respectively, and the network communication port is connected with an external communication network.
3. The method for constructing the utility tunnel risk automatic identification convolutional neural network model according to claim 2, which is characterized in that: the data processing units of the FPGA and the GPU image processing units are not less than two, the data processing units of the FPGAs and the GPU image processing units are all connected in parallel, wherein the data processing units of the FPGAs are connected in series by adopting a synchronous duplex circuit, and the GPU image processing units are connected in series by adopting an asynchronous duplex circuit.
4. The method for constructing the utility tunnel risk automatic identification convolutional neural network model according to claim 1, which is characterized in that: the bottom operating system is a program system based on an SOA system.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011374969.7A CN112613597B (en) | 2020-11-30 | 2020-11-30 | Comprehensive pipe rack risk automatic identification convolutional neural network model and construction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011374969.7A CN112613597B (en) | 2020-11-30 | 2020-11-30 | Comprehensive pipe rack risk automatic identification convolutional neural network model and construction method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112613597A CN112613597A (en) | 2021-04-06 |
CN112613597B true CN112613597B (en) | 2023-06-30 |
Family
ID=75228110
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011374969.7A Active CN112613597B (en) | 2020-11-30 | 2020-11-30 | Comprehensive pipe rack risk automatic identification convolutional neural network model and construction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112613597B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113172931B (en) * | 2021-04-29 | 2021-11-30 | 南京晓庄学院 | Remote monitoring type intelligent control system and method for hydraulic single-row punching machine |
CN114399619A (en) * | 2022-01-14 | 2022-04-26 | 南京苏胜天信息科技有限公司 | Machine vision image recognition system and processing method thereof |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108090560A (en) * | 2018-01-05 | 2018-05-29 | 中国科学技术大学苏州研究院 | The design method of LSTM recurrent neural network hardware accelerators based on FPGA |
KR20180065498A (en) * | 2016-12-08 | 2018-06-18 | 한국항공대학교산학협력단 | Method for deep learning and method for generating next prediction image using the same |
CN109376852A (en) * | 2017-04-21 | 2019-02-22 | 上海寒武纪信息科技有限公司 | Arithmetic unit and operation method |
CN110220725A (en) * | 2019-05-30 | 2019-09-10 | 河海大学 | A kind of wheel for metro vehicle health status prediction technique integrated based on deep learning and BP |
CN111105097A (en) * | 2019-12-25 | 2020-05-05 | 河海大学常州校区 | Dam deformation prediction system and method based on convolutional neural network |
CN111832704A (en) * | 2020-06-30 | 2020-10-27 | 东南大学 | Design method of convolution input type nested recurrent neural network |
CN111860773A (en) * | 2020-06-30 | 2020-10-30 | 北京百度网讯科技有限公司 | Processing apparatus and method for information processing |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10783900B2 (en) * | 2014-10-03 | 2020-09-22 | Google Llc | Convolutional, long short-term memory, fully connected deep neural networks |
-
2020
- 2020-11-30 CN CN202011374969.7A patent/CN112613597B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20180065498A (en) * | 2016-12-08 | 2018-06-18 | 한국항공대학교산학협력단 | Method for deep learning and method for generating next prediction image using the same |
CN109376852A (en) * | 2017-04-21 | 2019-02-22 | 上海寒武纪信息科技有限公司 | Arithmetic unit and operation method |
CN108090560A (en) * | 2018-01-05 | 2018-05-29 | 中国科学技术大学苏州研究院 | The design method of LSTM recurrent neural network hardware accelerators based on FPGA |
CN110220725A (en) * | 2019-05-30 | 2019-09-10 | 河海大学 | A kind of wheel for metro vehicle health status prediction technique integrated based on deep learning and BP |
CN111105097A (en) * | 2019-12-25 | 2020-05-05 | 河海大学常州校区 | Dam deformation prediction system and method based on convolutional neural network |
CN111832704A (en) * | 2020-06-30 | 2020-10-27 | 东南大学 | Design method of convolution input type nested recurrent neural network |
CN111860773A (en) * | 2020-06-30 | 2020-10-30 | 北京百度网讯科技有限公司 | Processing apparatus and method for information processing |
Non-Patent Citations (2)
Title |
---|
Network Fault Prediction Based on CNN-LSTM Hybrid Neural Network;Zilong Tan et al;《2019 International Conference on Communications, Information System and Computer Engineering (CISCE)》;20190822;486--490 * |
基于CNN-LSTM混合神经网络模型的短期负荷预测方法;陆继翔等;《电力系统自动化》;20190425;第43卷(第8期);131-137 * |
Also Published As
Publication number | Publication date |
---|---|
CN112613597A (en) | 2021-04-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112613597B (en) | Comprehensive pipe rack risk automatic identification convolutional neural network model and construction method | |
CN109543832B (en) | Computing device and board card | |
CN111401406B (en) | Neural network training method, video frame processing method and related equipment | |
CN106027300A (en) | System and method for parameter optimization of intelligent robot applying neural network | |
CN108345934B (en) | Activation device and method for neural network processor | |
US10885620B2 (en) | Neural network image processing system | |
CN109885452A (en) | Method for monitoring performance, device and terminal device | |
CN105956660A (en) | Neural network chip realization method used for real-time image identification | |
CN111126590B (en) | Device and method for artificial neural network operation | |
CN109739703B (en) | Debugging method and related product | |
CN114356578A (en) | Parallel computing method, device, equipment and medium for natural language processing model | |
CN116643877A (en) | Computing power resource scheduling method, training method and system of computing power resource scheduling model | |
CN114715145B (en) | Trajectory prediction method, device and equipment and automatic driving vehicle | |
CN114862209A (en) | Transport capacity scheduling method and device, electronic equipment and storage medium | |
CN111831354A (en) | Data precision configuration method, device, chip array, equipment and medium | |
CN115345285A (en) | GPU-based timing chart neural network training method and system and electronic equipment | |
CN111930681B (en) | Computing device and related product | |
CN111199276B (en) | Data processing method and related product | |
CN113591709A (en) | Motion recognition method, motion recognition device, motion recognition apparatus, motion recognition medium, and computer program product | |
US20230342626A1 (en) | Model processing method and related apparatus | |
CN115766475B (en) | Semi-asynchronous power federal learning network based on communication efficiency and communication method thereof | |
CN113253684B (en) | Multi-AGV (automatic guided vehicle) scheduling method and device based on graph convolution neural network and electronic equipment | |
CN112766475B (en) | Processing component and artificial intelligence processor | |
CN103217681A (en) | Tree-shaped topological mechanism multiprocessor sonar signal processing device and method | |
Wu et al. | An efficient defect detection system for printed circuit boards with edge-cloud fusion computing |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |