-
1.发明公开
公开(公告)号：EP3876166A3
公开(公告)日：2022-01-12
申请号：EP21180897.7
申请日：2021-06-22
发明人： XI, Teng , ZHANG, Gang , WEN, Shengzhao
摘要： Embodiments of the present disclosure disclose a method and apparatus for determining a network model pruning strategy, a device and a storage medium, relate to the field of artificial intelligence technology such as computer vision and deep learning, and may be used in image processing scenarios. The method includes: generating a BN threshold search space using configuration information of the BN threshold search space for a network model for a target hardware; generating a pruning strategy code generator using the BN threshold search space; randomly generating a BN threshold code using the pruning strategy code generator; decoding the BN threshold code to obtain a candidate BN threshold; and determining a target pruning strategy of the network model for the target hardware, based on a pruning accuracy loss of the network model corresponding to the candidate BN threshold. An optimal BN threshold is obtained through automatic search, and the BN threshold is used to replace channel importance to prune the network model, thereby greatly reducing the accuracy loss of the network model after pruning.
-
公开(公告)号：EP3879457A3
公开(公告)日：2022-01-12
申请号：EP21180631.0
申请日：2021-06-21
发明人： YANG, Fukui , WEN, Shengzhao , HAN, Junyu
IPC分类号： G06K9/62
摘要： The present disclosure provides a method, and an apparatus for model distillation, relates to the technical field of artificial intelligence, and in particular, relates to technical fields of deep learning and computer vision. A specific implementation includes: 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. The distillation process of the present disclosure may improve the detection capabilities of the models, reduce the delay of the execution devices, and reduce the occupation and consumption of computing resources such as memories.
-
公开(公告)号：EP3879457A2
公开(公告)日：2021-09-15
申请号：EP21180631.0
申请日：2021-06-21
发明人： YANG, Fukui , WEN, Shengzhao , HAN, Junyu
IPC分类号： G06K9/62
摘要： The present disclosure provides a method, and an apparatus for model distillation, relates to the technical field of artificial intelligence, and in particular, relates to technical fields of deep learning and computer vision. A specific implementation includes: 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. The distillation process of the present disclosure may improve the detection capabilities of the models, reduce the delay of the execution devices, and reduce the occupation and consumption of computing resources such as memories.
-
4.发明公开
公开(公告)号：EP3876166A2
公开(公告)日：2021-09-08
申请号：EP21180897.7
申请日：2021-06-22
发明人： XI, Teng , ZHANG, Gang , WEN, Shengzhao
摘要： Embodiments of the present disclosure disclose a method and apparatus for determining a network model pruning strategy, a device and a storage medium, relate to the field of artificial intelligence technology such as computer vision and deep learning, and may be used in image processing scenarios. The method includes: generating a BN threshold search space using configuration information of the BN threshold search space for a network model for a target hardware; generating a pruning strategy code generator using the BN threshold search space; randomly generating a BN threshold code using the pruning strategy code generator; decoding the BN threshold code to obtain a candidate BN threshold; and determining a target pruning strategy of the network model for the target hardware, based on a pruning accuracy loss of the network model corresponding to the candidate BN threshold. An optimal BN threshold is obtained through automatic search, and the BN threshold is used to replace channel importance to prune the network model, thereby greatly reducing the accuracy loss of the network model after pruning.
-
-
-
搜索结果
4 条记录 (耗时 0.01 秒)
