'importTensorFlowLayers' does not support 'tf.keras.​layers.Nor​malization​'

조회 수: 3 (최근 30일)
Sangmin Lee
Sangmin Lee 2021년 8월 13일
댓글: Rickard Andersson 2024년 3월 7일
Hi,
I am using the latest relase version of Matlab, and I think 'importTensorFlowLayer's does not support 'tf.keras.layers.Normalization'.
'tf.keras.layers.Normalization' layer remembers mean and variance in its layer to nomralize input data. Do you have a plan to support this layer, or how can I replace this layer to work in Matlab?
Thanks.

답변 (2개)

Bhuvanyu Dogra
Bhuvanyu Dogra 2021년 8월 17일
It is my understanding that you are looking for alternatives of the layer 'tf.keras.layers.Normalization' in MATLAB R2021a.
'layerNormalizationLayer' or 'batchNormalizationLayer' can serve as alternatives as they provide normalization across all channels or individual channels. Moreover, 'batchNormalizationLayer' has the ability to use mean or variance for prediction.
For more information, refer to the following documentation:
  1. layerNormalizationLayer: https://www.mathworks.com/help/deeplearning/ref/nnet.cnn.layer.layernormalizationlayer.html
  2. batchNormalizationLayer: https://www.mathworks.com/help/deeplearning/ref/nnet.cnn.layer.batchnormalizationlayer.html
For more information on the TensorFlow-Keras layers supported in MATLAB using 'importTensorFlowLayers', refer to the following documentation:

bob scott
bob scott 2023년 4월 20일
Hi, I would like to to take keras model file that contains a 'tf.keras.layers.Normalization' layer for normalizing inputs to the model and import it and have it 'work' without having to adjust anything on the model side.
Are there any plans to support 'tf.keras.layers.Normalization' when importing keras models?

카테고리

Help CenterFile Exchange에서 Deep Learning Toolbox에 대해 자세히 알아보기

제품


릴리스

R2021a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by