Data processing model construction method and device, equipment and storage medium

Abstract

The invention provides a data processing model construction method and device, equipment and a storage medium, and the method comprises the steps: determining an operator which is not supported by a second operation main body from a first data processing model operated on a first operation main body, and taking the operator as a to-be-optimized operator; an optimization operator corresponding to the to-be-optimized operator is constructed by using the operator supported by the second operation main body, so that the optimization operator can realize the same function as the corresponding to-be-optimized operator and can be supported by the second operation main body; replacing a to-be-optimized operator in the first data processing model with the optimization operator to obtain a second data processing model, so that the second data processing model and the first data processing model have the same function and framework and can be deployed to a second operation main body; the data processing models with the same function and framework can be deployed to the operation main bodies with different computing resources, the practicability of the data processing models is improved, and the development pressure on the data processing models is relieved.

Description

technical field [0001] The present application relates to the technical field of model construction, and in particular, to a data processing model construction method, apparatus, device, and storage medium. Background technique [0002] The data processing service usually deploys a pre-built data processing model (for example, a neural network model) on the running subject, so that the running subject uses the data processing model to realize the data processing service. Different running subjects have different hardware resources and processing performance, therefore, the same data processing model may not be deployed to different running subjects at the same time. [0003] For example, in a speech recognition scenario, due to the abundant computing resources in the cloud, speech recognition services are usually implemented by deploying a speech recognition system in the cloud. However, due to factors such as user privacy, bandwidth delay, network instability, etc., deploy...

AI Technical Summary

Problems solved by technology

However, the computing resources on the device side are significantly different from those on the cloud. The operators of the speech recognition system running on the cloud may not be supported on the device side. As a result, the speech recognition system on the cloud cannot be deployed to the device side application. Develop a Speech Recognition System

[0004] Therefore, data processing models with the same functions and frameworks cannot be deployed to operating entities with different computing resources, resulting in low practicability of the data processing models and unable to alleviate the pressure on the development of data processing models

Images