Cannot Run on Myriad Device #1
Comments
|
NCS and NCS 2 support very few layers. plugin = IEPlugin(device="MYRIAD")
#plugin = IEPlugin(device="GPU")
#plugin = IEPlugin(device="CPU")
#plugin.add_cpu_extension("lib/libcpu_extension.so")
exec_net = plugin.load(network=net)and, Please refer to the implementation of the next repository and offload the "ArgMax" layer. Probably you will need to train with your own dataset based on the following .prototxt. Same issue. |
|
Hi Pinto, Thanks a lot for the quick response. |
It's exactly as you say.
I have not confirmed the operation of "transposed convolution", but I wish you good luck. A list of layers supported by OpenVINO is listed below. It might be meaningless information for you, I know that using Intel's CPU will show better performance than NCS2. |
|
Hi Pinto, Thanks again for prompt reply. I gotta say I got more info from you than Intel itself. And although it is working fine. The performance of the inference, however, is not good. In fact it seems that the network is not trained at all and generating random results. Is your model trained or just initialized and converted to IR? Thanks again. Umer |
Unfortunately, this repository is the only repository I failed to convert. |
|
I had the similar problem when trying to convert deeplab3 - mobilenet2. The only layer out of that model that's unsupported by MYRIAD at the moment is Argmax. When you convert the model to IR format, you can specify the resized logits of the model as output: That way you'll get a |
|
@ikrets |
|
I changed the model by myself and excluded Argmax, but NCS / NCS 2 hangs during reasoning. .xml<?xml version="1.0" ?>
<net batch="1" name="semantic-segmentation-adas-0001" version="4">
<layers>
<layer id="0" name="data" precision="FP16" type="Input">
<output>
<port id="0">
<dim>1</dim>
<dim>3</dim>
<dim>256</dim>
<dim>512</dim>
</port>
</output>
</layer>
<layer id="1" name="Mul_/Fused_Mul_/FusedScaleShift_" precision="FP16" type="ScaleShift">
<input>
<port id="0">
<dim>1</dim>
<dim>3</dim>
<dim>256</dim>
<dim>512</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>3</dim>
<dim>256</dim>
<dim>512</dim>
</port>
</output>
<blobs>
<weights offset="0" size="6"/>
<biases offset="6" size="6"/>
</blobs>
</layer>
<layer id="2" name="AvgPool2DBackward2" precision="FP16" type="Pooling">
<data exclude-pad="false" kernel="2,2" pads_begin="0,0" pads_end="0,0" pool-method="avg" rounding_type="ceil" strides="2,2"/>
<input>
<port id="0">
<dim>1</dim>
<dim>3</dim>
<dim>256</dim>
<dim>512</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>3</dim>
<dim>128</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="3" name="ConvNdBackward3" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="32" pads_begin="1,1" pads_end="1,1" strides="2,2"/>
<input>
<port id="0">
<dim>1</dim>
<dim>3</dim>
<dim>128</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>32</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</output>
<blobs>
<weights offset="12" size="1728"/>
<biases offset="1740" size="64"/>
</blobs>
</layer>
<layer id="4" name="ThresholdBackward5" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>32</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="5" name="ConvNdBackward6" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="32" pads_begin="1,1" pads_end="1,1" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>32</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</output>
<blobs>
<weights offset="1804" size="18432"/>
<biases offset="20236" size="64"/>
</blobs>
</layer>
<layer id="6" name="ThresholdBackward8" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>32</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="7" name="ConvNdBackward9" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="64" pads_begin="1,1" pads_end="1,1" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>64</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</output>
<blobs>
<weights offset="20300" size="36864"/>
<biases offset="57164" size="128"/>
</blobs>
</layer>
<layer id="8" name="ThresholdBackward11" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>64</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="9" name="MaxPool2DBackward12" precision="FP16" type="Pooling">
<data exclude-pad="true" kernel="3,3" pads_begin="0,0" pads_end="0,0" pool-method="max" rounding_type="ceil" strides="2,2"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>64</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="10" name="ConvNdBackward13" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="32" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
<blobs>
<weights offset="57292" size="4096"/>
<biases offset="61388" size="64"/>
</blobs>
</layer>
<layer id="11" name="ThresholdBackward15" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="12" name="ConvNdBackward16" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="32" pads_begin="1,1" pads_end="1,1" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
<blobs>
<weights offset="61452" size="18432"/>
<biases offset="79884" size="64"/>
</blobs>
</layer>
<layer id="13" name="ThresholdBackward18" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="14" name="ConvNdBackward19" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="128" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
<blobs>
<weights offset="79948" size="8192"/>
<biases offset="88140" size="256"/>
</blobs>
</layer>
<layer id="15" name="ConvNdBackward22" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="128" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
<blobs>
<weights offset="88396" size="16384"/>
<biases offset="104780" size="256"/>
</blobs>
</layer>
<layer id="16" name="AddBackward124" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="17" name="ThresholdBackward25" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="18" name="ConvNdBackward26" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="32" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
<blobs>
<weights offset="105036" size="8192"/>
<biases offset="113228" size="64"/>
</blobs>
</layer>
<layer id="19" name="ThresholdBackward28" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="20" name="ConvNdBackward29" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="32" pads_begin="1,1" pads_end="1,1" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
<blobs>
<weights offset="113292" size="18432"/>
<biases offset="131724" size="64"/>
</blobs>
</layer>
<layer id="21" name="ThresholdBackward31" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="22" name="ConvNdBackward32" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="128" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
<blobs>
<weights offset="131788" size="8192"/>
<biases offset="139980" size="256"/>
</blobs>
</layer>
<layer id="23" name="AddBackward135" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="24" name="ThresholdBackward36" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="25" name="ConvNdBackward37" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="32" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
<blobs>
<weights offset="140236" size="8192"/>
<biases offset="148428" size="64"/>
</blobs>
</layer>
<layer id="26" name="ThresholdBackward39" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="27" name="ConvNdBackward40" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="32" pads_begin="1,1" pads_end="1,1" strides="2,2"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>32</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
<blobs>
<weights offset="148492" size="18432"/>
<biases offset="166924" size="64"/>
</blobs>
</layer>
<layer id="28" name="ThresholdBackward42" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>32</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="29" name="ConvNdBackward43" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="128" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
<blobs>
<weights offset="166988" size="8192"/>
<biases offset="175180" size="256"/>
</blobs>
</layer>
<layer id="30" name="Pooling_" precision="FP16" type="Pooling">
<data kernel="1,1" pads_begin="0,0" pads_end="0,0" pool-method="max" strides="2,2"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="31" name="AddBackward146" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="32" name="ThresholdBackward47" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="33" name="ConvNdBackward48" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="64" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>64</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
<blobs>
<weights offset="175436" size="16384"/>
<biases offset="191820" size="128"/>
</blobs>
</layer>
<layer id="34" name="ThresholdBackward50" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>64</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="35" name="ConvNdBackward51" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="64" pads_begin="1,1" pads_end="1,1" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>64</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
<blobs>
<weights offset="191948" size="73728"/>
<biases offset="265676" size="128"/>
</blobs>
</layer>
<layer id="36" name="ThresholdBackward53" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>64</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="37" name="ConvNdBackward54" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="256" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>256</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
<blobs>
<weights offset="265804" size="32768"/>
<biases offset="298572" size="512"/>
</blobs>
</layer>
<layer id="38" name="ConvNdBackward57" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="256" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>256</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
<blobs>
<weights offset="299084" size="65536"/>
<biases offset="364620" size="512"/>
</blobs>
</layer>
<layer id="39" name="AddBackward159" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>16</dim>
<dim>32</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>256</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="40" name="ThresholdBackward60" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="41" name="AvgPool2DBackward61" precision="FP16" type="Pooling">
<data exclude-pad="false" kernel="2,2" pads_begin="0,0" pads_end="0,0" pool-method="avg" rounding_type="ceil" strides="2,2"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="42" name="ConvNdBackward62" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="64" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="365132" size="32768"/>
<biases offset="397900" size="128"/>
</blobs>
</layer>
<layer id="43" name="ThresholdBackward64" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="44" name="ConvNdBackward65" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="64" pads_begin="1,1" pads_end="1,1" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="398028" size="73728"/>
<biases offset="471756" size="128"/>
</blobs>
</layer>
<layer id="45" name="ThresholdBackward67" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="46" name="ConvNdBackward68" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="256" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="471884" size="32768"/>
<biases offset="504652" size="512"/>
</blobs>
</layer>
<layer id="47" name="AddBackward171" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="48" name="ThresholdBackward72" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="49" name="ConvNdBackward73" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="64" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="505164" size="32768"/>
<biases offset="537932" size="128"/>
</blobs>
</layer>
<layer id="50" name="ThresholdBackward75" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="51" name="ConvNdBackward76" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="64" pads_begin="1,1" pads_end="1,1" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="538060" size="73728"/>
<biases offset="611788" size="128"/>
</blobs>
</layer>
<layer id="52" name="ThresholdBackward78" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="53" name="ConvNdBackward79" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="256" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="611916" size="32768"/>
<biases offset="644684" size="512"/>
</blobs>
</layer>
<layer id="54" name="AddBackward182" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="55" name="ThresholdBackward83" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="56" name="ConvNdBackward84" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="64" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="645196" size="32768"/>
<biases offset="677964" size="128"/>
</blobs>
</layer>
<layer id="57" name="ThresholdBackward86" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="58" name="ConvNdBackward87" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="64" pads_begin="1,1" pads_end="1,1" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="678092" size="73728"/>
<biases offset="751820" size="128"/>
</blobs>
</layer>
<layer id="59" name="ThresholdBackward89" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="60" name="ConvNdBackward90" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="256" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="751948" size="32768"/>
<biases offset="784716" size="512"/>
</blobs>
</layer>
<layer id="61" name="AddBackward193" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="62" name="ThresholdBackward94" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="63" name="ConvNdBackward95" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="128" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="785228" size="65536"/>
<biases offset="850764" size="256"/>
</blobs>
</layer>
<layer id="64" name="ThresholdBackward97" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="65" name="ConvNdBackward98" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="128" pads_begin="1,1" pads_end="1,1" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="851020" size="294912"/>
<biases offset="1145932" size="256"/>
</blobs>
</layer>
<layer id="66" name="ThresholdBackward100" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="67" name="ConvNdBackward101" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="512" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="1146188" size="131072"/>
<biases offset="1277260" size="1024"/>
</blobs>
</layer>
<layer id="68" name="ConvNdBackward104" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="512" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="1278284" size="262144"/>
<biases offset="1540428" size="1024"/>
</blobs>
</layer>
<layer id="69" name="AddBackward1106" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="70" name="ThresholdBackward107" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="71" name="ConvNdBackward108" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="128" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="1541452" size="131072"/>
<biases offset="1672524" size="256"/>
</blobs>
</layer>
<layer id="72" name="ThresholdBackward110" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="73" name="ConvNdBackward111" precision="FP16" type="Convolution">
<data dilations="2,2" group="1" kernel="3,3" output="128" pads_begin="2,2" pads_end="2,2" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="1672780" size="294912"/>
<biases offset="1967692" size="256"/>
</blobs>
</layer>
<layer id="74" name="ThresholdBackward113" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="75" name="ConvNdBackward114" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="512" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="1967948" size="131072"/>
<biases offset="2099020" size="1024"/>
</blobs>
</layer>
<layer id="76" name="AddBackward1117" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="77" name="ThresholdBackward118" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="78" name="ConvNdBackward119" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="128" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="2100044" size="131072"/>
<biases offset="2231116" size="256"/>
</blobs>
</layer>
<layer id="79" name="ThresholdBackward121" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="80" name="ConvNdBackward122" precision="FP16" type="Convolution">
<data dilations="2,2" group="1" kernel="3,3" output="128" pads_begin="2,2" pads_end="2,2" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="2231372" size="294912"/>
<biases offset="2526284" size="256"/>
</blobs>
</layer>
<layer id="81" name="ThresholdBackward124" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="82" name="ConvNdBackward125" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="512" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="2526540" size="131072"/>
<biases offset="2657612" size="1024"/>
</blobs>
</layer>
<layer id="83" name="AddBackward1128" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="84" name="ThresholdBackward129" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="85" name="ConvNdBackward130" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="128" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="2658636" size="131072"/>
<biases offset="2789708" size="256"/>
</blobs>
</layer>
<layer id="86" name="ThresholdBackward132" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="87" name="ConvNdBackward133" precision="FP16" type="Convolution">
<data dilations="2,2" group="1" kernel="3,3" output="128" pads_begin="2,2" pads_end="2,2" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="2789964" size="294912"/>
<biases offset="3084876" size="256"/>
</blobs>
</layer>
<layer id="88" name="ThresholdBackward135" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="89" name="ConvNdBackward136" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="512" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="3085132" size="131072"/>
<biases offset="3216204" size="1024"/>
</blobs>
</layer>
<layer id="90" name="AddBackward1139" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="91" name="ThresholdBackward140" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="92" name="ConvNdBackward141" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="128" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="3217228" size="131072"/>
<biases offset="3348300" size="256"/>
</blobs>
</layer>
<layer id="93" name="ThresholdBackward143" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="94" name="ConvNdBackward144" precision="FP16" type="Convolution">
<data dilations="2,2" group="1" kernel="3,3" output="128" pads_begin="2,2" pads_end="2,2" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="3348556" size="294912"/>
<biases offset="3643468" size="256"/>
</blobs>
</layer>
<layer id="95" name="ThresholdBackward146" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="96" name="ConvNdBackward147" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="512" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="3643724" size="131072"/>
<biases offset="3774796" size="1024"/>
</blobs>
</layer>
<layer id="97" name="AddBackward1150" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="98" name="ThresholdBackward151" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="99" name="ConvNdBackward152" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="128" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="3775820" size="131072"/>
<biases offset="3906892" size="256"/>
</blobs>
</layer>
<layer id="100" name="ThresholdBackward154" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="101" name="ConvNdBackward155" precision="FP16" type="Convolution">
<data dilations="2,2" group="1" kernel="3,3" output="128" pads_begin="2,2" pads_end="2,2" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="3907148" size="294912"/>
<biases offset="4202060" size="256"/>
</blobs>
</layer>
<layer id="102" name="ThresholdBackward157" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="103" name="ConvNdBackward158" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="512" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="4202316" size="131072"/>
<biases offset="4333388" size="1024"/>
</blobs>
</layer>
<layer id="104" name="AddBackward1161" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="105" name="ThresholdBackward162" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="106" name="ConvNdBackward163" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="256" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="4334412" size="262144"/>
<biases offset="4596556" size="512"/>
</blobs>
</layer>
<layer id="107" name="ThresholdBackward165" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="108" name="ConvNdBackward166" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="256" pads_begin="1,1" pads_end="1,1" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="4597068" size="1179648"/>
<biases offset="5776716" size="512"/>
</blobs>
</layer>
<layer id="109" name="ThresholdBackward168" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="110" name="ConvNdBackward169" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="1024" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="5777228" size="524288"/>
<biases offset="6301516" size="2048"/>
</blobs>
</layer>
<layer id="111" name="ConvNdBackward172" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="1024" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>512</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="6303564" size="1048576"/>
<biases offset="7352140" size="2048"/>
</blobs>
</layer>
<layer id="112" name="AddBackward1174" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="113" name="ThresholdBackward175" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="114" name="ConvNdBackward176" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="256" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="7354188" size="524288"/>
<biases offset="7878476" size="512"/>
</blobs>
</layer>
<layer id="115" name="ThresholdBackward178" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="116" name="ConvNdBackward179" precision="FP16" type="Convolution">
<data dilations="4,4" group="1" kernel="3,3" output="256" pads_begin="4,4" pads_end="4,4" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="7878988" size="1179648"/>
<biases offset="9058636" size="512"/>
</blobs>
</layer>
<layer id="117" name="ThresholdBackward181" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="118" name="ConvNdBackward182" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="1024" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="9059148" size="524288"/>
<biases offset="9583436" size="2048"/>
</blobs>
</layer>
<layer id="119" name="AddBackward1185" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="120" name="ThresholdBackward186" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="121" name="ConvNdBackward187" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="256" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="9585484" size="524288"/>
<biases offset="10109772" size="512"/>
</blobs>
</layer>
<layer id="122" name="ThresholdBackward189" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="123" name="ConvNdBackward190" precision="FP16" type="Convolution">
<data dilations="4,4" group="1" kernel="3,3" output="256" pads_begin="4,4" pads_end="4,4" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="10110284" size="1179648"/>
<biases offset="11289932" size="512"/>
</blobs>
</layer>
<layer id="124" name="ThresholdBackward192" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="125" name="ConvNdBackward193" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="1024" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="11290444" size="524288"/>
<biases offset="11814732" size="2048"/>
</blobs>
</layer>
<layer id="126" name="AddBackward1196" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="127" name="ThresholdBackward197" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="128" name="AdaptiveAvgPool2dBackward199" precision="FP16" type="Pooling">
<data exclude-pad="false" kernel="32,64" pads_begin="0,0" pads_end="0,0" pool-method="avg" rounding_type="ceil" strides="32,64"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</output>
</layer>
<layer id="129" name="UpsamplingBilinear2dBackward200" precision="FP16" type="Interp">
<data align_corners="1" height="32" pad_beg="0" pad_end="0" shrink_factor="1" width="64" zoom_factor="1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>1</dim>
<dim>1</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="130" name="AddBackward1201" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="131" name="AdaptiveAvgPool2dBackward203" precision="FP16" type="Pooling">
<data exclude-pad="false" kernel="16,32" pads_begin="0,0" pads_end="0,0" pool-method="avg" rounding_type="ceil" strides="16,32"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>2</dim>
<dim>2</dim>
</port>
</output>
</layer>
<layer id="132" name="UpsamplingBilinear2dBackward204" precision="FP16" type="Interp">
<data align_corners="1" height="32" pad_beg="0" pad_end="0" shrink_factor="1" width="64" zoom_factor="1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>2</dim>
<dim>2</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="133" name="AddBackward1205" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="134" name="AdaptiveAvgPool2dBackward207" precision="FP16" type="Pooling">
<data exclude-pad="false" kernel="12,22" pads_begin="0,0" pads_end="0,0" pool-method="avg" rounding_type="ceil" strides="11,22"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>3</dim>
<dim>3</dim>
</port>
</output>
</layer>
<layer id="135" name="UpsamplingBilinear2dBackward208" precision="FP16" type="Interp">
<data align_corners="1" height="32" pad_beg="0" pad_end="0" shrink_factor="1" width="64" zoom_factor="1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>3</dim>
<dim>3</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="136" name="AddBackward1209" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="137" name="AdaptiveAvgPool2dBackward211" precision="FP16" type="Pooling">
<data exclude-pad="false" kernel="6,12" pads_begin="0,0" pads_end="0,0" pool-method="avg" rounding_type="ceil" strides="6,11"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>6</dim>
<dim>6</dim>
</port>
</output>
</layer>
<layer id="138" name="UpsamplingBilinear2dBackward212" precision="FP16" type="Interp">
<data align_corners="1" height="32" pad_beg="0" pad_end="0" shrink_factor="1" width="64" zoom_factor="1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>6</dim>
<dim>6</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="139" name="AddBackward1213" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="140" name="ConvNdBackward214" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="256" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>1024</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
<blobs>
<weights offset="11816780" size="524288"/>
<biases offset="12341068" size="512"/>
</blobs>
</layer>
<layer id="141" name="ThresholdBackward216" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</output>
</layer>
<layer id="142" name="UpsamplingBilinear2dBackward217" precision="FP16" type="Interp">
<data align_corners="1" height="0" pad_beg="0" pad_end="0" shrink_factor="1" width="0" zoom_factor="2"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>8</dim>
<dim>16</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>256</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="143" name="ConvNdBackward218" precision="FP16" type="Convolution">
<data dilations="2,2" group="1" kernel="3,3" output="128" pads_begin="2,2" pads_end="2,2" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
<blobs>
<weights offset="12341580" size="589824"/>
<biases offset="12931404" size="256"/>
</blobs>
</layer>
<layer id="144" name="ConvNdBackward221" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="128" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>256</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
<blobs>
<weights offset="12931660" size="65536"/>
<biases offset="12997196" size="256"/>
</blobs>
</layer>
<layer id="145" name="AddBackward1223" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="146" name="ThresholdBackward224" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</output>
</layer>
<layer id="147" name="UpsamplingBilinear2dBackward225" precision="FP16" type="Interp">
<data align_corners="1" height="0" pad_beg="0" pad_end="0" shrink_factor="1" width="0" zoom_factor="2"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>16</dim>
<dim>32</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="148" name="ConvNdBackward226" precision="FP16" type="Convolution">
<data dilations="2,2" group="1" kernel="3,3" output="128" pads_begin="2,2" pads_end="2,2" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
<blobs>
<weights offset="12997452" size="294912"/>
<biases offset="13292364" size="256"/>
</blobs>
</layer>
<layer id="149" name="ConvNdBackward229" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="32" pads_begin="1,1" pads_end="1,1" strides="2,2"/>
<input>
<port id="0">
<dim>1</dim>
<dim>3</dim>
<dim>256</dim>
<dim>512</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>32</dim>
<dim>128</dim>
<dim>256</dim>
</port>
</output>
<blobs>
<weights offset="13292620" size="1728"/>
<biases offset="13294348" size="64"/>
</blobs>
</layer>
<layer id="150" name="ThresholdBackward231" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>128</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>32</dim>
<dim>128</dim>
<dim>256</dim>
</port>
</output>
</layer>
<layer id="151" name="ConvNdBackward232" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="32" pads_begin="1,1" pads_end="1,1" strides="2,2"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>128</dim>
<dim>256</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>32</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</output>
<blobs>
<weights offset="13294412" size="18432"/>
<biases offset="13312844" size="64"/>
</blobs>
</layer>
<layer id="152" name="ThresholdBackward234" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>32</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="153" name="ConvNdBackward235" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="3,3" output="64" pads_begin="1,1" pads_end="1,1" strides="2,2"/>
<input>
<port id="0">
<dim>1</dim>
<dim>32</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>64</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
<blobs>
<weights offset="13312908" size="36864"/>
<biases offset="13349772" size="128"/>
</blobs>
</layer>
<layer id="154" name="ThresholdBackward237" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>64</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="155" name="ConvNdBackward238" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="128" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>64</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
<blobs>
<weights offset="13349900" size="16384"/>
<biases offset="13366284" size="256"/>
</blobs>
</layer>
<layer id="156" name="AddBackward1240" precision="FP16" type="Eltwise">
<data coeff="" operation="sum"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="2">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="157" name="ThresholdBackward241" precision="FP16" type="ReLU">
<data negative_slope="0.0"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</output>
</layer>
<layer id="158" name="UpsamplingBilinear2dBackward242" precision="FP16" type="Interp">
<data align_corners="1" height="0" pad_beg="0" pad_end="0" shrink_factor="1" width="0" zoom_factor="2"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>32</dim>
<dim>64</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>128</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="159" name="ConvNdBackward243" precision="FP16" type="Convolution">
<data dilations="1,1" group="1" kernel="1,1" output="20" pads_begin="0,0" pads_end="0,0" strides="1,1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>128</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>20</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</output>
<blobs>
<weights offset="13366540" size="5120"/>
<biases offset="13371660" size="40"/>
</blobs>
</layer>
<layer id="160" name="LogSoftmaxBackward244" precision="FP16" type="SoftMax">
<data axis="1"/>
<input>
<port id="0">
<dim>1</dim>
<dim>20</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>20</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</output>
</layer>
<layer id="161" name="UpsamplingBilinear2dBackward245" precision="FP16" type="Interp">
<data align_corners="1" height="0" pad_beg="0" pad_end="0" shrink_factor="1" width="0" zoom_factor="4"/>
<input>
<port id="0">
<dim>1</dim>
<dim>20</dim>
<dim>64</dim>
<dim>128</dim>
</port>
</input>
<output>
<port id="1">
<dim>1</dim>
<dim>20</dim>
<dim>256</dim>
<dim>512</dim>
</port>
</output>
</layer>
<layer id="162" name="Rectify/Mul_" precision="FP16" type="ScaleShift">
<input>
<port id="0">
<dim>1</dim>
<dim>20</dim>
<dim>256</dim>
<dim>512</dim>
</port>
</input>
<output>
<port id="3">
<dim>1</dim>
<dim>20</dim>
<dim>256</dim>
<dim>512</dim>
</port>
</output>
<blobs>
<weights offset="13371700" size="40"/>
<biases offset="13371740" size="40"/>
</blobs>
</layer>
</layers>
<edges>
<edge from-layer="0" from-port="0" to-layer="1" to-port="0"/>
<edge from-layer="1" from-port="3" to-layer="2" to-port="0"/>
<edge from-layer="2" from-port="1" to-layer="3" to-port="0"/>
<edge from-layer="3" from-port="3" to-layer="4" to-port="0"/>
<edge from-layer="4" from-port="1" to-layer="5" to-port="0"/>
<edge from-layer="5" from-port="3" to-layer="6" to-port="0"/>
<edge from-layer="6" from-port="1" to-layer="7" to-port="0"/>
<edge from-layer="7" from-port="3" to-layer="8" to-port="0"/>
<edge from-layer="8" from-port="1" to-layer="9" to-port="0"/>
<edge from-layer="9" from-port="1" to-layer="10" to-port="0"/>
<edge from-layer="10" from-port="3" to-layer="11" to-port="0"/>
<edge from-layer="11" from-port="1" to-layer="12" to-port="0"/>
<edge from-layer="12" from-port="3" to-layer="13" to-port="0"/>
<edge from-layer="13" from-port="1" to-layer="14" to-port="0"/>
<edge from-layer="9" from-port="1" to-layer="15" to-port="0"/>
<edge from-layer="14" from-port="3" to-layer="16" to-port="0"/>
<edge from-layer="15" from-port="3" to-layer="16" to-port="1"/>
<edge from-layer="16" from-port="2" to-layer="17" to-port="0"/>
<edge from-layer="17" from-port="1" to-layer="18" to-port="0"/>
<edge from-layer="18" from-port="3" to-layer="19" to-port="0"/>
<edge from-layer="19" from-port="1" to-layer="20" to-port="0"/>
<edge from-layer="20" from-port="3" to-layer="21" to-port="0"/>
<edge from-layer="21" from-port="1" to-layer="22" to-port="0"/>
<edge from-layer="22" from-port="3" to-layer="23" to-port="0"/>
<edge from-layer="17" from-port="1" to-layer="23" to-port="1"/>
<edge from-layer="23" from-port="2" to-layer="24" to-port="0"/>
<edge from-layer="24" from-port="1" to-layer="25" to-port="0"/>
<edge from-layer="25" from-port="3" to-layer="26" to-port="0"/>
<edge from-layer="26" from-port="1" to-layer="27" to-port="0"/>
<edge from-layer="27" from-port="3" to-layer="28" to-port="0"/>
<edge from-layer="28" from-port="1" to-layer="29" to-port="0"/>
<edge from-layer="24" from-port="1" to-layer="30" to-port="0"/>
<edge from-layer="29" from-port="3" to-layer="31" to-port="0"/>
<edge from-layer="30" from-port="1" to-layer="31" to-port="1"/>
<edge from-layer="31" from-port="2" to-layer="32" to-port="0"/>
<edge from-layer="32" from-port="1" to-layer="33" to-port="0"/>
<edge from-layer="33" from-port="3" to-layer="34" to-port="0"/>
<edge from-layer="34" from-port="1" to-layer="35" to-port="0"/>
<edge from-layer="35" from-port="3" to-layer="36" to-port="0"/>
<edge from-layer="36" from-port="1" to-layer="37" to-port="0"/>
<edge from-layer="32" from-port="1" to-layer="38" to-port="0"/>
<edge from-layer="37" from-port="3" to-layer="39" to-port="0"/>
<edge from-layer="38" from-port="3" to-layer="39" to-port="1"/>
<edge from-layer="39" from-port="2" to-layer="40" to-port="0"/>
<edge from-layer="40" from-port="1" to-layer="41" to-port="0"/>
<edge from-layer="41" from-port="1" to-layer="42" to-port="0"/>
<edge from-layer="42" from-port="3" to-layer="43" to-port="0"/>
<edge from-layer="43" from-port="1" to-layer="44" to-port="0"/>
<edge from-layer="44" from-port="3" to-layer="45" to-port="0"/>
<edge from-layer="45" from-port="1" to-layer="46" to-port="0"/>
<edge from-layer="46" from-port="3" to-layer="47" to-port="0"/>
<edge from-layer="41" from-port="1" to-layer="47" to-port="1"/>
<edge from-layer="47" from-port="2" to-layer="48" to-port="0"/>
<edge from-layer="48" from-port="1" to-layer="49" to-port="0"/>
<edge from-layer="49" from-port="3" to-layer="50" to-port="0"/>
<edge from-layer="50" from-port="1" to-layer="51" to-port="0"/>
<edge from-layer="51" from-port="3" to-layer="52" to-port="0"/>
<edge from-layer="52" from-port="1" to-layer="53" to-port="0"/>
<edge from-layer="53" from-port="3" to-layer="54" to-port="0"/>
<edge from-layer="48" from-port="1" to-layer="54" to-port="1"/>
<edge from-layer="54" from-port="2" to-layer="55" to-port="0"/>
<edge from-layer="55" from-port="1" to-layer="56" to-port="0"/>
<edge from-layer="56" from-port="3" to-layer="57" to-port="0"/>
<edge from-layer="57" from-port="1" to-layer="58" to-port="0"/>
<edge from-layer="58" from-port="3" to-layer="59" to-port="0"/>
<edge from-layer="59" from-port="1" to-layer="60" to-port="0"/>
<edge from-layer="60" from-port="3" to-layer="61" to-port="0"/>
<edge from-layer="55" from-port="1" to-layer="61" to-port="1"/>
<edge from-layer="61" from-port="2" to-layer="62" to-port="0"/>
<edge from-layer="62" from-port="1" to-layer="63" to-port="0"/>
<edge from-layer="63" from-port="3" to-layer="64" to-port="0"/>
<edge from-layer="64" from-port="1" to-layer="65" to-port="0"/>
<edge from-layer="65" from-port="3" to-layer="66" to-port="0"/>
<edge from-layer="66" from-port="1" to-layer="67" to-port="0"/>
<edge from-layer="62" from-port="1" to-layer="68" to-port="0"/>
<edge from-layer="67" from-port="3" to-layer="69" to-port="0"/>
<edge from-layer="68" from-port="3" to-layer="69" to-port="1"/>
<edge from-layer="69" from-port="2" to-layer="70" to-port="0"/>
<edge from-layer="70" from-port="1" to-layer="71" to-port="0"/>
<edge from-layer="71" from-port="3" to-layer="72" to-port="0"/>
<edge from-layer="72" from-port="1" to-layer="73" to-port="0"/>
<edge from-layer="73" from-port="3" to-layer="74" to-port="0"/>
<edge from-layer="74" from-port="1" to-layer="75" to-port="0"/>
<edge from-layer="75" from-port="3" to-layer="76" to-port="0"/>
<edge from-layer="70" from-port="1" to-layer="76" to-port="1"/>
<edge from-layer="76" from-port="2" to-layer="77" to-port="0"/>
<edge from-layer="77" from-port="1" to-layer="78" to-port="0"/>
<edge from-layer="78" from-port="3" to-layer="79" to-port="0"/>
<edge from-layer="79" from-port="1" to-layer="80" to-port="0"/>
<edge from-layer="80" from-port="3" to-layer="81" to-port="0"/>
<edge from-layer="81" from-port="1" to-layer="82" to-port="0"/>
<edge from-layer="82" from-port="3" to-layer="83" to-port="0"/>
<edge from-layer="77" from-port="1" to-layer="83" to-port="1"/>
<edge from-layer="83" from-port="2" to-layer="84" to-port="0"/>
<edge from-layer="84" from-port="1" to-layer="85" to-port="0"/>
<edge from-layer="85" from-port="3" to-layer="86" to-port="0"/>
<edge from-layer="86" from-port="1" to-layer="87" to-port="0"/>
<edge from-layer="87" from-port="3" to-layer="88" to-port="0"/>
<edge from-layer="88" from-port="1" to-layer="89" to-port="0"/>
<edge from-layer="89" from-port="3" to-layer="90" to-port="0"/>
<edge from-layer="84" from-port="1" to-layer="90" to-port="1"/>
<edge from-layer="90" from-port="2" to-layer="91" to-port="0"/>
<edge from-layer="91" from-port="1" to-layer="92" to-port="0"/>
<edge from-layer="92" from-port="3" to-layer="93" to-port="0"/>
<edge from-layer="93" from-port="1" to-layer="94" to-port="0"/>
<edge from-layer="94" from-port="3" to-layer="95" to-port="0"/>
<edge from-layer="95" from-port="1" to-layer="96" to-port="0"/>
<edge from-layer="96" from-port="3" to-layer="97" to-port="0"/>
<edge from-layer="91" from-port="1" to-layer="97" to-port="1"/>
<edge from-layer="97" from-port="2" to-layer="98" to-port="0"/>
<edge from-layer="98" from-port="1" to-layer="99" to-port="0"/>
<edge from-layer="99" from-port="3" to-layer="100" to-port="0"/>
<edge from-layer="100" from-port="1" to-layer="101" to-port="0"/>
<edge from-layer="101" from-port="3" to-layer="102" to-port="0"/>
<edge from-layer="102" from-port="1" to-layer="103" to-port="0"/>
<edge from-layer="103" from-port="3" to-layer="104" to-port="0"/>
<edge from-layer="98" from-port="1" to-layer="104" to-port="1"/>
<edge from-layer="104" from-port="2" to-layer="105" to-port="0"/>
<edge from-layer="105" from-port="1" to-layer="106" to-port="0"/>
<edge from-layer="106" from-port="3" to-layer="107" to-port="0"/>
<edge from-layer="107" from-port="1" to-layer="108" to-port="0"/>
<edge from-layer="108" from-port="3" to-layer="109" to-port="0"/>
<edge from-layer="109" from-port="1" to-layer="110" to-port="0"/>
<edge from-layer="105" from-port="1" to-layer="111" to-port="0"/>
<edge from-layer="110" from-port="3" to-layer="112" to-port="0"/>
<edge from-layer="111" from-port="3" to-layer="112" to-port="1"/>
<edge from-layer="112" from-port="2" to-layer="113" to-port="0"/>
<edge from-layer="113" from-port="1" to-layer="114" to-port="0"/>
<edge from-layer="114" from-port="3" to-layer="115" to-port="0"/>
<edge from-layer="115" from-port="1" to-layer="116" to-port="0"/>
<edge from-layer="116" from-port="3" to-layer="117" to-port="0"/>
<edge from-layer="117" from-port="1" to-layer="118" to-port="0"/>
<edge from-layer="118" from-port="3" to-layer="119" to-port="0"/>
<edge from-layer="113" from-port="1" to-layer="119" to-port="1"/>
<edge from-layer="119" from-port="2" to-layer="120" to-port="0"/>
<edge from-layer="120" from-port="1" to-layer="121" to-port="0"/>
<edge from-layer="121" from-port="3" to-layer="122" to-port="0"/>
<edge from-layer="122" from-port="1" to-layer="123" to-port="0"/>
<edge from-layer="123" from-port="3" to-layer="124" to-port="0"/>
<edge from-layer="124" from-port="1" to-layer="125" to-port="0"/>
<edge from-layer="125" from-port="3" to-layer="126" to-port="0"/>
<edge from-layer="120" from-port="1" to-layer="126" to-port="1"/>
<edge from-layer="126" from-port="2" to-layer="127" to-port="0"/>
<edge from-layer="127" from-port="1" to-layer="128" to-port="0"/>
<edge from-layer="128" from-port="1" to-layer="129" to-port="0"/>
<edge from-layer="127" from-port="1" to-layer="130" to-port="0"/>
<edge from-layer="129" from-port="1" to-layer="130" to-port="1"/>
<edge from-layer="127" from-port="1" to-layer="131" to-port="0"/>
<edge from-layer="131" from-port="1" to-layer="132" to-port="0"/>
<edge from-layer="130" from-port="2" to-layer="133" to-port="0"/>
<edge from-layer="132" from-port="1" to-layer="133" to-port="1"/>
<edge from-layer="127" from-port="1" to-layer="134" to-port="0"/>
<edge from-layer="134" from-port="1" to-layer="135" to-port="0"/>
<edge from-layer="133" from-port="2" to-layer="136" to-port="0"/>
<edge from-layer="135" from-port="1" to-layer="136" to-port="1"/>
<edge from-layer="127" from-port="1" to-layer="137" to-port="0"/>
<edge from-layer="137" from-port="1" to-layer="138" to-port="0"/>
<edge from-layer="136" from-port="2" to-layer="139" to-port="0"/>
<edge from-layer="138" from-port="1" to-layer="139" to-port="1"/>
<edge from-layer="139" from-port="2" to-layer="140" to-port="0"/>
<edge from-layer="140" from-port="3" to-layer="141" to-port="0"/>
<edge from-layer="141" from-port="1" to-layer="142" to-port="0"/>
<edge from-layer="142" from-port="1" to-layer="143" to-port="0"/>
<edge from-layer="40" from-port="1" to-layer="144" to-port="0"/>
<edge from-layer="143" from-port="3" to-layer="145" to-port="0"/>
<edge from-layer="144" from-port="3" to-layer="145" to-port="1"/>
<edge from-layer="145" from-port="2" to-layer="146" to-port="0"/>
<edge from-layer="146" from-port="1" to-layer="147" to-port="0"/>
<edge from-layer="147" from-port="1" to-layer="148" to-port="0"/>
<edge from-layer="1" from-port="3" to-layer="149" to-port="0"/>
<edge from-layer="149" from-port="3" to-layer="150" to-port="0"/>
<edge from-layer="150" from-port="1" to-layer="151" to-port="0"/>
<edge from-layer="151" from-port="3" to-layer="152" to-port="0"/>
<edge from-layer="152" from-port="1" to-layer="153" to-port="0"/>
<edge from-layer="153" from-port="3" to-layer="154" to-port="0"/>
<edge from-layer="154" from-port="1" to-layer="155" to-port="0"/>
<edge from-layer="148" from-port="3" to-layer="156" to-port="0"/>
<edge from-layer="155" from-port="3" to-layer="156" to-port="1"/>
<edge from-layer="156" from-port="2" to-layer="157" to-port="0"/>
<edge from-layer="157" from-port="1" to-layer="158" to-port="0"/>
<edge from-layer="158" from-port="1" to-layer="159" to-port="0"/>
<edge from-layer="159" from-port="3" to-layer="160" to-port="0"/>
<edge from-layer="160" from-port="1" to-layer="161" to-port="0"/>
<edge from-layer="161" from-port="1" to-layer="162" to-port="0"/>
</edges>
<meta_data>
<MO_version value="1.5.4.dacdc0a0"/>
<cli_parameters>
<data_type value="FP16"/>
<disable_fusing value="False"/>
<disable_gfusing value="False"/>
<disable_nhwc_to_nchw value="False"/>
<disable_omitting_optional value="False"/>
<disable_resnet_optimization value="False"/>
<enable_flattening_nested_params value="False"/>
<extensions value="DIR"/>
<framework value="caffe"/>
<generate_deprecated_IR_V2 value="False"/>
<input value="data"/>
<input_model value="DIR/model.caffemodel"/>
<input_model_is_text value="False"/>
<input_proto value="DIR/model.prototxt"/>
<input_shape value="[1,3,1024,2048]"/>
<k value="DIR/CustomLayersMapping.xml"/>
<legacy_mxnet_model value="False"/>
<log_level value="ERROR"/>
<mean_values value="()"/>
<model_name value="semantic-segmentation-adas-0001"/>
<move_to_preprocess value="False"/>
<offload_unsupported_operations_to_tf value="False"/>
<output value="Rectify/Mul_"/>
<output_dir value="DIR"/>
<remove_output_softmax value="False"/>
<reverse_input_channels value="False"/>
<save_params_from_nd value="False"/>
<scale_values value="()"/>
<silent value="False"/>
<version value="False"/>
<unset unset_cli_parameters="batch, counts, finegrain_fusing, freeze_placeholder_with_value, input_checkpoint, input_meta_graph, input_symbol, mean_file, mean_file_offsets, nd_prefix_name, pretrained_model_name, saved_model_dir, saved_model_tags, scale, tensorboard_logdir, tensorflow_custom_layer_libraries, tensorflow_custom_operations_config_update, tensorflow_object_detection_api_pipeline_config, tensorflow_operation_patterns, tensorflow_subgraph_patterns, tensorflow_use_custom_operations_config"/>
</cli_parameters>
</meta_data>
</net> |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi
I changed the code (changed plugin device and changed model directory to FP16) to run the segmentation on NCS device and I am getting this error:
Traceback (most recent call last):
File "openvino_test_CPU-2.py", line 33, in
exec_net = plugin.load(network=net)
File "ie_api.pyx", line 305, in inference_engine.ie_api.IEPlugin.load
File "ie_api.pyx", line 318, in inference_engine.ie_api.IEPlugin.load
RuntimeError: Cannot convert layer "argmax" due to unsupported layer type "ArgMax"
/teamcity/work/scoring_engine_build/releases_openvino-2018-r4/ie_bridges/python/inference_engine/ie_api_impl.cpp:260
The default segmentation on CPU works just fine.
Kindly advise.
The text was updated successfully, but these errors were encountered: