WO2021227787A1 - Procédé et appareil de formation de prédicteur de réseau neuronal, et procédé et appareil de traitement d'image - Google Patents

Procédé et appareil de formation de prédicteur de réseau neuronal, et procédé et appareil de traitement d'image Download PDF

Info

Publication number
WO2021227787A1
WO2021227787A1 PCT/CN2021/088254 CN2021088254W WO2021227787A1 WO 2021227787 A1 WO2021227787 A1 WO 2021227787A1 CN 2021088254 W CN2021088254 W CN 2021088254W WO 2021227787 A1 WO2021227787 A1 WO 2021227787A1
Authority
WO
WIPO (PCT)
Prior art keywords
network structure
network
neural network
feature vector
training
Prior art date
Application number
PCT/CN2021/088254
Other languages
English (en)
Chinese (zh)
Inventor
许奕星
唐业辉
王云鹤
许春景
Original Assignee
华为技术有限公司
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 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2021227787A1 publication Critical patent/WO2021227787A1/fr

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/045Combinations of networks
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/24Classification techniques
    • G06F18/241Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods
    • G06N3/084Backpropagation, e.g. using gradient descent

Definitions

  • This application relates to the field of artificial intelligence, and more specifically, to a method, image processing method and device for training a neural network predictor.
  • Computer vision is an inseparable part of various intelligent/autonomous systems in various application fields, such as manufacturing, inspection, document analysis, medical diagnosis, and military. It is about how to use cameras/video cameras and computers to obtain What we need is the knowledge of the data and information of the subject. To put it vividly, it is to install eyes (camera/camcorder) and brain (algorithm) on the computer to replace the human eye to identify, track and measure the target, so that the computer can perceive the environment. Because perception can be seen as extracting information from sensory signals, computer vision can also be seen as a science that studies how to make artificial systems "perceive" from images or multi-dimensional data.
  • computer vision is to use various imaging systems to replace the visual organs to obtain input information, and then the computer replaces the brain to complete the processing and interpretation of the input information.
  • the ultimate research goal of computer vision is to enable computers to observe and understand the world through vision like humans, and have the ability to adapt to the environment autonomously.
  • neural networks for example, convolutional neural networks
  • the performance of a neural network is often closely related to the network structure of the neural network.
  • the neural network structure search neural architecture search, NAS
  • NAS neural architecture search
  • one method is to use a neural network predictor to predict the performance of the network structure on a specified data set.
  • This application provides a method for training a neural network predictor, an image processing method, and a device, which help improve the prediction accuracy of the neural network predictor.
  • a method for training a neural network predictor includes:
  • a first network structure of a first neural network and a second network structure of a second neural network where the first network structure is a labeled network structure, and the label is used to indicate the performance of the first network structure; acquiring The similarity between the first network structure and the second network structure; training the neural network predictor according to the first network structure, the second network structure, the similarity and the label The neural network predictor is used to predict the performance of the network structure.
  • the relationship between the first network structure and the second network structure (for example, the similarity between the first network structure and the second network structure) is used to assist the training institute
  • the neural network predictor can improve the training effect of the neural network predictor when using a small amount of labeled data (for example, at least one labeled network structure), that is, improve the training effect of the neural network predictor after training. Forecast accuracy.
  • the acquiring the similarity between the first network structure and the second network structure includes: acquiring the first network structure according to the first network structure.
  • a feature vector, the first feature vector is used to represent the first network structure;
  • a second feature vector is obtained according to the second network structure, and the second feature vector is used to represent the second network structure;
  • the first feature vector and the second feature vector obtain the similarity.
  • the similarity is obtained through the first feature vector and the second feature vector, so that the similarity can more accurately describe the relationship between the two network structures. Therefore, use The similarity can further improve the training effect of the neural network predictor, that is, improve the prediction accuracy of the trained neural network predictor.
  • the obtaining a first feature vector according to the first network structure includes: encoding the first network structure using an encoder to obtain the first network structure A feature vector, the encoder is used for encoding to obtain a feature vector representing a network structure; the obtaining a second feature vector according to the second network structure includes: encoding the second network structure using the encoder, Obtain the second feature vector.
  • the network structure is encoded by the encoder, and the feature vector of the network structure can be easily obtained.
  • the encoder may be implemented by a neural network (NN).
  • the encoder may be a recurrent neural network (RNN).
  • RNN recurrent neural network
  • the encoder is obtained after training by the following method: using a decoder to decode the second feature vector to obtain a third network structure, The decoder is used for decoding to obtain the network structure represented by the feature vector; training the encoder according to the difference between the second network structure and the third network structure.
  • the second feature vector is decoded by the decoder to obtain the third network structure, and data is not required to be annotated (for example, the feature vector output by the encoder is not required to be annotated), and according to the The difference between the second network structure and the third network structure can easily train the encoder.
  • the decoder may be trained in the process of training the encoder.
  • the encoder and the decoder may be trained at the same time according to the difference between the second network structure and the third network structure.
  • the encoder (and the decoder) can be easily trained.
  • training the encoder through a learning method can improve the accuracy of the feature vectors extracted by the encoder (of the first network structure and the second network structure), that is, it can make different
  • the feature vector of more accurately reflects the characteristics of different network structures (for example, the computing power of different network structures).
  • the training the neural network predictor according to the first network structure, the second network structure, the similarity and the label It includes: determining the performance of the first network structure according to the first feature vector, the second feature vector, and the similarity; training the nerve according to the performance of the first network structure and the label Network predictor.
  • the first feature vector, the second feature vector, and the relationship between the first network structure and the second network structure are used (for example, the relationship between the two network structures Similarity) predicting the performance of the first network structure can make (predicted) the performance of the first network structure more accurate.
  • the neural network predictor can improve the training effect of the neural network predictor, that is, the prediction accuracy of the trained neural network predictor can be improved.
  • the similarity is the distance between the first feature vector and the second feature vector.
  • the distance between the first feature vector and the second feature vector can more accurately represent the similarity between the first network structure and the second network structure. Training the neural network predictor with the similarity can further improve the training effect of the neural network predictor.
  • the neural network predictor is a graph convolutional neural network (GCN).
  • GCN graph convolutional neural network
  • the relationship between the first network structure and the second network structure can be better utilized in the training process (for example, the first network structure The similarity between the structure of the second network and the second network structure), so that the training effect of the neural network predictor can be improved.
  • the neural network predictor is used to predict the performance of the network structure of the target neural network, and the target neural network is used for image processing.
  • the target neural network may be convolutional neural networks (convolutional neural networks, CNN).
  • the target neural network can be used for image classification, image segmentation, image detection, image super-segmentation, and so on.
  • an image processing method which includes:
  • the relationship between the first network structure and the second network structure (for example, the similarity between the first network structure and the second network structure) is used to assist the training institute
  • the neural network predictor can improve the training effect of the neural network predictor when using a small amount of labeled data (for example, at least one labeled network structure), that is, improve the training effect of the neural network predictor after training. Forecast accuracy.
  • the neural network is determined according to the neural network predictor, and the use of the neural network can improve the effect of image processing.
  • the image processing may include image classification, image segmentation, image detection, image super-segmentation, and the like.
  • a device for training a neural network predictor including:
  • the first acquisition module is used to acquire the first network structure of the first neural network and the second network structure of the second neural network, the first network structure is a labeled network structure, and the label is used to indicate the first network structure.
  • Performance of a network structure a second acquisition module for acquiring the similarity between the first network structure and the second network structure; a training module for acquiring the similarity between the first network structure and the second network structure;
  • the network structure, the similarity and the label train the neural network predictor, and the neural network predictor is used to predict the performance of the network structure.
  • the relationship between multiple network structures (for example, the similarity between the first network structure and the second network structure) is used to assist in training the neural network predictor.
  • the training effect of the neural network predictor is improved, that is, the prediction accuracy of the trained neural network predictor is improved.
  • the second acquisition module is specifically configured to: acquire a first feature vector according to the first network structure, and the first feature vector is used to represent the The first network structure; the second feature vector is obtained according to the second network structure, the second feature vector is used to represent the second network structure; the second feature vector is obtained according to the first feature vector and the second feature vector The similarity.
  • the similarity is obtained through the first feature vector and the second feature vector, so that the similarity can more accurately describe the relationship between the two network structures. Therefore, use The similarity can further improve the training effect of the neural network predictor, that is, improve the prediction accuracy of the trained neural network predictor.
  • the second acquisition module is specifically configured to: use an encoder to encode the first network structure to obtain the first feature vector, and the encoding The encoder is used for encoding to obtain a feature vector representing the network structure; the encoder is used to encode the second network structure to obtain the second feature vector.
  • the network structure is encoded by the encoder, and the feature vector of the network structure can be easily obtained.
  • the encoder may be implemented by a neural network (NN).
  • the encoder may be a recurrent neural network (RNN).
  • RNN recurrent neural network
  • the encoder is obtained after training by the following method: using a decoder to decode the second feature vector to obtain a third network structure, The decoder is used for decoding to obtain the network structure represented by the feature vector; training the encoder according to the difference between the second network structure and the third network structure.
  • the second feature vector is decoded by the decoder to obtain the third network structure, and data is not required to be annotated (for example, the feature vector output by the encoder is not required to be annotated), and according to the The difference between the second network structure and the third network structure can easily train the encoder.
  • the decoder may be trained in the process of training the encoder.
  • the encoder and the decoder may be trained at the same time according to the difference between the second network structure and the third network structure.
  • the encoder (and the decoder) can be easily trained.
  • training the encoder through a learning method can improve the accuracy of the feature vectors extracted by the encoder (of the first network structure and the second network structure), that is, it can make different
  • the feature vector of more accurately reflects the characteristics of different network structures (for example, the computing power of different network structures).
  • the training module is specifically configured to: determine the first network according to the first feature vector, the second feature vector, and the similarity The performance of the structure; training the neural network predictor according to the performance of the first network structure and the label.
  • the first feature vector, the second feature vector, and the relationship between the first network structure and the second network structure are used (for example, the relationship between the two network structures Similarity) predicting the performance of the first network structure can make (predicted) the performance information of the first network structure more accurate.
  • the neural network predictor according to the performance of the first network structure and the label Training the neural network predictor can improve the training effect of the neural network predictor, that is, improve the prediction accuracy of the trained neural network predictor.
  • the similarity is the distance between the first feature vector and the second feature vector.
  • the distance between the first feature vector and the second feature vector can more accurately represent the similarity between the first network structure and the second network structure. Training the neural network predictor with the similarity can further improve the training effect of the neural network predictor.
  • the neural network predictor is a graph convolutional neural network.
  • the relationship between the first network structure and the second network structure can be better utilized in the training process (for example, the first network structure The similarity between the structure of the second network and the second network structure), so that the training effect of the neural network predictor can be improved.
  • the neural network predictor is used to predict the performance of the network structure of the target neural network, and the target neural network is used for image processing.
  • the target neural network may be convolutional neural networks (convolutional neural networks, CNN).
  • the target neural network can be used for image classification, image segmentation, image detection, image super-segmentation, and so on.
  • an image processing device including:
  • the acquisition module is used to obtain the image to be processed; the image processing module is used to perform image processing on the image to be processed using a neural network; wherein the neural network is determined according to a neural network predictor, the neural network predictor It is obtained after training by the method in any one of the above-mentioned first aspects.
  • the relationship between the first network structure and the second network structure (for example, the similarity between the first network structure and the second network structure) is used to assist the training institute
  • the neural network predictor can improve the training effect of the neural network predictor when using a small amount of labeled data (for example, at least one labeled network structure), that is, improve the training effect of the neural network predictor after training. Forecast accuracy.
  • the neural network is determined according to the neural network predictor, and the use of the neural network can improve the effect of image processing.
  • the image processing may include image classification, image segmentation, image detection, image super-segmentation, and the like.
  • a method for training a neural network includes:
  • a first network structure of a first neural network and a second network structure of a second neural network where the first network structure is a labeled network structure, and the label is used to indicate the performance of the first network structure; acquiring The similarity between the first network structure and the second network structure; training the neural network according to the first network structure, the second network structure, the similarity and the label.
  • the relationship between the first network structure and the second network structure (for example, the similarity between the first network structure and the second network structure) is used to assist the training institute
  • the neural network can improve the training effect of the neural network while using a small amount of labeled data (for example, at least one network structure with performance labels).
  • the acquiring the similarity between the first network structure and the second network structure includes: acquiring the first network structure according to the first network structure.
  • a feature vector, the first feature vector is used to represent the first network structure;
  • a second feature vector is obtained according to the second network structure, and the second feature vector is used to represent the second network structure;
  • the first feature vector and the second feature vector obtain the similarity.
  • the similarity is obtained through the first feature vector and the second feature vector, so that the similarity can more accurately describe the relationship between the two network structures. Therefore, use The similarity can further improve the training effect of the neural network.
  • the obtaining the first feature vector according to the first network structure includes: encoding the first network structure using an encoder to obtain the first network structure A feature vector, the encoder is used for encoding to obtain a feature vector representing a network structure; the obtaining a second feature vector according to the second network structure includes: encoding the second network structure using the encoder, Obtain the second feature vector.
  • the network structure is encoded by the encoder, and the feature vector of the network structure can be easily obtained.
  • the encoder may be implemented by a neural network (NN).
  • the encoder may be a recurrent neural network (RNN).
  • RNN recurrent neural network
  • the encoder is obtained after training by the following method: using a decoder to decode the second feature vector to obtain a third network structure, The decoder is used for decoding to obtain the network structure represented by the feature vector; training the encoder according to the difference between the second network structure and the third network structure.
  • the second feature vector is decoded by the decoder to obtain the third network structure, and data is not required to be annotated (for example, the feature vector output by the encoder is not required to be annotated), and according to the The difference between the second network structure and the third network structure can easily train the encoder.
  • the decoder may be trained in the process of training the encoder.
  • the encoder and the decoder may be trained at the same time according to the difference between the second network structure and the third network structure.
  • the encoder (and the decoder) can be easily trained.
  • training the encoder through a learning method can improve the accuracy of the feature vectors extracted by the encoder (of the first network structure and the second network structure), that is, it can make different
  • the feature vector of more accurately reflects the characteristics of different network structures (for example, the computing power of different network structures).
  • the training the neural network predictor according to the first network structure, the second network structure, the similarity and the label It includes: determining the performance of the first network structure according to the first feature vector, the second feature vector, and the similarity; training the nerve according to the performance of the first network structure and the label Network predictor.
  • the first feature vector, the second feature vector, and the relationship between the first network structure and the second network structure are used (for example, the relationship between the two network structures Similarity) predicting the performance of the first network structure can make (predicted) the performance of the first network structure more accurate.
  • the neural network can improve the training effect of the neural network.
  • the similarity is the distance between the first feature vector and the second feature vector.
  • the distance between the first feature vector and the second feature vector can more accurately represent the similarity between the first network structure and the second network structure. Training the neural network with the similarity can further improve the training effect of the neural network.
  • the neural network is a graph convolutional network (GCN).
  • GCN graph convolutional network
  • the relationship between the first network structure and the second network structure can be better utilized in the training process (for example, the first network structure And the similarity between the structure of the second network), so that the training effect of the neural network can be improved.
  • the neural network is used to predict the performance of the network structure of the target neural network, and the target neural network is used for image processing.
  • the target neural network may be convolutional neural networks (convolutional neural networks, CNN).
  • the target neural network can be used for image classification, image segmentation, image detection, image super-segmentation, and so on.
  • a device for training a neural network predictor includes: a memory for storing a program; a processor for executing the program stored in the memory, and when the program stored in the memory is executed , The processor is configured to execute the method in any one of the implementation manners in the first aspect.
  • the processor in the sixth aspect mentioned above can be either a central processing unit (CPU), or a combination of a CPU and a neural network computing processor.
  • the neural network computing processor here can include a graphics processing unit (graphics processing unit). unit, GPU), neural-network processing unit (NPU), tensor processing unit (TPU), and so on.
  • graphics processing unit graphics processing unit
  • NPU neural-network processing unit
  • TPU tensor processing unit
  • TPU is an artificial intelligence accelerator application specific integrated circuit fully customized by Google for machine learning.
  • an image processing device in a seventh aspect, includes: a memory for storing a program; a processor for executing the program stored in the memory, and when the program stored in the memory is executed, the processing The device is used to execute the method in any one of the implementation manners in the second aspect.
  • the processor in the seventh aspect mentioned above can be either a central processing unit (CPU) or a combination of a CPU and a neural network computing processor.
  • the neural network computing processor here can include a graphics processing unit (graphics processing unit). unit, GPU), neural-network processing unit (NPU), tensor processing unit (TPU), and so on.
  • graphics processing unit graphics processing unit
  • NPU neural-network processing unit
  • TPU tensor processing unit
  • TPU is an artificial intelligence accelerator application specific integrated circuit fully customized by Google for machine learning.
  • a device for training a neural network includes: a memory for storing a program; a processor for executing the program stored in the memory.
  • the device The processor is used to execute the method in any one of the implementation manners of the fifth aspect.
  • the processor in the eighth aspect mentioned above can be either a central processing unit (CPU), or a combination of a CPU and a neural network computing processor.
  • the neural network computing processor here can include a graphics processing unit (graphics processing unit). unit, GPU), neural-network processing unit (NPU), tensor processing unit (TPU), and so on.
  • graphics processing unit graphics processing unit
  • NPU neural-network processing unit
  • TPU tensor processing unit
  • TPU is an artificial intelligence accelerator application specific integrated circuit fully customized by Google for machine learning.
  • a computer-readable medium stores program code for device execution.
  • the program code includes an implementation for executing the first aspect, the second aspect, or the third aspect. The method in the way.
  • a computer program product containing instructions is provided.
  • the computer program product runs on a computer, the computer executes the method in any one of the foregoing first aspect or second aspect.
  • a chip in an eleventh aspect, includes a processor and a data interface.
  • the processor reads instructions stored in a memory through the data interface and executes the first aspect or the second aspect or the third aspect. The method in any one of the implementation modes.
  • the chip may further include a memory in which instructions are stored, and the processor is configured to execute instructions stored on the memory.
  • the processor is configured to execute the method in any one of the implementation manners of the first aspect, the second aspect, or the third aspect.
  • the aforementioned chip may specifically be a field-programmable gate array (FPGA) or an application-specific integrated circuit (ASIC).
  • FPGA field-programmable gate array
  • ASIC application-specific integrated circuit
  • the relationship between the first network structure and the second network structure (for example, the similarity between the first network structure and the second network structure) is used to assist the training institute
  • the neural network predictor can improve the training effect of the neural network predictor when using a small amount of labeled data (for example, at least one labeled network structure), that is, improve the training effect of the neural network predictor after training. Forecast accuracy.
  • FIG. 1 is a schematic structural diagram of a system architecture provided by an embodiment of the present application.
  • Fig. 2 is a schematic structural diagram of a convolutional neural network provided by an embodiment of the present application.
  • FIG. 3 is a schematic diagram of a chip hardware structure provided by an embodiment of the present application.
  • FIG. 4 is a schematic structural diagram of another system architecture provided by an embodiment of the present application.
  • Fig. 5 is a schematic flowchart of a method for training a neural network predictor provided by an embodiment of the present application.
  • Fig. 6 is a schematic flowchart of a method for training a neural network predictor provided by another embodiment of the present application.
  • Fig. 7 is a schematic block diagram of a method for training a neural network predictor provided by an embodiment of the present application.
  • FIG. 8 is a schematic flowchart of an image processing method provided by an embodiment of the present application.
  • Fig. 9 is a schematic block diagram of an apparatus for training a neural network predictor provided by an embodiment of the present application.
  • Fig. 10 is a schematic block diagram of an image processing device provided by an embodiment of the present application.
  • the embodiments of this application can be applied in many fields of artificial intelligence, for example, smart manufacturing, smart transportation, smart home, smart medical, smart security, automatic driving, smart cities and other fields.
  • the embodiments of the present application can be applied to photographing, video recording, smart city, human-computer interaction, and other scenes that require image processing, such as image classification, image segmentation, image detection, and image super-segmentation.
  • the method of training a neural network predictor in the embodiments of this application can be applied to neural network architecture search (NAS), and the trained neural network predictor can quickly and accurately predict the performance of the network structure. This saves the time spent searching for the neural network structure.
  • the network structure obtained through the neural network structure search can be used to construct a neural network applied to image processing scenes to improve the effect of image processing.
  • the neural network constructed by the method in the embodiments of the present application can be applied to the scene of image classification, and the use of the neural network can improve the accuracy of image classification and the efficiency of image classification, thereby improving user experience.
  • the neural network constructed by the method in the embodiments of the present application can be applied to the scene of image recognition, and the use of the neural network can improve the accuracy of image recognition and the efficiency of image recognition, thereby improving user experience.
  • the method in the embodiment of the present application is not limited to the above two scenarios when applied, and the neural network constructed by the method in the embodiment of the present application can also be used for photographing, video recording, smart city, human-computer interaction, and other needs.
  • Scenarios for image processing such as image classification, image segmentation, image detection, image super-segmentation, etc.
  • the method of training a neural network predictor in the embodiment of this application can also be applied to other scenarios that need to predict the performance of the network structure, or the method of training a neural network predictor in the embodiment of this application can also be applied to other training neural networks.
  • Network scenarios, or the method of training neural network predictors in the embodiments of this application can also be applied to other scenarios that require the use of neural networks (for example, speech recognition, machine translation, semantic segmentation, etc.). This is not limited.
  • the image in the embodiments of this application can be a static image (or called a static picture) or a dynamic image (or called a dynamic picture).
  • the image in this application can be a video or a dynamic picture.
  • the images in this application can also be static pictures or photos.
  • static images or dynamic images are collectively referred to as images.
  • the embodiments of the present application involve a large number of related applications of neural networks.
  • a neural network can be composed of neural units.
  • a neural unit can refer to an arithmetic unit that takes x s and intercept 1 as inputs.
  • the output of the arithmetic unit can be as shown in the following formula:
  • s 1, 2,...n, n is a natural number greater than 1
  • W s is the weight of x s
  • b is the bias of the neural unit.
  • f is the activation function of the neural unit, which is used to introduce nonlinear characteristics into the neural network to convert the input signal in the neural unit into an output signal.
  • the output signal of the activation function can be used as the input of the next convolutional layer, and the activation function can be a sigmoid function.
  • a neural network is a network formed by connecting multiple above-mentioned single neural units together, that is, the output of one neural unit can be the input of another neural unit.
  • the input of each neural unit can be connected with the local receptive field of the previous layer to extract the characteristics of the local receptive field.
  • the local receptive field can be a region composed of several neural units.
  • Deep neural network also known as multi-layer neural network
  • the DNN is divided according to the positions of different layers.
  • the neural network inside the DNN can be divided into three categories: input layer, hidden layer, and output layer.
  • the first layer is the input layer
  • the last layer is the output layer
  • the number of layers in the middle are all hidden layers.
  • the layers are fully connected, that is to say, any neuron in the i-th layer must be connected to any neuron in the i+1th layer.
  • DNN looks complicated, it is not complicated in terms of the work of each layer. Simply put, it is the following linear relationship expression: in, Is the input vector, Is the output vector, Is the offset vector, W is the weight matrix (also called coefficient), and ⁇ () is the activation function.
  • Each layer is just the input vector After such a simple operation, the output vector is obtained Due to the large number of DNN layers, the coefficient W and the offset vector The number is also relatively large.
  • DNN The definition of these parameters in DNN is as follows: Take coefficient W as an example: Suppose in a three-layer DNN, the linear coefficients from the fourth neuron in the second layer to the second neuron in the third layer are defined as The superscript 3 represents the number of layers where the coefficient W is located, and the subscript corresponds to the output third-level index 2 and the input second-level index 4.
  • the coefficient from the kth neuron in the L-1th layer to the jth neuron in the Lth layer is defined as
  • Convolutional neural network (convolutional neuron network, CNN) is a deep neural network with a convolutional structure.
  • the convolutional neural network contains a feature extractor composed of a convolutional layer and a sub-sampling layer.
  • the feature extractor can be regarded as a filter.
  • the convolutional layer refers to the neuron layer that performs convolution processing on the input signal in the convolutional neural network.
  • a neuron can be connected to only part of the neighboring neurons.
  • a convolutional layer usually contains several feature planes, and each feature plane can be composed of some rectangularly arranged neural units. Neural units in the same feature plane share weights, and the shared weights here are the convolution kernels.
  • Sharing weight can be understood as the way of extracting image information has nothing to do with location.
  • the convolution kernel can be initialized in the form of a matrix of random size. In the training process of the convolutional neural network, the convolution kernel can obtain reasonable weights through learning. In addition, the direct benefit of sharing weights is to reduce the connections between the layers of the convolutional neural network, and at the same time reduce the risk of overfitting.
  • RNN Recurrent Neural Networks
  • RNN can process sequence data of any length.
  • the training of RNN is the same as the training of traditional CNN or DNN.
  • the neural network can use the back propagation (BP) algorithm to modify the size of the parameters in the initial neural network model during the training process, so that the reconstruction error loss of the neural network model becomes smaller and smaller. Specifically, forwarding the input signal to the output will cause error loss, and the parameters in the initial neural network model are updated by backpropagating the error loss information, so that the error loss is converged.
  • the back-propagation algorithm is a back-propagation motion dominated by error loss, and aims to obtain the optimal parameters of the neural network model, such as the weight matrix.
  • the pixel value of the image can be a red-green-blue (RGB) color value, and the pixel value can be a long integer representing the color.
  • the pixel value is 256*Red+100*Green+76Blue, where * represents a multiplication operation, Blue represents the blue component, Green represents the green component, and Red represents the red component. In each color component, the smaller the value, the lower the brightness, and the larger the value, the higher the brightness.
  • the pixel values can be grayscale values.
  • an embodiment of the present application provides a system architecture 100.
  • a data collection device 160 is used to collect training data.
  • the training data may include an unlabeled network structure, an labeled network structure, and a ground truth (GT) corresponding to the labeled network structure, where the labeled network structure
  • GT ground truth
  • the corresponding truth value may be the performance of the labeling network structure (for example, the performance of the labeling network structure on a specified data set) that is manually pre-labeled.
  • the data collection device 160 stores the training data in the database 130, and the training device 120 trains to obtain the target model/rule 101 based on the training data maintained in the database 130.
  • the training device 120 processes the input network structure (for example, the unlabeled network structure and the labeled network structure) to obtain the performance of the network structure.
  • the performance of the network structure is compared with the true value corresponding to the network structure (for example, the performance of the labeled network structure is compared with the true value of the labeled network structure), until the performance of the network structure output by the training device 120 is compared with the true value of the network structure.
  • the difference between the two is smaller than a certain threshold, thereby completing the training of the target model/rule 101 (ie, neural network predictor).
  • the above-mentioned target model/rule 101 can be used to implement the neural network predictor obtained after training, that is, the network structure is input into the target model/rule 101 after relevant preprocessing, and the performance of the network structure can be predicted.
  • the predicted performance of the network structure can be used to determine the neural network, and the determined neural network can be used for image processing.
  • the training data maintained in the database 130 may not all come from the collection of the data collection device 160, and may also be received from other devices.
  • the training device 120 does not necessarily perform the training of the target model/rule 101 completely based on the training data maintained by the database 130. It may also obtain training data from the cloud or other places for model training. The above description should not be used as a reference to this application. Limitations of the embodiment.
  • the target model/rule 101 trained according to the training device 120 can be applied to different systems or devices, such as the execution device 110 shown in FIG. 1, which can be a terminal, such as a mobile phone terminal, a tablet computer, notebook computers, augmented reality (AR)/virtual reality (VR), vehicle-mounted terminals, etc., can also be servers or cloud devices.
  • the execution device 110 is configured with an input/output (input/output, I/O) interface 112 for data interaction with external devices.
  • the user can input data to the I/O interface 112 through the client device 140.
  • the input data in this embodiment of the application may include: the network structure input by the client device.
  • the preprocessing module 113 and the preprocessing module 114 are used to perform preprocessing according to the input data (such as network structure) received by the I/O interface 112.
  • the preprocessing module 113 and the preprocessing module 114 may not be provided. (There can also be only one of the preprocessing modules), and the calculation module 111 is directly used to process the input data.
  • the execution device 110 may call data, codes, etc. in the data storage system 150 for corresponding processing .
  • the data, instructions, etc. obtained by corresponding processing may also be stored in the data storage system 150.
  • the I/O interface 112 returns the processing result, such as the performance of the network structure obtained above, to the client device 140, so as to provide it to the user.
  • the training device 120 can generate corresponding target models/rules 101 based on different training data for different goals or tasks, and the corresponding target models/rules 101 can be used to achieve the above goals or complete The above tasks provide users with the desired results.
  • the target model/rule 101 in the embodiment of the present application may specifically be the image processing device in the embodiment of the present application, and the image processing device may be determined according to the performance of the network structure predicted by the neural network predictor.
  • the training data may include the image to be processed and the truth value corresponding to the image to be processed.
  • the user can manually set input data, and the manual setting can be operated through the interface provided by the I/O interface 112.
  • the client device 140 can automatically send input data to the I/O interface 112. If the client device 140 is required to automatically send the input data and the user's authorization is required, the user can set the corresponding authority in the client device 140.
  • the user can view the result output by the execution device 110 on the client device 140, and the specific presentation form may be a specific manner such as display, sound, and action.
  • the client device 140 can also be used as a data collection terminal to collect the input data of the input I/O interface 112 and the output result of the output I/O interface 112 as new sample data and store it in the database 130 as shown in the figure.
  • the I/O interface 112 directly uses the input data input to the I/O interface 112 and the output result of the output I/O interface 112 as a new sample as shown in the figure.
  • the data is stored in the database 130.
  • FIG. 1 is only a schematic diagram of a system architecture provided by an embodiment of the present application, and the positional relationship between the devices, devices, modules, etc. shown in the figure does not constitute any limitation.
  • the data The storage system 150 is an external memory relative to the execution device 110. In other cases, the data storage system 150 may also be placed in the execution device 110.
  • the target model/rule 101 is obtained by training according to the training device 120.
  • the target model/rule 101 may be a neural network obtained after training based on the method of training a neural network predictor in this application.
  • the predictor, or the target model/rule 101 in the embodiment of the present application may also be the image processing device in the embodiment of the present application.
  • a neural network predictor obtained after training based on the method of training a neural network predictor in the present application can be used for searching a neural network, and the neural network can be used for image processing, speech processing, natural language processing, and the like.
  • the neural network predictor can be used to search for convolutional neural networks (CNN), deep convolutional neural networks (DCNN), and/or recurrent neural networks (RNNS), etc. Wait.
  • CNN convolutional neural networks
  • DCNN deep convolutional neural networks
  • RNNS recurrent neural networks
  • CNN is a very common neural network
  • the structure of CNN will be introduced in detail below in conjunction with Figure 2 and taking image processing as an example.
  • a convolutional neural network is a deep neural network with a convolutional structure. It is a deep learning architecture.
  • the deep learning architecture refers to the algorithm of machine learning. Multi-level learning is carried out on the abstract level of.
  • CNN is a feed-forward artificial neural network, and each neuron in the feed-forward artificial neural network can respond to input data (for example, images).
  • a convolutional neural network (CNN) 200 may include an input layer 210, a convolutional layer/pooling layer 220 (the pooling layer is optional), and a neural network layer 230.
  • CNN convolutional neural network
  • the convolutional layer/pooling layer 220 as shown in FIG. 2 may include layers 221-226 as shown in Examples.
  • layer 221 is a convolutional layer
  • layer 222 is a pooling layer
  • layer 223 is Convolutional layer
  • 224 is a pooling layer
  • 225 is a convolutional layer
  • 226 is a pooling layer
  • 221 and 222 are convolutional layers
  • 223 is a pooling layer
  • 224 and 225 are convolutions.
  • the accumulation layer, 226 is the pooling layer. That is, the output of the convolutional layer can be used as the input of the subsequent pooling layer, or as the input of another convolutional layer to continue the convolution operation.
  • the convolution layer 221 can include many convolution operators.
  • the convolution operator is also called a kernel. Its function in image processing is equivalent to a filter that extracts specific information from the input image matrix.
  • the convolution operator is essentially It can be a weight matrix. This weight matrix is usually pre-defined. In the process of convolution on the image, the weight matrix is usually one pixel after one pixel (or two pixels after two pixels) along the horizontal direction on the input image. ...It depends on the value of stride) to complete the work of extracting specific features from the image.
  • the size of the weight matrix should be related to the size of the image. It should be noted that the depth dimension of the weight matrix and the depth dimension of the input image are the same.
  • the weight matrix will extend to Enter the entire depth of the image. Therefore, convolution with a single weight matrix will produce a single depth dimension convolution output, but in most cases, a single weight matrix is not used, but multiple weight matrices of the same size (row ⁇ column) are applied. That is, multiple homogeneous matrices.
  • the output of each weight matrix is stacked to form the depth dimension of the convolutional image, where the dimension can be understood as determined by the "multiple" mentioned above.
  • Different weight matrices can be used to extract different features in the image. For example, one weight matrix is used to extract edge information of the image, another weight matrix is used to extract specific colors of the image, and another weight matrix is used to eliminate unwanted noise in the image.
  • the multiple weight matrices have the same size (row ⁇ column), and the feature maps extracted by the multiple weight matrices of the same size have the same size, and then the multiple extracted feature maps of the same size are combined to form a convolution operation. Output.
  • weight values in these weight matrices need to be obtained through a lot of training in practical applications.
  • Each weight matrix formed by the weight values obtained through training can be used to extract information from the input image, so that the convolutional neural network 200 can make correct predictions. .
  • the initial convolutional layer (such as 221) often extracts more general features, which can also be called low-level features; with the convolutional neural network
  • the features extracted by the subsequent convolutional layers (for example, 226) become more and more complex, such as features such as high-level semantics, and features with higher semantics are more suitable for the problem to be solved.
  • the pooling layer can be a convolutional layer followed by a layer.
  • the pooling layer can also be a multi-layer convolutional layer followed by one or more pooling layers.
  • the sole purpose of the pooling layer is to reduce the size of the image space.
  • the pooling layer may include an average pooling operator and/or a maximum pooling operator for sampling the input image to obtain an image with a smaller size.
  • the average pooling operator can calculate the pixel values in the image within a specific range to generate an average value as the result of the average pooling.
  • the maximum pooling operator can take the pixel with the largest value within a specific range as the result of the maximum pooling.
  • the operators in the pooling layer should also be related to the image size.
  • the size of the image output after processing by the pooling layer can be smaller than the size of the image of the input pooling layer, and each pixel in the image output by the pooling layer represents the average value or the maximum value of the corresponding sub-region of the image input to the pooling layer.
  • the convolutional neural network 200 After processing by the convolutional layer/pooling layer 220, the convolutional neural network 200 is not enough to output the required output information. Because as mentioned above, the convolutional layer/pooling layer 220 only extracts features and reduces the parameters brought by the input image. However, in order to generate the final output information (required class information or other related information), the convolutional neural network 200 needs to use the neural network layer 230 to generate one or a group of required classes of output. Therefore, the neural network layer 230 may include multiple hidden layers (231, 232 to 23n as shown in FIG. 2) and an output layer 240. The parameters contained in the multiple hidden layers can be based on specific task types. The relevant training data of the, for example, the task type can include image recognition, image classification, image super-resolution reconstruction and so on.
  • the output layer 240 After the multiple hidden layers in the neural network layer 230, that is, the final layer of the entire convolutional neural network 200 is the output layer 240.
  • the output layer 240 has a loss function similar to the classification cross entropy, which is specifically used to calculate the prediction error.
  • the convolutional neural network 200 shown in FIG. 2 is only used as an example of a convolutional neural network. In specific applications, the convolutional neural network may also exist in the form of other network models.
  • the neural network predictor obtained after training based on the method of training neural network predictor in this application can be used to search for (the network structure of) the neural network, and the neural network obtained through the neural network structure search can include The convolutional neural network 200 shown in FIG. 2; or, the image processing device in the embodiment of the present application may include the convolutional neural network 200 shown in FIG. Image processing result.
  • FIG. 3 is a hardware structure of a chip provided by an embodiment of the application, and the chip includes a neural network processor 50.
  • the chip may be set in the execution device 110 as shown in FIG. 1 to complete the calculation work of the calculation module 111.
  • the chip can also be set in the training device 120 as shown in FIG. 1 to complete the training work of the training device 120 and output the target model/rule 101.
  • the algorithms of each layer in the convolutional neural network as shown in Figure 2 can be implemented in the chip as shown in Figure 3.
  • the neural network processor NPU 50 is mounted on the host CPU (host CPU) as a coprocessor, and the host CPU distributes tasks.
  • the core part of the NPU is the arithmetic circuit 503.
  • the controller 504 controls the arithmetic circuit 503 to extract data from the memory (weight memory or input memory) and perform calculations.
  • the arithmetic circuit 503 includes multiple processing units (process engines, PE). In some implementations, the arithmetic circuit 503 is a two-dimensional systolic array. The arithmetic circuit 503 may also be a one-dimensional systolic array or other electronic circuit capable of performing mathematical operations such as multiplication and addition. In some implementations, the arithmetic circuit 503 is a general-purpose matrix processor.
  • the arithmetic circuit 503 fetches the data corresponding to matrix B from the weight memory 502 and buffers it on each PE in the arithmetic circuit 503.
  • the arithmetic circuit 503 fetches the matrix A data and matrix B from the input memory 501 to perform matrix operations, and the partial result or final result of the obtained matrix is stored in an accumulator 508.
  • the vector calculation unit 507 can perform further processing on the output of the arithmetic circuit 503, such as vector multiplication, vector addition, exponential operation, logarithmic operation, size comparison, and so on.
  • the vector calculation unit 507 can be used for network calculations in the non-convolutional/non-FC layer of the neural network, such as pooling, batch normalization, local response normalization, etc. .
  • the vector calculation unit 507 can store the processed output vector to the unified buffer 506.
  • the vector calculation unit 507 may apply a nonlinear function to the output of the arithmetic circuit 503, such as a vector of accumulated values, to generate the activation value.
  • the vector calculation unit 507 generates a normalized value, a combined value, or both.
  • the processed output vector can be used as an activation input to the arithmetic circuit 503, for example for use in a subsequent layer in a neural network.
  • the unified memory 506 is used to store input data and output data.
  • the weight data directly transfers the input data in the external memory to the input memory 501 and/or the unified memory 506 through the storage unit access controller 505 (direct memory access controller, DMAC), and stores the weight data in the external memory into the weight memory 502, And the data in the unified memory 506 is stored in the external memory.
  • DMAC direct memory access controller
  • the bus interface unit (BIU) 510 is used to implement interaction between the main CPU, the DMAC, and the instruction fetch memory 509 through the bus.
  • An instruction fetch buffer 509 connected to the controller 504 is used to store instructions used by the controller 504;
  • the controller 504 is used to call the instructions cached in the memory 509 to control the working process of the computing accelerator.
  • the unified memory 506, the input memory 501, the weight memory 502, and the instruction fetch memory 509 are all on-chip (On-Chip) memories.
  • the external memory is a memory external to the NPU.
  • the external memory can be a double data rate synchronous dynamic random access memory.
  • Memory double data rate synchronous dynamic random access memory, DDR SDRAM), high bandwidth memory (HBM) or other readable and writable memory.
  • each layer in the convolutional neural network shown in FIG. 2 can be executed by the arithmetic circuit 503 or the vector calculation unit 307.
  • the training device 120 in FIG. 1 introduced above can execute each step of the method for training a neural network predictor according to an embodiment of the present application, and the execution device 110 in FIG. 1 can execute each step of the image processing method according to an embodiment of the present application.
  • the CNN model shown in FIG. 2 and the chip shown in FIG. 3 can also be used to execute each step of the image processing method of the embodiment of the present application, and the chip shown in FIG. 3 can also be used to execute the training neural network of the embodiment of the present application.
  • the various steps of the predictor method can be used to execute each step of the image processing method of the embodiment of the present application.
  • an embodiment of the present application provides a system architecture 300.
  • the system architecture includes a local device 301, a local device 302, an execution device 210 and a data storage system 250, where the local device 301 and the local device 302 are connected to the execution device 210 through a communication network.
  • the execution device 210 may be implemented by one or more servers.
  • the execution device 210 can be used in conjunction with other computing devices, such as data storage, routers, load balancers and other devices.
  • the execution device 210 may be arranged on one physical site or distributed on multiple physical sites.
  • the execution device 210 may use the data in the data storage system 250 or call the program code in the data storage system 250 to implement the method for training a neural network predictor or the image processing method in the embodiment of the present application.
  • the execution device 210 may execute the following process:
  • a first network structure of a first neural network and a second network structure of a second neural network where the first network structure is a labeled network structure, and the label is used to indicate the performance of the first network structure; acquiring The similarity between the first network structure and the second network structure; training the neural network predictor according to the first network structure, the second network structure, the similarity and the label The neural network predictor is used to predict the performance of the network structure.
  • a neural network predictor can be built.
  • the neural network predictor can be used for searching neural networks, and the neural network can be used for image processing, speech processing, natural language processing, and the like.
  • the execution device 210 may also execute the following process:
  • One of the methods described is obtained after training.
  • an image processing device can be constructed, and the image processing device can be used for image processing.
  • Each local device can represent any computing device, such as personal computers, computer workstations, smart phones, tablets, smart cameras, smart cars or other types of cellular phones, media consumption devices, wearable devices, set-top boxes, game consoles, etc.
  • the local device of each user can interact with the execution device 210 through a communication network of any communication mechanism/communication standard.
  • the communication network can be a wide area network, a local area network, a point-to-point connection, etc., or any combination thereof.
  • the local device 301 and the local device 302 obtain the relevant parameters of the neural network predictor from the execution device 210, deploy the neural network predictor on the local device 301 and the local device 302, and use the neural network predictor Predict the performance of the network structure.
  • a neural network predictor can be directly deployed on the execution device 210, and the execution device 210 obtains the network structure from the local device 301 and the local device 302, and uses the neural network predictor to predict the performance of the network structure.
  • the local equipment 301 and the local equipment 302 obtain the relevant parameters of the image processing apparatus from the execution equipment 210, deploy the image processing apparatus on the local equipment 301 and the local equipment 302, and use the image processing apparatus to process the image Perform image processing.
  • an image processing device may be directly deployed on the execution device 210, and the execution device 210 obtains the image to be processed from the local device 301 and the local device 302, and uses the image processing device to perform image processing on the image to be processed.
  • the above-mentioned execution device 210 may also be a cloud device.
  • the execution device 210 may be deployed in the cloud; or, the above-mentioned execution device 210 may also be a terminal device. In this case, the execution device 210 may be deployed on the user terminal side.
  • the embodiments of the present application are not limited to this.
  • Fig. 5 is a schematic flow chart of the method for training a neural network predictor of this application.
  • the method 500 for training a neural network predictor in FIG. 5 may include step 510, step 520, and step 530.
  • the method 500 may be executed by the execution device 120 in FIG. 1, the chip shown in FIG. 3, and the execution device 210 in FIG. 4 and other devices.
  • S510 Acquire a first network structure of the first neural network and a second network structure of the second neural network.
  • the first network structure may be a labeled network structure, and the label may be used to indicate the performance of the first network structure.
  • the label here can be understood as a real label corresponding to the network structure, and the label can be used to represent the real performance corresponding to the network structure.
  • a neural network composed of a network structure can be trained, and when the neural network is trained to converge, the true performance corresponding to the network structure can be determined according to the converged neural network, that is, the performance label corresponding to the network structure can be obtained.
  • the network structure with the performance label can be obtained by manual labeling.
  • the second network structure may be an untagged network structure.
  • multiple network structures may also be acquired in the above S510, where the multiple network structures may include a small number of network structures with performance labels and a large number of network structures without performance labels.
  • the degree of similarity may be used to indicate the degree of similarity between the first network structure and the second network structure.
  • multiple network structures can be acquired in the above S510.
  • the similarity between some of the multiple network structures (for example, at least two network structures) can be acquired.
  • it is also possible to obtain the similarity between all the network structures in the multiple network structures which is not limited in the embodiment of the present application.
  • the similarity between any two network structures in the plurality of network structures can be obtained.
  • a first feature vector can be obtained according to the first network structure; a second feature vector can be obtained according to the second network structure, and the second feature vector is used to represent the second network structure; according to the The first feature vector and the second feature vector obtain the similarity.
  • the first feature vector can be used to represent the first network structure, for example, the first feature vector can represent the characteristics of the first network structure (ie, network features), or the first feature vector can be used Yu represents the computing power of the first network structure.
  • the second feature vector may be used to represent the second network structure, for example, the second feature vector may be used to represent the characteristics of the second network structure (ie, network features), or the second feature vector may be used to represent The computing power of the second network structure.
  • feature vectors of different network structures can more accurately reflect the characteristics of different network structures (for example, the computing capabilities of different network structures).
  • the feature vector can be used to represent the network structure, and the feature vector is also convenient for a neural network (for example, a neural network predictor) to process the network structure (represented by the feature vector).
  • a neural network for example, a neural network predictor
  • the similarity is obtained through the first feature vector and the second feature vector, so that the similarity can more accurately describe the relationship between the two network structures. Therefore, use The similarity can further improve the training effect of the neural network predictor, that is, improve the prediction accuracy of the trained neural network predictor.
  • the first feature vector of the first network structure may be embedding or a vector feature similar to embedding.
  • the second feature vector of the second network structure is similar to the first feature vector of the first network structure, and will not be repeated here.
  • the similarity may be the distance between the first feature vector and the second feature vector.
  • the similarity may be the cosine distance between the first feature vector and the second feature vector.
  • the distance between the first feature vector and the second feature vector can more accurately represent the similarity between the first network structure and the second network structure. Training the neural network predictor with the similarity can further improve the training effect of the neural network predictor.
  • an encoder may be used to encode the first network structure to obtain the first feature vector; the encoder may be used to encode the second network structure to obtain the second feature vector.
  • the encoder may be used for encoding to obtain a feature vector representing the network structure, and the encoder may be implemented by a neural network (NN).
  • NN neural network
  • the encoder may be an encoder in an auto-encoder (AE), and the auto-encoder may also include a decoder.
  • AE auto-encoder
  • the encoder may be a recurrent neural network (RNN).
  • RNN recurrent neural network
  • the network structure is encoded by the encoder, and the feature vector of the network structure can be obtained in a portable manner.
  • the encoder may be obtained after training by the following method:
  • the decoder may be trained in the process of training the encoder.
  • the encoder and the decoder may be trained at the same time according to the difference between the second network structure and the third network structure.
  • the second feature vector is decoded by the decoder to obtain the third network structure, and data does not need to be annotated (for example, the feature vector output by the encoder does not need to be annotated), and according to the The difference between the second network structure and the third network structure can easily train the encoder and the decoder.
  • the encoder and the decoder can be easily trained.
  • training the encoder through a learning method can improve the accuracy of the feature vectors extracted by the encoder (of the first network structure and the second network structure), that is, it can make different
  • the feature vector of more accurately reflects the characteristics of different network structures (for example, the computing power of different network structures).
  • the neural network predictor can be used to predict the performance of the network structure.
  • the neural network predictor may be used to predict the performance of the network structure of the target neural network, and the target neural network may be used for image processing, speech processing, natural language processing, and the like.
  • the target neural network may be the convolutional neural network shown in FIG. 2, and the target neural network may be used for image classification, image segmentation, image detection, image super-segmentation, and the like.
  • the neural network predictor may be a graph convolutional network (GCN).
  • GCN graph convolutional network
  • the relationship between the first network structure and the second network structure can be better utilized in the training process (for example, the first network structure The similarity between the structure of the second network and the second network structure), so that the training effect of the neural network predictor can be improved.
  • the training of the neural network predictor according to the first network structure, the second network structure, the similarity and the label may include:
  • multiple network structures may be obtained in the above S510, and the performance of the multiple network structures can be obtained according to the feature vectors of the multiple network structures and the similarity;
  • the difference between the performance of the network structure and the corresponding label of the network structure is used as the loss value to train the neural network predictor.
  • the first feature vector, the second feature vector, and the relationship between the first network structure and the second network structure are used (for example, the two networks Similarity between structures) predicting the performance of the first network structure can make the performance information (predicted) of the first network structure more accurate.
  • Training the neural network predictor with the label can improve the training effect of the neural network predictor, that is, improve the prediction accuracy of the trained neural network predictor.
  • the relationship between the first network structure and the second network structure is used (for example, the similarity between the first network structure and the second network structure )
  • the neural network predictor which can improve the training effect of the neural network predictor when a small amount of labeled data (for example, at least one labeled network structure) is used, that is, to improve the trained neural network predictor.
  • the prediction accuracy of the network predictor is used.
  • Fig. 6 is a schematic flow chart of the method for training a neural network predictor of this application.
  • the method 600 for training a neural network predictor in FIG. 6 may include step 610, step 620, and step 630.
  • the method 600 may be executed by the execution device 120 in FIG. 1, the chip shown in FIG. 3, and the execution device 210 in FIG. 4 and other devices.
  • network features of multiple network structures in the network structure set may be extracted, and the multiple network structures may include a small number of network structures with performance labels and a large number of network structures without performance labels.
  • the network feature of the network structure may be the feature vector of the network structure in the method 500 in FIG. 5, and for details, reference may be made to the description in the method 500, which will not be repeated here.
  • a self-encoder may be used to extract network features of N network structures in the network structure set X, and the self-encoder may include an encoder E and a decoder D.
  • both the encoder E and the decoder D can be implemented by a neural network (NN).
  • the encoder E and the decoder D may be a recurrent neural network (RNN).
  • the encoder E may be used to extract and encode N network structures in the network structure set X to obtain network features of the N network structures.
  • the decoder D can be used to decode the network features of the N network structures to obtain N candidate network structures. These N candidate network structures correspond to the N network structures; Structure, training the encoder E and the decoder D.
  • the decoder D can also be used to decode only the network features of the N l network structures with performance labels to obtain N l candidate network structures, and these N l candidate network structures correspond to the N l N network structures with performance labels; the encoder E and the decoder D can be trained according to N l candidate network structures.
  • the following loss function can be constructed to train the autoencoder (that is, the encoder E and the decoder D):
  • W e is the parameter of the encoder
  • W d is the parameter of the decoder
  • N l, N u are positive integers.
  • the output of the encoder can be As the network feature of the extracted network structure, for ease of description, the output of the encoder can be subsequently Abbreviated as E(x i ).
  • the encoder may be the encoder in the method 500 in FIG. 5, and the decoder may be the decoder in the method 500 in FIG.
  • S620 Construct a network relationship graph according to the network characteristics.
  • a network relationship graph may be constructed according to the network characteristics obtained in S610.
  • N network features corresponding to N network structures can be obtained in S610, and then based on these N network features, an NxN network relationship graph can be constructed.
  • the NxN network relationship graph may include the similarity between each of the N network features and other N-1 network features, and the similarity between each network feature and its own network structure.
  • the prepaid range of the similarity can be [0,1], where 0 can mean that the two are completely different (or the two have the lowest similarity), and 1 can mean that the two are completely the same (or the two are the same).
  • each network feature is the same as its own network structure, the similarity between each network feature and its own network structure can be 1.
  • s(x i , x j ) can be used to represent the similarity between the network structure x i and the network structure x j , s(x i , x j ) can be calculated by the following distance formula:
  • d( ⁇ ) is an arbitrary distance measurement function
  • is a hyperparameter
  • exp( ⁇ ) is an exponential function
  • the meaning of the above formula is that for a given network structure x i , the network characteristic E(x i ) and the network structure x j ’s network characteristic E(x j ), the network characteristic E(x i ) and the network characteristic E(x j The greater the distance between ), the lower the similarity; conversely, the closer the distance between the network feature E(x i ) and the network feature E(x j ), the higher the similarity.
  • the similarity may be the similarity in the method 500 in FIG. 5, and the specific description may refer to the embodiment in the method 500, which will not be repeated here.
  • each of these N network structures can calculate a similarity according to the above method, and we get an NxN relationship graph.
  • Each element in the relationship graph represents the similarity between two network structures (the network characteristics).
  • the graph convolutional neural network can be regarded as a neural network predictor.
  • the graph convolutional neural network may be the neural network predictor in the method 500 in FIG. 5, and the specific description may refer to the embodiment in the method 500, which will not be repeated here.
  • the network characteristics E(x i ) of the N network structures output by the encoder and the network relationship diagram of NxN (for example, the network relationship diagram of NxN can be an NxN matrix) input into the graph convolutional neural network , You can get the performance of N network structures.
  • the performance label is The predicted performance is Hope the predicted performance is Can be as close as possible to the true value Then the following loss function can be constructed to train the graph convolutional neural network (ie, neural network predictor):
  • the following loss function can also be constructed to train the encoder, decoder, and graph convolutional neural network at the same time:
  • W e is the parameter of the encoder
  • W d is the parameter of the decoder
  • W p is the parameter of the graph convolutional neural network
  • L rc is the loss function of the autoencoder
  • L rg is the parameter of the graph convolutional neural network
  • is a hyperparameter
  • is used to adjust the weight of the two loss functions.
  • the data set NAS-Bench-101 can contain about 423,000 different network structures, and these network structures (that is, about 423,000 different network structures included in the data set NAS-Bench-101) are trained on the data set CIFAR-10 The true accuracy obtained afterwards.
  • r is the correlation coefficient, with a value range of [-1,1], used to evaluate the correlation between the predicted value (predicted performance) and the true value. The larger the value of r, the more accurate the predicted value (prediction performance).
  • Table 1 shows the KTau, MSE, and r values of the neural network predictor obtained by different methods when using different numbers of labeled data in the NAS-Bench-101 data set, as shown in the following table 1 shows:
  • the labeled samples in the foregoing Table 1 are the network structure with performance tags mentioned in the foregoing embodiment.
  • Method 2 1000 labeled samples and all unlabeled samples can be used as training data to train the neural network predictor, and the obtained predictor is used to search the neural network.
  • the search results are shown in Table 2 below:
  • the accuracy in Table 2 above may refer to the accuracy of the searched network structure, for example, the accuracy may be Top-1 Accuracy (%), and the ranking position may be the ranking position of the network structure in the current search space, for example, The ranking position can be Ranking (%).
  • the performance of the method in this application for the unknown search space can be verified.
  • 1000 data can be randomly selected from the data set NAS-Bench-101 (that is, the network structure in the data set NAS-Bench-101), and the 1000 data can be trained on the CIFAR-100 data set to obtain this
  • the true accuracy of 1000 data is used to train the predictor.
  • the predictor obtained after training can be used to predict the performance of the network model on the CIFAR-100 data set.
  • the prediction results are shown in Table 3 below:
  • FIG. 8 is a schematic flowchart of the image processing method of this application.
  • the method 800 in FIG. 8 includes step 810 and step 820.
  • the method 800 may be executed by the execution device 120 in FIG. 1, the chip shown in FIG. 3, and the execution device 210 in FIG. 4 and other devices.
  • S810 Acquire an image to be processed.
  • S820 Perform image processing on the to-be-processed image using a neural network.
  • the neural network may be determined according to a neural network predictor, and the neural network predictor is obtained after training by the method 500 in FIG. 5 or the method 600 in FIG. 6 described above.
  • the neural network may be a neural network that meets performance requirements and is searched in a preset search space through a neural network structure search method.
  • the neural network predictor obtained after training in the method 500 in FIG. 5 or the method 600 in FIG. 6 can be used to predict the performance of the network structure.
  • FIG. 9 is a schematic diagram of the hardware structure of an apparatus for training a neural network predictor provided by an embodiment of the present application.
  • the apparatus 3000 for training a neural network predictor shown in FIG. 9 includes a memory 3001, a processor 3002, a communication interface 3003, and a bus 3004.
  • the memory 3001, the processor 3002, and the communication interface 3003 implement communication connections between each other through the bus 3004.
  • the memory 3001 may be a read only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM).
  • the memory 3001 may store a program. When the program stored in the memory 3001 is executed by the processor 3002, the processor 3002 is configured to execute each step of the method for training a neural network predictor in the embodiment of the present application.
  • the processor 3002 may adopt a general central processing unit (CPU), a microprocessor, an application specific integrated circuit (ASIC), a graphics processing unit (GPU), or one or more
  • the integrated circuit is used to execute related programs to implement the method for training a neural network predictor in the method embodiment of the present application.
  • the processor 3002 may also be an integrated circuit chip with signal processing capability. For example, it may be the chip shown in FIG. 2.
  • each step of the method for training a neural network predictor of the present application can be completed by an integrated logic circuit of hardware in the processor 3002 or instructions in the form of software.
  • the above-mentioned processor 3002 may also be a general-purpose processor, a digital signal processing (digital signal processing, DSP), an application specific integrated circuit (ASIC), an off-the-shelf programmable gate array (field programmable gate array, FPGA) or other programmable logic devices, Discrete gates or transistor logic devices, discrete hardware components.
  • DSP digital signal processing
  • ASIC application specific integrated circuit
  • FPGA field programmable gate array
  • the general-purpose processor may be a microprocessor or the processor may also be any conventional processor or the like.
  • the steps of the method disclosed in the embodiments of the present application may be directly embodied as being executed and completed by a hardware decoding processor, or executed and completed by a combination of hardware and software modules in the decoding processor.
  • the software module can be located in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, or electrically erasable programmable memory, registers.
  • the storage medium is located in the memory 3001, and the processor 3002 reads the information in the memory 3001, and combines its hardware to complete the functions required by the units included in the device for training neural network predictors, or execute the training neural network in the method embodiment of the application. Method of network predictor.
  • the communication interface 3003 uses a transceiver device such as but not limited to a transceiver to implement communication between the device 3000 and other devices or communication networks. For example, the information of the neural network predictor to be constructed and the training data required in the process of training the neural network predictor can be obtained through the communication interface 3003.
  • a transceiver device such as but not limited to a transceiver to implement communication between the device 3000 and other devices or communication networks.
  • the information of the neural network predictor to be constructed and the training data required in the process of training the neural network predictor can be obtained through the communication interface 3003.
  • the bus 3004 may include a path for transferring information between various components of the device 3000 (for example, the memory 3001, the processor 3002, and the communication interface 3003).
  • FIG. 10 is a schematic diagram of the hardware structure of an image processing apparatus according to an embodiment of the present application.
  • the image processing apparatus 4000 shown in FIG. 10 includes a memory 4001, a processor 4002, a communication interface 4003, and a bus 4004.
  • the memory 4001, the processor 4002, and the communication interface 4003 implement communication connections between each other through the bus 4004.
  • the memory 4001 may be ROM, static storage device and RAM.
  • the memory 4001 may store a program. When the program stored in the memory 4001 is executed by the processor 4002, the processor 4002 and the communication interface 4003 are used to execute each step of the image processing method of the embodiment of the present application.
  • the processor 4002 may adopt a general-purpose CPU, a microprocessor, an ASIC, a GPU, or one or more integrated circuits to execute related programs to realize the functions required by the units in the image processing apparatus of the embodiment of the present application. Or execute the image processing method in the method embodiment of this application.
  • the processor 4002 may also be an integrated circuit chip with signal processing capability. For example, it may be the chip shown in FIG. 2. In the implementation process, each step of the image processing method of the embodiment of the present application can be completed by an integrated logic circuit of hardware in the processor 4002 or instructions in the form of software.
  • the aforementioned processor 4002 may also be a general-purpose processor, DSP, ASIC, FPGA or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component.
  • the methods, steps, and logical block diagrams disclosed in the embodiments of the present application can be implemented or executed.
  • the general-purpose processor may be a microprocessor or the processor may also be any conventional processor or the like.
  • the steps of the method disclosed in the embodiments of the present application may be directly embodied as being executed and completed by a hardware decoding processor, or executed and completed by a combination of hardware and software modules in the decoding processor.
  • the software module can be located in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, or electrically erasable programmable memory, registers.
  • the storage medium is located in the memory 4001, and the processor 4002 reads the information in the memory 4001, and combines its hardware to complete the functions required by the units included in the image processing apparatus of the embodiment of the present application, or perform the image processing of the method embodiment of the present application method.
  • the communication interface 4003 uses a transceiver device such as but not limited to a transceiver to implement communication between the device 4000 and other devices or a communication network.
  • a transceiver device such as but not limited to a transceiver to implement communication between the device 4000 and other devices or a communication network.
  • the image to be processed can be acquired through the communication interface 4003.
  • the bus 4004 may include a path for transferring information between various components of the device 4000 (for example, the memory 4001, the processor 4002, and the communication interface 4003).
  • the processor in the embodiment of the present application may be a central processing unit (central processing unit, CPU), and the processor may also be other general-purpose processors, digital signal processors (digital signal processors, DSP), and application-specific integrated circuits. (application specific integrated circuit, ASIC), ready-made programmable gate array (field programmable gate array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc.
  • the general-purpose processor may be a microprocessor or the processor may also be any conventional processor or the like.
  • the memory in the embodiments of the present application may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory.
  • the non-volatile memory can be read-only memory (ROM), programmable read-only memory (programmable ROM, PROM), erasable programmable read-only memory (erasable PROM, EPROM), and electrically available Erase programmable read-only memory (electrically EPROM, EEPROM) or flash memory.
  • the volatile memory may be random access memory (RAM), which is used as an external cache.
  • RAM random access memory
  • static random access memory static random access memory
  • DRAM dynamic random access memory
  • DRAM synchronous dynamic random access memory
  • Access memory synchronous DRAM, SDRAM
  • double data rate synchronous dynamic random access memory double data rate SDRAM, DDR SDRAM
  • enhanced synchronous dynamic random access memory enhanced SDRAM, ESDRAM
  • synchronous connection dynamic random access memory Take memory (synchlink DRAM, SLDRAM) and direct memory bus random access memory (direct rambus RAM, DR RAM).
  • the above-mentioned embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination.
  • the above-mentioned embodiments may be implemented in the form of a computer program product in whole or in part.
  • the computer program product includes one or more computer instructions or computer programs.
  • the processes or functions described in the embodiments of the present application are generated in whole or in part.
  • the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
  • the computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.
  • the computer instructions may be transmitted from a website, computer, server, or data center. Transmission to another website, computer, server or data center via wired (such as infrared, wireless, microwave, etc.).
  • the computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more sets of available media.
  • the usable medium may be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium.
  • the semiconductor medium may be a solid state drive.
  • At least one refers to one or more, and “multiple” refers to two or more.
  • the following at least one item (a)” or similar expressions refers to any combination of these items, including any combination of a single item (a) or a plurality of items (a).
  • at least one item (a) of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple .
  • the size of the sequence number of the above-mentioned processes does not mean the order of execution, and the execution order of each process should be determined by its function and internal logic, and should not correspond to the embodiments of the present application.
  • the implementation process constitutes any limitation.
  • the disclosed system, device, and method can be implemented in other ways.
  • the device embodiments described above are merely illustrative, for example, the division of the units is only a logical function division, and there may be other divisions in actual implementation, for example, multiple units or components may be combined or It can be integrated into another system, or some features can be ignored or not implemented.
  • the displayed or discussed mutual coupling or direct coupling or communication connection may be indirect coupling or communication connection through some interfaces, devices or units, and may be in electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the solutions of the embodiments.
  • the functional units in the various embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit.
  • the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium.
  • the technical solution of the present application essentially or the part that contributes to the existing technology or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including Several instructions are used to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in the various embodiments of the present application.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic disk or optical disk and other media that can store program code .

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Image Analysis (AREA)

Abstract

La présente invention concerne un procédé et un appareil de formation de prédicteur de réseau neuronal, et un procédé et un appareil de traitement d'image qui se rapportent au domaine de l'intelligence artificielle. Le procédé de formation de prédicteur de réseau neuronal consiste : à acquérir une première structure de réseau d'un premier réseau neuronal et une seconde structure de réseau d'un second réseau neuronal (S510), la première structure de réseau étant une structure de réseau ayant une étiquette et l'étiquette étant utilisée pour indiquer les performances de la première structure de réseau ; à acquérir la similarité entre la première structure de réseau et la seconde structure de réseau (S520) ; et à former un prédicteur de réseau neuronal en fonction de la première structure de réseau, de la seconde structure de réseau, de la similarité et de l'étiquette (S530), le prédicteur de réseau neuronal étant utilisé pour prédire les performances des structures de réseau. Dans le procédé, la relation entre des structures de réseau est utilisée pour aider à la formation d'un prédicteur de réseau neuronal et, dans la mesure où une petite quantité de données étiquetées est utilisée, la précision de prédiction du prédicteur de réseau neuronal formé peut être améliorée.
PCT/CN2021/088254 2020-05-09 2021-04-20 Procédé et appareil de formation de prédicteur de réseau neuronal, et procédé et appareil de traitement d'image WO2021227787A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010387976.4 2020-05-09
CN202010387976.4A CN111695673B (zh) 2020-05-09 2020-05-09 训练神经网络预测器的方法、图像处理方法及装置

Publications (1)

Publication Number Publication Date
WO2021227787A1 true WO2021227787A1 (fr) 2021-11-18

Family

ID=72477513

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/088254 WO2021227787A1 (fr) 2020-05-09 2021-04-20 Procédé et appareil de formation de prédicteur de réseau neuronal, et procédé et appareil de traitement d'image

Country Status (2)

Country Link
CN (1) CN111695673B (fr)
WO (1) WO2021227787A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114462524A (zh) * 2022-01-19 2022-05-10 北京工业大学 一种面向数据中心批处理作业的聚类方法

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111695673B (zh) * 2020-05-09 2024-05-24 华为技术有限公司 训练神经网络预测器的方法、图像处理方法及装置
CN112381147B (zh) * 2020-11-16 2024-04-26 虎博网络技术(上海)有限公司 动态图片相似度模型建立、相似度计算方法和装置
CN112957013B (zh) * 2021-02-05 2022-11-11 江西国科美信医疗科技有限公司 一种动态生命体征信号采集系统、监测装置及设备
CN113807183B (zh) * 2021-08-17 2024-06-14 华为技术有限公司 模型训练方法及相关设备

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110210558A (zh) * 2019-05-31 2019-09-06 北京市商汤科技开发有限公司 评估神经网络性能的方法及装置
CN110232434A (zh) * 2019-04-28 2019-09-13 吉林大学 一种基于属性图优化的神经网络架构评估方法
US20190370648A1 (en) * 2018-05-29 2019-12-05 Google Llc Neural architecture search for dense image prediction tasks
CN110555514A (zh) * 2019-08-20 2019-12-10 北京迈格威科技有限公司 神经网络模型搜索方法、图像识别方法和装置
CN111695673A (zh) * 2020-05-09 2020-09-22 华为技术有限公司 训练神经网络预测器的方法、图像处理方法及装置

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9792492B2 (en) * 2015-07-07 2017-10-17 Xerox Corporation Extracting gradient features from neural networks
KR101725684B1 (ko) * 2015-12-01 2017-04-27 연세대학교 산학협력단 에너지 인센티브 정책을 위한 건축물 에너지 성능 평가 장치 및 방법
KR101951595B1 (ko) * 2018-05-18 2019-02-22 한양대학교 산학협력단 모듈형 순환 신경망 구조 기반 차량 경로 예측 시스템 및 방법
CN110309856A (zh) * 2019-05-30 2019-10-08 华为技术有限公司 图像分类方法、神经网络的训练方法及装置
CN110532871B (zh) * 2019-07-24 2022-05-10 华为技术有限公司 图像处理的方法和装置
CN111046907B (zh) * 2019-11-02 2023-10-27 国网天津市电力公司 一种基于多头注意力机制的半监督卷积网络嵌入方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190370648A1 (en) * 2018-05-29 2019-12-05 Google Llc Neural architecture search for dense image prediction tasks
CN110232434A (zh) * 2019-04-28 2019-09-13 吉林大学 一种基于属性图优化的神经网络架构评估方法
CN110210558A (zh) * 2019-05-31 2019-09-06 北京市商汤科技开发有限公司 评估神经网络性能的方法及装置
CN110555514A (zh) * 2019-08-20 2019-12-10 北京迈格威科技有限公司 神经网络模型搜索方法、图像识别方法和装置
CN111695673A (zh) * 2020-05-09 2020-09-22 华为技术有限公司 训练神经网络预测器的方法、图像处理方法及装置

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114462524A (zh) * 2022-01-19 2022-05-10 北京工业大学 一种面向数据中心批处理作业的聚类方法

Also Published As

Publication number Publication date
CN111695673B (zh) 2024-05-24
CN111695673A (zh) 2020-09-22

Similar Documents

Publication Publication Date Title
US20220092351A1 (en) Image classification method, neural network training method, and apparatus
WO2021120719A1 (fr) Procédé de mise à jour de modèle de réseau neuronal, procédé et dispositif de traitement d'image
WO2021043168A1 (fr) Procédé d'entraînement de réseau de ré-identification de personnes et procédé et appareil de ré-identification de personnes
WO2020221200A1 (fr) Procédé de construction de réseau neuronal, procédé et dispositifs de traitement d'image
WO2020216227A9 (fr) Procédé et appareil de classification d'image et procédé et appareil de traitement de données
WO2021042828A1 (fr) Procédé et appareil de compression de modèle de réseau neuronal, ainsi que support de stockage et puce
WO2021043273A1 (fr) Procédé et appareil d'amélioration d'image
WO2021227787A1 (fr) Procédé et appareil de formation de prédicteur de réseau neuronal, et procédé et appareil de traitement d'image
WO2021043112A1 (fr) Procédé et appareil de classification d'images
WO2021057056A1 (fr) Procédé de recherche d'architecture neuronale, procédé et dispositif de traitement d'image, et support de stockage
WO2021018245A1 (fr) Procédé et appareil de classification d'images
WO2021022521A1 (fr) Procédé de traitement de données et procédé et dispositif d'apprentissage de modèle de réseau neuronal
WO2022001805A1 (fr) Procédé et dispositif de distillation de réseau neuronal
WO2021013095A1 (fr) Procédé et appareil de classification d'image, et procédé et appareil pour apprentissage de modèle de classification d'image
US20230206069A1 (en) Deep Learning Training Method for Computing Device and Apparatus
WO2021155792A1 (fr) Appareil de traitement, procédé et support de stockage
CN110222717B (zh) 图像处理方法和装置
WO2022001372A1 (fr) Procédé et appareil d'entraînement de réseau neuronal, et procédé et appareil de traitement d'image
WO2021008206A1 (fr) Procédé de recherche d'architecture neuronale, et procédé et dispositif de traitement d'images
WO2021018251A1 (fr) Procédé et dispositif de classification d'image
CN110222718B (zh) 图像处理的方法及装置
WO2021073311A1 (fr) Procédé et appareil de reconnaissance d'image, support d'enregistrement lisible par ordinateur et puce
WO2021103731A1 (fr) Procédé de segmentation sémantique et procédé et appareil d'apprentissage de modèle
WO2021136058A1 (fr) Procédé et dispositif de traitement vidéo
CN113011562A (zh) 一种模型训练方法及装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21804719

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21804719

Country of ref document: EP

Kind code of ref document: A1