Intel® Distribution of OpenVINO™ Toolkit
Community assistance about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all aspects of computer vision-related on Intel® platforms.
6392 Discussions

caffe scale dosn't surpoort two bottom as input????

idata
Employee
692 Views

caffe scale dosn't surpoort two bottom as input????

0 Kudos
2 Replies
idata
Employee
389 Views

@xiaoshenhua The scale layer is meant to scale some input values by dividing by the standard deviation of those values. I've only seen it used with one bottom(input) in the networks that I've used.

0 Kudos
idata
Employee
389 Views

thank you but

 

in the caffe/src/caffe/proto/caffe.proto:

 

support two inputs

 

so how can i realize (N*C*W*H) * (N*C)?

 

message ScaleParameter {

 

// The first axis of bottom[0] (the first input Blob) along which to apply

 

// bottom[1] (the second input Blob). May be negative to index from the end

 

// (e.g., -1 for the last axis).

 

//

 

// For example, if bottom[0] is 4D with shape 100x3x40x60, the output

 

// top[0] will have the same shape, and bottom[1] may have any of the

 

// following shapes (for the given value of axis):

 

// (axis == 0 == -4) 100; 100x3; 100x3x40; 100x3x40x60

 

// (axis == 1 == -3) 3; 3x40; 3x40x60

 

// (axis == 2 == -2) 40; 40x60

 

// (axis == 3 == -1) 60

 

// Furthermore, bottom[1] may have the empty shape (regardless of the value of

 

// "axis") -- a scalar multiplier.

 

optional int32 axis = 1 [default = 1];

 

// (num_axes is ignored unless just one bottom is given and the scale is

 

// a learned parameter of the layer. Otherwise, num_axes is determined by the

 

// number of axes by the second bottom.)

 

// The number of axes of the input (bottom[0]) covered by the scale

 

// parameter, or -1 to cover all axes of bottom[0] starting from axis.

 

// Set num_axes := 0, to multiply with a zero-axis Blob: a scalar.

 

optional int32 num_axes = 2 [default = 1];

 

}
0 Kudos
Reply