Deep learning environment building method and device and server

Abstract

The embodiment of the invention provides a deep learning environment building method and device and a server. The method is applied to a management server. The method comprises the steps of receivinga deep learning environment building request sent by a terminal, wherein the deep learning environment building request carries recognition information of a target deep learning framework; determiningtarget resource information required for building a target deep learning environment; allocating a target computing server to the target deep learning environment according to the target resource information required by the target deep learning environment; and sending a deep learning environment building request to a target computing server to enable the target computing server to create a container corresponding to the target deep learning framework according to the recognition information, thereby forming a target deep learning environment. Through the scheme provided by the embodiment ofthe invention, the dependence isolation of different deep learning frameworks can be realized, and the resource utilization rate is effectively improved.

Description

technical field [0001] The present invention relates to the technical field of deep learning, in particular to a method, device and server for building a deep learning environment. Background technique [0002] The concept of deep learning originated from the study of artificial neural networks. A multi-layer perceptron with multiple hidden layers is a deep learning structure. Deep learning combines low-level features to form more abstract high-level representation attribute categories or features to discover distributed feature representations of data. [0003] There are many deep learning frameworks, for example, the deep learning framework can be caffe, tensorflow, pytorch, mxnet, etc. Different deep learning frameworks depend on different deep learning environments. Specifically, training deep learning models based on different deep learning frameworks requires different dependency packages, which can include several libraries, such as function libraries, and operating...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the embodiment of the present invention is to provide a deep learning environment construction method, device and server, which can realize the isolation of dependent packages corresponding to different deep learning frameworks, and solve the problem that the dependent packages corresponding to different deep learning frameworks conflict with each other , effectively improving the resource utilization of the cluster system

Images