Caffe Deconvolution 的定义与计算过程

Caffe Deconvolution 的定义与计算过程

1. Deconvolution

1.1 Deconvolution Layer

https://caffe.berkeleyvision.org/tutorial/layers/deconvolution.html

• Input: $(N,C_{in},H_{in},W_{in})$
• Filter: $(C_{in},C_{out}/group,K_h,K_w)$
• Output: $(N,C_{out},H_{out},W_{out})$

1.2 Deconvolution Layer

https://caffe.berkeleyvision.org/doxygen/classcaffe_1_1DeconvolutionLayer.html

1.3 BilinearFiller

https://caffe.berkeleyvision.org/doxygen/classcaffe_1_1BilinearFiller.html

2. Deconvolution

https://support.huawei.com/enterprise/en/doc/EDOC1100234056/f1efbdcc/deconvolution

CANN Documentation
https://support.huawei.com/enterprise/en/ascend-computing/cann-pid-251168373
https://support.huawei.com/enterprise/zh/ascend-computing/cann-pid-251168373

CANN 5.0.4 Caffe, TensorFlow, and ONNX Operator Support 01
https://support.huawei.com/enterprise/en/doc/EDOC1100234056?idPath=23710424%7C251366513%7C22892968%7C251168373

2.1 General Restrictions

input shape: $(N_y,C_y,H_y,W_y)$
filter shape (k=kernel): $(N_k,C_k/group,H_k,W_k)$
output shape: $(N_x,C_x,H_x,W_x)$

• $N_x=N_y,C_x=C_k,N_k=C_y$
• $W_x=(W_y–1)\ast strid{e}_w–2\ast pa{d}_w+dilatio{n}_h\ast (W_k–1)+1$
• $H_x=(H_y–1)\ast strid{e}_h–2\ast pa{d}_h+dilatio{n}_w\ast (H_k–1)+1$

References

https://yongqiang.blog.csdn.net/