提交 abeb3aae 编写于 作者: D dangqingqing

Minor changes

上级 f56084b4
...@@ -79,10 +79,10 @@ void PadGrad<DEVICE_TYPE_CPU>(real* inGrad, ...@@ -79,10 +79,10 @@ void PadGrad<DEVICE_TYPE_CPU>(real* inGrad,
* *
* Argument in this Function: * Argument in this Function:
* \param pad_ A struct object contains the padding size in each dimension. * \param pad_ A struct object contains the padding size in each dimension.
* It has six integers. The channelStart and channelEnd indicates * It has six integers. The channelStart and channelEnd indicate
* how many zeros to add before and after the input in channel * how many zeros to add before and after the input in channel
* dimension. And the heightStart and heightEnd indicates padding * dimension. And the heightStart and heightEnd indicate padding
* in height dimension. The widthStart and widthEnd indicates the * in height dimension. The widthStart and widthEnd indicate the
* padding in width dimension. * padding in width dimension.
* \param inputs A 4D tensor, only one input. * \param inputs A 4D tensor, only one input.
* \param outputs A 4D tensor, the output value after padding. * \param outputs A 4D tensor, the output value after padding.
......
...@@ -19,8 +19,9 @@ limitations under the License. */ ...@@ -19,8 +19,9 @@ limitations under the License. */
namespace paddle { namespace paddle {
/** /**
* @brief response normalization across feature maps * \brief This layer pads zeros to inputs according to the specify dimension.
* namely normalize in number of size_ channels * The input and output is a 4D tensor. Padding zeros from the 2nd to
* the 4th dimenstion according padc_, padh_ and padw_.
*/ */
class PadLayer : public Layer { class PadLayer : public Layer {
public: public:
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册