-
公开(公告)号：US20210312264A1
公开(公告)日：2021-10-07
申请号：US17354430
申请日：2021-06-22
Inventor： Fukui YANG , Shengzhao WEN , Junyu HAN
Abstract: A method, and an apparatus for model distillation are provided. The method may include: obtaining a batch of teacher features corresponding to a teacher model and a batch of student features corresponding to a student model; determining a set of teacher similarities corresponding to the batch of teacher features and a set of student similarities corresponding to the batch of student features; determining weights of loss values of features of images based on difference values corresponding to the images; and weighting a loss value of a feature of each image in a batch of images, training the student model by using a weighting result. The present disclosure may use the difference values between the feature similarities of the student model and the feature similarities of the teacher model to determine the weights of the loss values.
-
公开(公告)号：US20190065251A1
公开(公告)日：2019-02-28
申请号：US16116624
申请日：2018-08-29
Inventor： Shengzhao WEN , Hanqing ZHOU , Chuanxiu LIU , Jiajun ZHANG
Abstract: The present disclosure provides a method and apparatus for processing a heterogeneous cluster-oriented task. The method comprises: receiving a task request and a basic execution environment; scheduling a heterogeneous device according to the task request; compiling the basic execution environment into an execution environment corresponding to the scheduled heterogeneous device, and deploying on the scheduled heterogeneous device; triggering the scheduled heterogeneous device to execute the task request. It is only necessary for the user to provide the basic execution environment with respect to the task, and unnecessary for the user to respectively write a version of execution environment for each type of hardware platform, thereby implementing quick development of the heterogeneous device code and reducing the development and maintenance costs.
-
Search Results
2
records
(took 0.021 seconds)
