Different Learning Rate for Weight and Bias in one layer

144 views
Skip to first unread message

pmpa...@gmail.com

unread,
Sep 7, 2018, 9:00:33 AM9/7/18
to Keras-users
In Caffe, we can set different learning rate for weight and bias in one layer.


layer { name: "conv1" type: "Convolution" bottom: "conv2" top: "conv1" param { lr_mult: 1 } param { lr_mult: 2 } convolution_param { num_output: 64 kernel_size: 3 stride: 1 pad: 1 weight_filler { type: "gaussian" std: 0.01 } bias_filler { type: "constant" value: 0.1 } } }


the learning rate of weight and bias is 

learning_rate(defined in solver prototxt) lr_mult.


In Keras, is it possible to set different learning rate for weight and bias in one layer?

Ted Yu

unread,
Sep 7, 2018, 1:25:58 PM9/7/18
to pmpa...@gmail.com, keras...@googlegroups.com
A brief search led to an old issue:

--
You received this message because you are subscribed to the Google Groups "Keras-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to keras-users...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/keras-users/f04f8aee-d80d-42a3-a0cc-33288b661a43%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Reply all
Reply to author
Forward
0 new messages