how i can make value range for weight in keras?

Asked Jul 02 '19 at 02:53

Active Jul 02 '19 at 02:53

Viewed 225 times

I am a beginner learning in Deep learning framework with keras backend tensorflow and start learning with this sample code https://machinelearningmastery.com/handwritten-digit-recognition-using-convolutional-neural-networks-python-keras/.

I want to make boundaries for my weight in keras model. I want to make my weights value in the range from -1 to 1. How can i do that?

I read about weight decay, weight constraint, weight clipping, making condition boundaries in file keras optimize.py , but because my lack of experience i can't decided which should i use.

asked Jul 02 '19 at 02:53

astri

Why do you want to do that? – gmds Jul 02 '19 at 03:02
Thank you for your fast response. My reason is i am researching with custom sigmoid activation function and i want to modify my weight too – astri Jul 02 '19 at 03:13
try this link , it may help. https://stackoverflow.com/questions/54435998/keras-limiting-output-min-max-value – LOrD_ARaGOrN Jul 02 '19 at 08:18
@RishiBansal Thank you for your response, i will try it – astri Jul 03 '19 at 00:04

how i can make value range for weight in keras?

0 Answers0