Distillation learning method and device, computer equipment and storage medium

A learning method and computer program technology, applied in neural learning methods, neural architecture, biological neural network models, etc., can solve problems such as the performance difference between the student network and the teacher network, and the poor distillation learning effect.

CN111898735APending Publication Date: 2020-11-06上海眼控科技股份有限公司
0 Cites 5 Cited by

Patent Information

Authority / Receiving Office
CN · China
Current Assignee / Owner
Publication Date
2020-11-06

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention relates to a distillation learning method and device, computer equipment and a storage medium. The method comprises the following steps: respectively inputting a target image into a teacher network and a student network to obtain a teacher feature map output by the teacher network and a student feature map output by the student network; performing channel matching on the teacher feature map and the student feature map, and obtaining a target loss value between the teacher feature map and the student feature map according to a matching result; and adjusting parameters in the student network according to the target loss value to obtain a target student network. By adopting the method, the distillation learning effect of the student network on the teacher network can be improved, and the performance difference between the student network and the teacher network is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The present application relates to the technical field of machine learning, in particular to a distillation learning method, device, computer equipment and storage medium. Background technique

[0002] With the development of machine learning technology, distillation learning technology has emerged. Distillation learning uses migration learning, using the output of a pre-trained complex network model (teacher model) as a supervisory signal to train another simple network model. (Student network, Student model) to obtain a student network with a simplified structure and low computational complexity, while having the knowledge of the teacher network.

[0003] In the traditional technology, most of them are carried out from the three aspects of pixel-level distillation, relational distillation and discriminator-based global distillation, but the distillation learning effect is poor, and the performance of the obtained student network and the teacher netwo...

Examples

Embodiment Construction

[0050] In order to make the purpose, technical solution and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application, and are not intended to limit the present application.

[0051] In one embodiment, such as figure 1 As shown, a distillation learning method is provided, and this embodiment is illustrated by applying the method to a terminal. It can be understood that this method can also be applied to a server, and can also be applied to a system including a terminal and a server, and through The interactive realization of terminal and server. Wherein, the terminal can be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers and portable wearable devices, and the server can be realized by an indep...