Abstract: A human-computer interaction method, a human-computer interaction apparatus, an electronic device and a storage medium are provided, which relate to a field of artificial intelligence technology, and in particular to fields of deep learning, natural language processing and large model technologies. The human-computer interaction method includes: determining, in response to a human-computer interaction request, a first target plug-in related to a first dialogue text contained in the human-computer interaction request from a plurality of plug-ins registered in a large language model based on the first dialogue text contained in the human-computer interaction request; obtaining a second dialogue text based on the first dialogue text and a description text of the first target plug-in; and inputting the second dialogue text into the large language model to obtain a response text.

Abstract: A method and apparatus for generating an operator are provided. The method includes: constructing a group of basic application programming interfaces for providing one of the following basic functions: an access function, a storage function, and a computing function; constructing a kernel application programming interface for invoking the basic application programming interfaces to implement an operator logic; and generating a target kernel operator based on the group of basic application programming interfaces and the kernel application programming interface.

Abstract: A method for controlling a video memory for model training, an electronic device and a storage medium are provided, relating to the field of artificial intelligence technology, and in particular to the fields of neural network, large model, training optimization and other technologies. The method includes: reconstructing a video memory space for one or more backward calculations during model training according to grouping information of parameter gradient information required for the one or more backward calculations; performing the one or more backward calculations to obtain one or more backward calculation results; storing the one or more backward calculation results into the video memory space reconstructed for the one or more backward calculations; and releasing the video memory space reconstructed for the one or more backward calculations.

Abstract: The disclosure provides a neural network training method and apparatus, an electronic device, a medium and a program product, and relates to the field of artificial intelligence, in particular to the fields of deep learning and distributed learning. The method includes: acquiring a neural network for deep learning; constructing a deep reinforcement learning model for the neural network; and determining, through the deep reinforcement learning model, a processing unit selection for the plurality of the network layers based on a duration for training each of the network layers by each type of the plurality of types of the processing units, and a cost of each type of the plurality of types of the processing units.

Abstract: The present disclosure provides a method and apparatus for recognizing a text, a device and a storage medium, and relates to the field of deep learning technology. A specific implementation comprises: receiving a target image; performing a text detection on the target image using a pre-trained lightweight text detection network, to obtain a text detection box; and recognizing a text in the text detection box using a pre-trained lightweight text recognition network, to obtain a text recognition result.

Abstract: A method for training a deep learning model may include: acquiring model description information and configuration information of a deep learning model; segmenting the model description information into at least two sections based on segmentation point variable in the configuration information, and loading the model description information to a corresponding resource to run; inputting a batch of training samples into a resource corresponding to a first section of model description information, then starting training and using obtained context information as an input of a resource corresponding to a subsequent section of model description information; and so on until an operation result of a resource corresponding to a final section of model description information is obtained; if a training completion condition is met, outputting a trained deep learning model; and otherwise, keeping on acquiring a subsequent batch of training samples and performing the above training steps until the condition is met.

Abstract: The disclosure provides an access method, an access apparatus, an electronic device and a computer storage medium, and relates to a field of computer technologies, in particular to a field of artificial intelligence technologies such as chip and deep learning. The method includes: determining a computational graph for calling an access device based on operator representations in a target model; optimizing the computational graph based on information of the access device; and performing relevant running operations of the target model on the access device based on the computational graph and an interface for the access device to access to a model framework of the target model, the interface being determined based on kit data of the access device.

Abstract: Provided is a quantization parameter storage method, a model inference method, an electronic device and a storage medium, relating to the fields of large model technology, artificial intelligence technology and model quantization technology. The quantization parameter storage method includes: obtaining, by a calculation unit of a processor, a statistical value of a first quantization parameter of a model statistically based on benchmark data; searching for, by the calculation unit, a target value of the first quantization parameter and a target value of a second quantization parameter of the model in a search space based on the statistical value of the first quantization parameter; and storing, by the calculation unit, the target value of the first quantization parameter and the target value of the second quantization parameter into a memory.

Abstract: A method for detecting a fault, an electronic device and a storage medium are provided, relating to the field of computer technology, and in particular to the fields of deep learning, large model training, fault detection and other technologies. The method includes: determining a plurality of computing devices, where the plurality of computing devices are used to perform model training based on a pipeline parallelism strategy; determining a parameter and a scheduling strategy used by the pipeline parallelism strategy; determining idle time of each computing device among the plurality of computing devices in a model training process based on the parameter and the scheduling strategy; and performing fault detection on each computing device during the idle time of each computing device in the model training process.

Abstract: An arrow tooth saw having an arrow-shaped cutter tooth includes an arrow tooth unit, the arrow tooth unit includes an arrow tooth, an arrow tooth angle, an arrow tooth tip, arrow tooth edges, an arrow tooth head and an arrow tooth tail. Due to a sharp arrow tooth angle and long arrow tooth edges, a shear resistance of the arrow tooth is reduced to only one tenth or one in tens of a shear resistance of a straight saw with the same width, and the actual thickness of an arrow tooth edge is more than one order of magnitude thinner than a blunting value of a corresponding straight tooth required by sawing regulations. The present invention created a saw tooth that can continue sawing without the need for grinding repair while maintaining better sawing quality than that of a straight tooth.

Abstract: A session recommendation method, a device and an electronic device are provided, related to the field of graph neural network technology. The session recommendation method includes: acquiring a session control sequence, and acquiring a first embedding vector matrix based on an embedding vector of each of items in the session control sequence; generating a position information sequence based on an arrangement sequence of the items in the session control sequence, and acquiring a second embedding vector matrix based on an embedding vector of each piece of position information in the position information sequence; determining a target embedding vector matrix based on the first embedding vector matrix and the second embedding vector matrix; and determining a recommended item, based on the target embedding vector matrix and through a Session-based Recommendation Graph Neural Network.

Abstract: A method for distributed operation based on a neural network model and a related apparatus are provided, relating to the field of computer technology and in particular to the fields of artificial intelligence, deep learning, machine learning, distributed training and other technologies. The method includes: parsing code of the neural network model to construct an operator topology graph corresponding to the neural network model; generating a distributed operation strategy of the neural network model based on the operator topology graph and a preset resource constraint; and modifying the code of the neural network model based on the distributed operation strategy to obtain target code; where the target code is used to operate the neural network model based on the distributed operation strategy on a computing device corresponding to the resource constraint.

Abstract: A large model-based text generation method, electronic device, and storage medium in the field of artificial intelligence technologies such as large models and natural language processing are provided. The specific implementation includes: obtaining a matching prefix, where the matching prefix includes at least one consecutive token; obtaining a draft token sequence based on the matching prefix according to a pre-configured draft token sequence length, where the draft token sequence includes at least one token; performing validity verification on the draft token sequence using a pre-trained large model based on a speculative decoding algorithm; and in response to passing the verification, using the draft token sequence as generated text.

Abstract: A scheduling method for a deep learning framework, a scheduling apparatus, an electronic device, a storage medium, and a program product is provided, and can be used in the field of artificial intelligence, especially in the fields of machine learning, deep learning, etc. The method includes: receiving a processing request for processing a plurality of tasks by using a dedicated processing unit, the processing request including scheduling requirements for the plurality of tasks, and each of the plurality of tasks being associated with execution of multi-batch data processing; and scheduling, based on the scheduling requirements for the plurality of tasks in batches of data, the dedicated processing unit to process the plurality of tasks.

Abstract: Embodiments of the present disclosure provide a method and apparatus of training a model, an electronic device, a storage medium and a development system, which relate to a field of deep learning. The method may include calling a training preparation component to set at least a loss function and an optimization function for training the model, in response to determining that a training preparation instruction is received. The method further includes calling a training component to set a first data reading component, in response to determining that a training instruction is received. The first data reading component is configured to load a training data set for training the model. In addition, the method may further include training the model based on the training data set from the first data reading component, by using the loss function and the optimization function through the training component.

Abstract: The present application discloses a distributed training method and system, a device and a storage medium, and relates to technical fields of deep learning and cloud computing. The method includes: sending, by a task information server, a first training request and information of an available first computing server to at least a first data server; sending, by the first data server, a first batch of training data to the first computing server, according to the first training request; performing, by the first computing server, model training according to the first batch of training data, sending model parameters to the first data server so as to be stored after the training is completed, and sending identification information of the first batch of training data to the task information server so as to be recorded; wherein the model parameters are not stored at any one of the computing servers.

Abstract: A method and apparatus of constructing a network model for deep learning, a device, and a storage medium, which relate to artificial intelligence, and in particular to a field of deep learning. The method of constructing the network model for deep learning includes: determining an execution mode for executing codes, based on a mode parameter; executing the codes by using a first component, which is executable in a first execution mode, through a syntax element in the codes, in response to determining that the execution mode is the first execution mode; and executing the codes by using a second component, which is executable in a second execution mode, through the syntax element, in response to determining that the execution mode is the second execution mode; wherein the first component and the second component have the same component interface, and the syntax element corresponds to the component interface.

Abstract: The present disclosure provides a method for training a feature extraction model, a method for classifying an image and related apparatuses, and relates to the field of artificial intelligence technology such as deep learning and image recognition. The scheme comprises: extracting an image feature of each sample image in a sample image set using a basic feature extraction module of an initial feature extraction model, to obtain an initial feature vector set; performing normalization processing on each initial feature vector in the initial feature vector set using a normalization processing module of the initial feature extraction model, to obtain each normalized feature vector; and guiding training for the initial feature extraction model through a preset high discriminative loss function, to obtain a target feature extraction model as a training result.

Abstract: The present disclosure provides an image processing method and apparatus, and relates to the field of image processing, and in particular to the field of image annotation. An implementation is: obtaining an image to be processed including a target region to be annotated; in response to a first click on the target region, performing a first operation to expand a predicted region for the target region based on a click position of the first click; in response to a second click in a position where the predicted region exceeds the target region, performing a second operation to reduce the predicted region based on a click position of the second click; and in response to determining that a difference between the predicted region and the target region meets a preset condition, obtaining an outline of the predicted region to annotate the target region.

Abstract: Provided is a lightweight model training method, an image processing method, a device and a medium. The lightweight model training method includes: acquiring first and second augmentation probabilities and a target weight adopted in an e-th iteration; performing data augmentation on a data set based on the first and second augmentation probabilities respectively, to obtain first and second data sets; obtaining a first output value of a student model and a second output value of a teacher model based on the first data set; obtaining a third output value and a fourth output value based on the second data set; determining a distillation loss function, a truth-value loss function and a target loss function; training the student model based on the target loss function; and determining a first augmentation probability or target weight to be adopted in an (e+1)-th iteration in a case of e is less than E.