deepin-ocr/3rdparty/ncnn/benchmark/mobilenet_yolo.param
wangzhengyang 718c41634f feat: 切换后端至PaddleOCR-NCNN,切换工程为CMake
1.项目后端整体迁移至PaddleOCR-NCNN算法,已通过基本的兼容性测试
2.工程改为使用CMake组织,后续为了更好地兼容第三方库,不再提供QMake工程
3.重整权利声明文件,重整代码工程,确保最小化侵权风险

Log: 切换后端至PaddleOCR-NCNN,切换工程为CMake
Change-Id: I4d5d2c5d37505a4a24b389b1a4c5d12f17bfa38c
2022-05-10 10:22:11 +08:00

42 lines
5.7 KiB
Plaintext

7767517
39 41
Input data 0 1 data -23330=4,3,416,416,3 0=416 1=416 2=3
Convolution conv0 1 1 data conv0_conv0/relu -23330=4,3,208,208,32 0=32 1=3 3=2 4=1 5=1 6=864 9=1
ConvolutionDepthWise conv1/dw 1 1 conv0_conv0/relu conv1/dw_conv1/dw/relu -23330=4,3,208,208,32 0=32 1=3 4=1 5=1 6=288 7=32 9=1
Convolution conv1 1 1 conv1/dw_conv1/dw/relu conv1_conv1/relu -23330=4,3,208,208,64 0=64 1=1 5=1 6=2048 9=1
ConvolutionDepthWise conv2/dw 1 1 conv1_conv1/relu conv2/dw_conv2/dw/relu -23330=4,3,104,104,64 0=64 1=3 3=2 4=1 5=1 6=576 7=64 9=1
Convolution conv2 1 1 conv2/dw_conv2/dw/relu conv2_conv2/relu -23330=4,3,104,104,128 0=128 1=1 5=1 6=8192 9=1
ConvolutionDepthWise conv3/dw 1 1 conv2_conv2/relu conv3/dw_conv3/dw/relu -23330=4,3,104,104,128 0=128 1=3 4=1 5=1 6=1152 7=128 9=1
Convolution conv3 1 1 conv3/dw_conv3/dw/relu conv3_conv3/relu -23330=4,3,104,104,128 0=128 1=1 5=1 6=16384 9=1
ConvolutionDepthWise conv4/dw 1 1 conv3_conv3/relu conv4/dw_conv4/dw/relu -23330=4,3,52,52,128 0=128 1=3 3=2 4=1 5=1 6=1152 7=128 9=1
Convolution conv4 1 1 conv4/dw_conv4/dw/relu conv4_conv4/relu -23330=4,3,52,52,256 0=256 1=1 5=1 6=32768 9=1
ConvolutionDepthWise conv5/dw 1 1 conv4_conv4/relu conv5/dw_conv5/dw/relu -23330=4,3,52,52,256 0=256 1=3 4=1 5=1 6=2304 7=256 9=1
Convolution conv5 1 1 conv5/dw_conv5/dw/relu conv5_conv5/relu -23330=4,3,52,52,256 0=256 1=1 5=1 6=65536 9=1
ConvolutionDepthWise conv6/dw 1 1 conv5_conv5/relu conv6/dw_conv6/dw/relu -23330=4,3,26,26,256 0=256 1=3 3=2 4=1 5=1 6=2304 7=256 9=1
Convolution conv6 1 1 conv6/dw_conv6/dw/relu conv6_conv6/relu -23330=4,3,26,26,512 0=512 1=1 5=1 6=131072 9=1
ConvolutionDepthWise conv7/dw 1 1 conv6_conv6/relu conv7/dw_conv7/dw/relu -23330=4,3,26,26,512 0=512 1=3 4=1 5=1 6=4608 7=512 9=1
Convolution conv7 1 1 conv7/dw_conv7/dw/relu conv7_conv7/relu -23330=4,3,26,26,512 0=512 1=1 5=1 6=262144 9=1
ConvolutionDepthWise conv8/dw 1 1 conv7_conv7/relu conv8/dw_conv8/dw/relu -23330=4,3,26,26,512 0=512 1=3 4=1 5=1 6=4608 7=512 9=1
Convolution conv8 1 1 conv8/dw_conv8/dw/relu conv8_conv8/relu -23330=4,3,26,26,512 0=512 1=1 5=1 6=262144 9=1
ConvolutionDepthWise conv9/dw 1 1 conv8_conv8/relu conv9/dw_conv9/dw/relu -23330=4,3,26,26,512 0=512 1=3 4=1 5=1 6=4608 7=512 9=1
Convolution conv9 1 1 conv9/dw_conv9/dw/relu conv9_conv9/relu -23330=4,3,26,26,512 0=512 1=1 5=1 6=262144 9=1
ConvolutionDepthWise conv10/dw 1 1 conv9_conv9/relu conv10/dw_conv10/dw/relu -23330=4,3,26,26,512 0=512 1=3 4=1 5=1 6=4608 7=512 9=1
Convolution conv10 1 1 conv10/dw_conv10/dw/relu conv10_conv10/relu -23330=4,3,26,26,512 0=512 1=1 5=1 6=262144 9=1
ConvolutionDepthWise conv11/dw 1 1 conv10_conv10/relu conv11/dw_conv11/dw/relu -23330=4,3,26,26,512 0=512 1=3 4=1 5=1 6=4608 7=512 9=1
Convolution conv11 1 1 conv11/dw_conv11/dw/relu conv11_conv11/relu -23330=4,3,26,26,512 0=512 1=1 5=1 6=262144 9=1
Split splitncnn_0 1 2 conv11_conv11/relu conv11_conv11/relu_splitncnn_0 conv11_conv11/relu_splitncnn_1 -23330=8,3,26,26,512,3,26,26,512
ConvolutionDepthWise conv12/dw 1 1 conv11_conv11/relu_splitncnn_1 conv12/dw_conv12/dw/relu -23330=4,3,13,13,512 0=512 1=3 3=2 4=1 5=1 6=4608 7=512 9=1
Convolution conv12 1 1 conv12/dw_conv12/dw/relu conv12_conv12/relu -23330=4,3,13,13,1024 0=1024 1=1 5=1 6=524288 9=1
ConvolutionDepthWise conv13/dw 1 1 conv12_conv12/relu conv13/dw_conv13/dw/relu -23330=4,3,13,13,1024 0=1024 1=3 4=1 5=1 6=9216 7=1024 9=1
Convolution conv13 1 1 conv13/dw_conv13/dw/relu conv13_conv13/relu -23330=4,3,13,13,1024 0=1024 1=1 5=1 6=1048576 9=1
ConvolutionDepthWise conv16/dw 1 1 conv13_conv13/relu conv16/dw_conv16/dw/relu -23330=4,3,13,13,1024 0=1024 1=3 4=1 5=1 6=9216 7=1024 9=1
Convolution conv17 1 1 conv16/dw_conv16/dw/relu conv17_conv17/relu -23330=4,3,13,13,1024 0=1024 1=1 5=1 6=1048576 9=1
Split splitncnn_1 1 2 conv17_conv17/relu conv17_conv17/relu_splitncnn_0 conv17_conv17/relu_splitncnn_1 -23330=8,3,13,13,1024,3,13,13,1024
DeconvolutionDepthWise upsample 1 1 conv17_conv17/relu_splitncnn_1 upsample -23330=4,3,26,26,512 0=512 1=4 3=2 4=1 6=16384 7=512
Eltwise conv_18/sum 2 1 conv11_conv11/relu_splitncnn_0 upsample conv_18/sum -23330=4,3,26,26,512 0=1
ConvolutionDepthWise conv19/dw 1 1 conv_18/sum conv19/dw_conv19/dw/relu -23330=4,3,26,26,512 0=512 1=3 4=1 5=1 6=4608 7=512 9=1
Convolution conv20 1 1 conv19/dw_conv19/dw/relu conv20_conv20/relu -23330=4,3,26,26,1024 0=1024 1=1 5=1 6=524288 9=1
Convolution conv22_indoor 1 1 conv17_conv17/relu_splitncnn_0 conv22 -23330=4,3,13,13,125 0=125 1=1 5=1 6=128000
Convolution conv23_indoor 1 1 conv20_conv20/relu conv23 -23330=4,3,26,26,125 0=125 1=1 5=1 6=128000
YoloDetectionOutput detection_out 2 1 conv22 conv23 output -23330=4,3,13,13,125 2=4.000000e-01 -23304=10,1.080000e+00,1.190000e+00,3.420000e+00,4.410000e+00,6.630000e+00,1.138000e+01,9.420000e+00,5.110000e+00,1.662000e+01,1.052000e+01