Can not open OpenCL library on this device [Jetson Nano]

Question

Can not open OpenCL library on this device [Jetson Nano]

jamesee opened this issue 4 years ago · 16 comments

Hi Terry

I followed your instructions closely, compiled the gl2handpose on Jetson Nano and got the following error messages:

How do I install opencl?
[ see https://forums.developer.nvidia.com/t/opencl-support-on-jetson-nano/72584, opencl is not supported on Jetson Nano]

===========================================
$ export LD_LIBRARY_PATH=~/lib:$LD_LIBRARY_PATH
$ ./gl2handpose
INFO: Created TensorFlow Lite delegate for GPU.
ERROR: Can not open OpenCL library on this device - libOpenCL.so: cannot open shared object file: No such file or directory
ERROR: Falling back to OpenGL
INFO: Initialized OpenGL-based API.
INFO: Created 1 GPU delegate kernels.

Answer 1 · 2020-10-10T16:57:53.000Z

I used the below command to compile the example
make -j4 ENABLE_VDEC=true ENABLE_IMGUI=true TARGET_ENV=jetson_nano TFLITE_DELEGATE=GPU_DELEGATEV2

How do you compiled your code in jetson nano ?

Answer 2 · 2020-10-10T17:13:57.000Z

I used the below command to compile the example
make -j4 ENABLE_VDEC=true ENABLE_IMGUI=true TARGET_ENV=jetson_nano TFLITE_DELEGATE=GPU_DELEGATEV2

previously, I used
$ make -j4 TARGET_ENV=jetson_nano TFLITE_DELEGATE=GPU_DELEGATEV2

with the new compile command, it is still the same error: - cannot open OpenCL Library ...libOpenCL.so is missing

how did you install opencl on jetson nano?

=======================================================
$ rm gl2handpose
$ make -j4 ENABLE_VDEC=true ENABLE_IMGUI=true TARGET_ENV=jetson_nano TFLITE_DELEGATE=GPU_DELEGATEV2
g++ -o gl2handpose -Wl,--whole-archive main.o tflite_handpose.o render_handpose.o shapes.o touch_event.o custom_ops/transpose_conv_bias.o /home/james/work/tflite_gles_app/common/assertgl.o /home/james/work/tflite_gles_app/common/assertegl.o /home/james/work/tflite_gles_app/common/util_egl.o /home/james/work/tflite_gles_app/common/util_shader.o /home/james/work/tflite_gles_app/common/util_matrix.o /home/james/work/tflite_gles_app/common/util_texture.o /home/james/work/tflite_gles_app/common/util_render2d.o /home/james/work/tflite_gles_app/common/util_debugstr.o /home/james/work/tflite_gles_app/common/util_pmeter.o /home/james/work/tflite_gles_app/common/util_tflite.o /home/james/work/tflite_gles_app/common/winsys/winsys_x11.o camera_capture.o /home/james/work/tflite_gles_app/common/util_v4l2.o /home/james/work/tflite_gles_app/common/util_drm.o render_imgui.o /home/james/work/tflite_gles_app/third_party/imgui/imgui.o /home/james/work/tflite_gles_app/third_party/imgui/imgui_draw.o /home/james/work/tflite_gles_app/third_party/imgui/imgui_widgets.o /home/james/work/tflite_gles_app/third_party/imgui/examples/imgui_impl_opengl3.o -Wl,--allow-multiple-definition -L/home/james/work/tflite_gles_app/third_party/tensorflow/current/lite/lib/current/ -L/home/james/lib -lm -lEGL -lGLESv2 -lX11 -pthread -ldrm -ltensorflowlite -ltensorflowlite_gpu_delegate -Wl,--no-whole-archive -rdynamic

$ export DISPLAY=:0
$ xhost +
access control disabled, clients can connect from any host

$ ./gl2handpose
INFO: Created TensorFlow Lite delegate for GPU.
ERROR: Can not open OpenCL library on this device - libOpenCL.so: cannot open shared object file: No such file or directory
ERROR: Falling back to OpenGL
INFO: Initialized OpenGL-based API.
INFO: Created 1 GPU delegate kernels.

Answer 3 · 2020-10-10T17:19:54.000Z

unfortunately, OpenCL is not supported on Jetson Nano.
https://forums.developer.nvidia.com/t/opencl-support/74071

Answer 4 · 2020-10-10T17:21:49.000Z

unfortunately, OpenCL is not supported on Jetson Nano.
https://forums.developer.nvidia.com/t/opencl-support/74071

Your example code gl2handpose used opencl..., how do we overcome this issue?

Answer 5 · 2020-10-10T17:26:00.000Z

TensorFlow Lite GPU Delegate uses OpenGLES instead of OpenCL if a device doesn't support OpenCL. It's correct behavior. not an error.

Answer 6 · 2020-10-10T17:26:52.000Z

But I am getting a black window.

Answer 7 · 2020-10-10T17:30:16.000Z

hmm. it seems an another issue.
could you try below simple app ?
can you see a triangle ? If not, could you confirm your OpenGLES environment ?

cd misc/gl2tri
make
./gl2tri

Answer 8 · 2020-10-10T17:32:09.000Z

My Jetson nano setup compiles and run well.
Only thing i done differently from below steps
https://github.com/terryky/tflite_gles_app#build_for_aarch64

is that I have compiled and installed tflite on the target directly (jetson nano) and no cross compile

Answer 9 · 2020-10-10T17:38:42.000Z

can you see a triangle ?

Yes, i can see a rainbow triangle

Answer 10 · 2020-10-10T17:40:05.000Z

@palcode

How do you install bazel 3.1.0 on jetson nano? can you share with me the script? I want to try what you did.

Answer 11 · 2020-10-10T17:44:36.000Z

Hmm, I don't know why you see black window on gl2handpose.
Is there any another log ?
could you try with -x option ?

gl2handpose -x

Answer 12 · 2020-10-10T17:48:30.000Z

$ ./gl2handpose -x
INFO: Created TensorFlow Lite delegate for GPU.
ERROR: Can not open OpenCL library on this device - libOpenCL.so: cannot open shared object file: No such file or directory
ERROR: Falling back to OpenGL
INFO: Initialized OpenGL-based API.
INFO: Created 1 GPU delegate kernels.

   T E N S O R S

tensors size : 380
nodes size : 194
number of inputs : 1
number of outputs: 2

                 name                     bytes  type  scale   zero_point

Tensor[ 0] 786432, 1(fp32), ( 0, 0.000000) input Tensor[ 1] 3456, 1(fp32), ( 0, 0.000000) conv2d/Kernel Tensor[ 2] 128, 1(fp32), ( 0, 0.000000) conv2d/Bias Tensor[ 3] 2097152, 1(fp32), ( 0, 0.000000) conv2d Tensor[ 4] 2097152, 1(fp32), ( 0, 0.000000) activation Tensor[ 5] 1152, 1(fp32), ( 0, 0.000000) depthwise_conv2d/Kernel Tensor[ 6] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d/Bias Tensor[ 7] 2097152, 1(fp32), ( 0, 0.000000) depthwise_conv2d Tensor[ 8] 4096, 1(fp32), ( 0, 0.000000) conv2d_1/Kernel Tensor[ 9] 128, 1(fp32), ( 0, 0.000000) conv2d_1/Bias Tensor[ 10] 2097152, 1(fp32), ( 0, 0.000000) conv2d_1 Tensor[ 11] 2097152, 1(fp32), ( 0, 0.000000) add Tensor[ 12] 2097152, 1(fp32), ( 0, 0.000000) activation_1 Tensor[ 13] 1152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_1/Kernel Tensor[ 14] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d_1/Bias Tensor[ 15] 2097152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_1 Tensor[ 16] 4096, 1(fp32), ( 0, 0.000000) conv2d_2/Kernel Tensor[ 17] 128, 1(fp32), ( 0, 0.000000) conv2d_2/Bias Tensor[ 18] 2097152, 1(fp32), ( 0, 0.000000) conv2d_2 Tensor[ 19] 2097152, 1(fp32), ( 0, 0.000000) add_1 Tensor[ 20] 2097152, 1(fp32), ( 0, 0.000000) activation_2 Tensor[ 21] 1152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_2/Kernel Tensor[ 22] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d_2/Bias Tensor[ 23] 2097152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_2 Tensor[ 24] 4096, 1(fp32), ( 0, 0.000000) conv2d_3/Kernel Tensor[ 25] 128, 1(fp32), ( 0, 0.000000) conv2d_3/Bias Tensor[ 26] 2097152, 1(fp32), ( 0, 0.000000) conv2d_3 Tensor[ 27] 2097152, 1(fp32), ( 0, 0.000000) add_2 Tensor[ 28] 2097152, 1(fp32), ( 0, 0.000000) activation_3 Tensor[ 29] 1152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_3/Kernel Tensor[ 30] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d_3/Bias Tensor[ 31] 2097152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_3 Tensor[ 32] 4096, 1(fp32), ( 0, 0.000000) conv2d_4/Kernel Tensor[ 33] 128, 1(fp32), ( 0, 0.000000) conv2d_4/Bias Tensor[ 34] 2097152, 1(fp32), ( 0, 0.000000) conv2d_4 Tensor[ 35] 2097152, 1(fp32), ( 0, 0.000000) add_3 Tensor[ 36] 2097152, 1(fp32), ( 0, 0.000000) activation_4 Tensor[ 37] 1152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_4/Kernel Tensor[ 38] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d_4/Bias Tensor[ 39] 2097152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_4 Tensor[ 40] 4096, 1(fp32), ( 0, 0.000000) conv2d_5/Kernel Tensor[ 41] 128, 1(fp32), ( 0, 0.000000) conv2d_5/Bias Tensor[ 42] 2097152, 1(fp32), ( 0, 0.000000) conv2d_5 Tensor[ 43] 2097152, 1(fp32), ( 0, 0.000000) add_4 Tensor[ 44] 2097152, 1(fp32), ( 0, 0.000000) activation_5 Tensor[ 45] 1152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_5/Kernel Tensor[ 46] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d_5/Bias Tensor[ 47] 2097152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_5 Tensor[ 48] 4096, 1(fp32), ( 0, 0.000000) conv2d_6/Kernel Tensor[ 49] 128, 1(fp32), ( 0, 0.000000) conv2d_6/Bias Tensor[ 50] 2097152, 1(fp32), ( 0, 0.000000) conv2d_6 Tensor[ 51] 2097152, 1(fp32), ( 0, 0.000000) add_5 Tensor[ 52] 2097152, 1(fp32), ( 0, 0.000000) activation_6 Tensor[ 53] 1152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_6/Kernel Tensor[ 54] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d_6/Bias Tensor[ 55] 2097152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_6 Tensor[ 56] 4096, 1(fp32), ( 0, 0.000000) conv2d_7/Kernel Tensor[ 57] 128, 1(fp32), ( 0, 0.000000) conv2d_7/Bias Tensor[ 58] 2097152, 1(fp32), ( 0, 0.000000) conv2d_7 Tensor[ 59] 2097152, 1(fp32), ( 0, 0.000000) add_6 Tensor[ 60] 2097152, 1(fp32), ( 0, 0.000000) activation_7 Tensor[ 61] 1152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_7/Kernel Tensor[ 62] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d_7/Bias Tensor[ 63] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_7 Tensor[ 64] 524288, 1(fp32), ( 0, 0.000000) max_pooling2d Tensor[ 65] 8192, 1(fp32), ( 0, 0.000000) conv2d_8/Kernel Tensor[ 66] 256, 1(fp32), ( 0, 0.000000) conv2d_8/Bias Tensor[ 67] 1048576, 1(fp32), ( 0, 0.000000) conv2d_8 Tensor[ 68] 32, 2( i32), ( 0, 0.000000) channel_padding/Paddings Tensor[ 69] 1048576, 1(fp32), ( 0, 0.000000) channel_padding Tensor[ 70] 1048576, 1(fp32), ( 0, 0.000000) add_7 Tensor[ 71] 1048576, 1(fp32), ( 0, 0.000000) activation_8 Tensor[ 72] 2304, 1(fp32), ( 0, 0.000000) depthwise_conv2d_8/Kernel Tensor[ 73] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_8/Bias Tensor[ 74] 1048576, 1(fp32), ( 0, 0.000000) depthwise_conv2d_8 Tensor[ 75] 16384, 1(fp32), ( 0, 0.000000) conv2d_9/Kernel Tensor[ 76] 256, 1(fp32), ( 0, 0.000000) conv2d_9/Bias Tensor[ 77] 1048576, 1(fp32), ( 0, 0.000000) conv2d_9 Tensor[ 78] 1048576, 1(fp32), ( 0, 0.000000) add_8 Tensor[ 79] 1048576, 1(fp32), ( 0, 0.000000) activation_9 Tensor[ 80] 2304, 1(fp32), ( 0, 0.000000) depthwise_conv2d_9/Kernel Tensor[ 81] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_9/Bias Tensor[ 82] 1048576, 1(fp32), ( 0, 0.000000) depthwise_conv2d_9 Tensor[ 83] 16384, 1(fp32), ( 0, 0.000000) conv2d_10/Kernel Tensor[ 84] 256, 1(fp32), ( 0, 0.000000) conv2d_10/Bias Tensor[ 85] 1048576, 1(fp32), ( 0, 0.000000) conv2d_10 Tensor[ 86] 1048576, 1(fp32), ( 0, 0.000000) add_9 Tensor[ 87] 1048576, 1(fp32), ( 0, 0.000000) activation_10 Tensor[ 88] 2304, 1(fp32), ( 0, 0.000000) depthwise_conv2d_10/Kernel Tensor[ 89] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_10/Bias Tensor[ 90] 1048576, 1(fp32), ( 0, 0.000000) depthwise_conv2d_10 Tensor[ 91] 16384, 1(fp32), ( 0, 0.000000) conv2d_11/Kernel Tensor[ 92] 256, 1(fp32), ( 0, 0.000000) conv2d_11/Bias Tensor[ 93] 1048576, 1(fp32), ( 0, 0.000000) conv2d_11 Tensor[ 94] 1048576, 1(fp32), ( 0, 0.000000) add_10 Tensor[ 95] 1048576, 1(fp32), ( 0, 0.000000) activation_11 Tensor[ 96] 2304, 1(fp32), ( 0, 0.000000) depthwise_conv2d_11/Kernel Tensor[ 97] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_11/Bias Tensor[ 98] 1048576, 1(fp32), ( 0, 0.000000) depthwise_conv2d_11 Tensor[ 99] 16384, 1(fp32), ( 0, 0.000000) conv2d_12/Kernel Tensor[100] 256, 1(fp32), ( 0, 0.000000) conv2d_12/Bias Tensor[101] 1048576, 1(fp32), ( 0, 0.000000) conv2d_12 Tensor[102] 1048576, 1(fp32), ( 0, 0.000000) add_11 Tensor[103] 1048576, 1(fp32), ( 0, 0.000000) activation_12 Tensor[104] 2304, 1(fp32), ( 0, 0.000000) depthwise_conv2d_12/Kernel Tensor[105] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_12/Bias Tensor[106] 1048576, 1(fp32), ( 0, 0.000000) depthwise_conv2d_12 Tensor[107] 16384, 1(fp32), ( 0, 0.000000) conv2d_13/Kernel Tensor[108] 256, 1(fp32), ( 0, 0.000000) conv2d_13/Bias Tensor[109] 1048576, 1(fp32), ( 0, 0.000000) conv2d_13 Tensor[110] 1048576, 1(fp32), ( 0, 0.000000) add_12 Tensor[111] 1048576, 1(fp32), ( 0, 0.000000) activation_13 Tensor[112] 2304, 1(fp32), ( 0, 0.000000) depthwise_conv2d_13/Kernel Tensor[113] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_13/Bias Tensor[114] 1048576, 1(fp32), ( 0, 0.000000) depthwise_conv2d_13 Tensor[115] 16384, 1(fp32), ( 0, 0.000000) conv2d_14/Kernel Tensor[116] 256, 1(fp32), ( 0, 0.000000) conv2d_14/Bias Tensor[117] 1048576, 1(fp32), ( 0, 0.000000) conv2d_14 Tensor[118] 1048576, 1(fp32), ( 0, 0.000000) add_13 Tensor[119] 1048576, 1(fp32), ( 0, 0.000000) activation_14 Tensor[120] 2304, 1(fp32), ( 0, 0.000000) depthwise_conv2d_14/Kernel Tensor[121] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_14/Bias Tensor[122] 1048576, 1(fp32), ( 0, 0.000000) depthwise_conv2d_14 Tensor[123] 16384, 1(fp32), ( 0, 0.000000) conv2d_15/Kernel Tensor[124] 256, 1(fp32), ( 0, 0.000000) conv2d_15/Bias Tensor[125] 1048576, 1(fp32), ( 0, 0.000000) conv2d_15 Tensor[126] 1048576, 1(fp32), ( 0, 0.000000) add_14 Tensor[127] 1048576, 1(fp32), ( 0, 0.000000) activation_15 Tensor[128] 2304, 1(fp32), ( 0, 0.000000) depthwise_conv2d_15/Kernel Tensor[129] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_15/Bias Tensor[130] 262144, 1(fp32), ( 0, 0.000000) depthwise_conv2d_15 Tensor[131] 262144, 1(fp32), ( 0, 0.000000) max_pooling2d_1 Tensor[132] 32768, 1(fp32), ( 0, 0.000000) conv2d_16/Kernel Tensor[133] 512, 1(fp32), ( 0, 0.000000) conv2d_16/Bias Tensor[134] 524288, 1(fp32), ( 0, 0.000000) conv2d_16 Tensor[135] 32, 2( i32), ( 0, 0.000000) channel_padding_1/Paddings Tensor[136] 524288, 1(fp32), ( 0, 0.000000) channel_padding_1 Tensor[137] 524288, 1(fp32), ( 0, 0.000000) add_15 Tensor[138] 524288, 1(fp32), ( 0, 0.000000) activation_16 Tensor[139] 4608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_16/Kernel Tensor[140] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_16/Bias Tensor[141] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_16 Tensor[142] 65536, 1(fp32), ( 0, 0.000000) conv2d_17/Kernel Tensor[143] 512, 1(fp32), ( 0, 0.000000) conv2d_17/Bias Tensor[144] 524288, 1(fp32), ( 0, 0.000000) conv2d_17 Tensor[145] 524288, 1(fp32), ( 0, 0.000000) add_16 Tensor[146] 524288, 1(fp32), ( 0, 0.000000) activation_17 Tensor[147] 4608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_17/Kernel Tensor[148] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_17/Bias Tensor[149] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_17 Tensor[150] 65536, 1(fp32), ( 0, 0.000000) conv2d_18/Kernel Tensor[151] 512, 1(fp32), ( 0, 0.000000) conv2d_18/Bias Tensor[152] 524288, 1(fp32), ( 0, 0.000000) conv2d_18 Tensor[153] 524288, 1(fp32), ( 0, 0.000000) add_17 Tensor[154] 524288, 1(fp32), ( 0, 0.000000) activation_18 Tensor[155] 4608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_18/Kernel Tensor[156] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_18/Bias Tensor[157] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_18 Tensor[158] 65536, 1(fp32), ( 0, 0.000000) conv2d_19/Kernel Tensor[159] 512, 1(fp32), ( 0, 0.000000) conv2d_19/Bias Tensor[160] 524288, 1(fp32), ( 0, 0.000000) conv2d_19 Tensor[161] 524288, 1(fp32), ( 0, 0.000000) add_18 Tensor[162] 524288, 1(fp32), ( 0, 0.000000) activation_19 Tensor[163] 4608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_19/Kernel Tensor[164] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_19/Bias Tensor[165] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_19 Tensor[166] 65536, 1(fp32), ( 0, 0.000000) conv2d_20/Kernel Tensor[167] 512, 1(fp32), ( 0, 0.000000) conv2d_20/Bias Tensor[168] 524288, 1(fp32), ( 0, 0.000000) conv2d_20 Tensor[169] 524288, 1(fp32), ( 0, 0.000000) add_19 Tensor[170] 524288, 1(fp32), ( 0, 0.000000) activation_20 Tensor[171] 4608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_20/Kernel Tensor[172] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_20/Bias Tensor[173] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_20 Tensor[174] 65536, 1(fp32), ( 0, 0.000000) conv2d_21/Kernel Tensor[175] 512, 1(fp32), ( 0, 0.000000) conv2d_21/Bias Tensor[176] 524288, 1(fp32), ( 0, 0.000000) conv2d_21 Tensor[177] 524288, 1(fp32), ( 0, 0.000000) add_20 Tensor[178] 524288, 1(fp32), ( 0, 0.000000) activation_21 Tensor[179] 4608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_21/Kernel Tensor[180] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_21/Bias Tensor[181] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_21 Tensor[182] 65536, 1(fp32), ( 0, 0.000000) conv2d_22/Kernel Tensor[183] 512, 1(fp32), ( 0, 0.000000) conv2d_22/Bias Tensor[184] 524288, 1(fp32), ( 0, 0.000000) conv2d_22 Tensor[185] 524288, 1(fp32), ( 0, 0.000000) add_21 Tensor[186] 524288, 1(fp32), ( 0, 0.000000) activation_22 Tensor[187] 4608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_22/Kernel Tensor[188] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_22/Bias Tensor[189] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_22 Tensor[190] 65536, 1(fp32), ( 0, 0.000000) conv2d_23/Kernel Tensor[191] 512, 1(fp32), ( 0, 0.000000) conv2d_23/Bias Tensor[192] 524288, 1(fp32), ( 0, 0.000000) conv2d_23 Tensor[193] 524288, 1(fp32), ( 0, 0.000000) add_22 Tensor[194] 524288, 1(fp32), ( 0, 0.000000) activation_23 Tensor[195] 4608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_23/Kernel Tensor[196] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_23/Bias Tensor[197] 131072, 1(fp32), ( 0, 0.000000) depthwise_conv2d_23 Tensor[198] 131072, 1(fp32), ( 0, 0.000000) max_pooling2d_2 Tensor[199] 131072, 1(fp32), ( 0, 0.000000) conv2d_24/Kernel Tensor[200] 1024, 1(fp32), ( 0, 0.000000) conv2d_24/Bias Tensor[201] 262144, 1(fp32), ( 0, 0.000000) conv2d_24 Tensor[202] 32, 2( i32), ( 0, 0.000000) channel_padding_2/Paddings Tensor[203] 262144, 1(fp32), ( 0, 0.000000) channel_padding_2 Tensor[204] 262144, 1(fp32), ( 0, 0.000000) add_23 Tensor[205] 262144, 1(fp32), ( 0, 0.000000) activation_24 Tensor[206] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_24/Kernel Tensor[207] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_24/Bias Tensor[208] 262144, 1(fp32), ( 0, 0.000000) depthwise_conv2d_24 Tensor[209] 262144, 1(fp32), ( 0, 0.000000) conv2d_25/Kernel Tensor[210] 1024, 1(fp32), ( 0, 0.000000) conv2d_25/Bias Tensor[211] 262144, 1(fp32), ( 0, 0.000000) conv2d_25 Tensor[212] 262144, 1(fp32), ( 0, 0.000000) add_24 Tensor[213] 262144, 1(fp32), ( 0, 0.000000) activation_25 Tensor[214] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_25/Kernel Tensor[215] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_25/Bias Tensor[216] 262144, 1(fp32), ( 0, 0.000000) depthwise_conv2d_25 Tensor[217] 262144, 1(fp32), ( 0, 0.000000) conv2d_26/Kernel Tensor[218] 1024, 1(fp32), ( 0, 0.000000) conv2d_26/Bias Tensor[219] 262144, 1(fp32), ( 0, 0.000000) conv2d_26 Tensor[220] 262144, 1(fp32), ( 0, 0.000000) add_25 Tensor[221] 262144, 1(fp32), ( 0, 0.000000) activation_26 Tensor[222] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_26/Kernel Tensor[223] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_26/Bias Tensor[224] 262144, 1(fp32), ( 0, 0.000000) depthwise_conv2d_26 Tensor[225] 262144, 1(fp32), ( 0, 0.000000) conv2d_27/Kernel Tensor[226] 1024, 1(fp32), ( 0, 0.000000) conv2d_27/Bias Tensor[227] 262144, 1(fp32), ( 0, 0.000000) conv2d_27 Tensor[228] 262144, 1(fp32), ( 0, 0.000000) add_26 Tensor[229] 262144, 1(fp32), ( 0, 0.000000) activation_27 Tensor[230] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_27/Kernel Tensor[231] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_27/Bias Tensor[232] 262144, 1(fp32), ( 0, 0.000000) depthwise_conv2d_27 Tensor[233] 262144, 1(fp32), ( 0, 0.000000) conv2d_28/Kernel Tensor[234] 1024, 1(fp32), ( 0, 0.000000) conv2d_28/Bias Tensor[235] 262144, 1(fp32), ( 0, 0.000000) conv2d_28 Tensor[236] 262144, 1(fp32), ( 0, 0.000000) add_27 Tensor[237] 262144, 1(fp32), ( 0, 0.000000) activation_28 Tensor[238] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_28/Kernel Tensor[239] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_28/Bias Tensor[240] 262144, 1(fp32), ( 0, 0.000000) depthwise_conv2d_28 Tensor[241] 262144, 1(fp32), ( 0, 0.000000) conv2d_29/Kernel Tensor[242] 1024, 1(fp32), ( 0, 0.000000) conv2d_29/Bias Tensor[243] 262144, 1(fp32), ( 0, 0.000000) conv2d_29 Tensor[244] 262144, 1(fp32), ( 0, 0.000000) add_28 Tensor[245] 262144, 1(fp32), ( 0, 0.000000) activation_29 Tensor[246] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_29/Kernel Tensor[247] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_29/Bias Tensor[248] 262144, 1(fp32), ( 0, 0.000000) depthwise_conv2d_29 Tensor[249] 262144, 1(fp32), ( 0, 0.000000) conv2d_30/Kernel Tensor[250] 1024, 1(fp32), ( 0, 0.000000) conv2d_30/Bias Tensor[251] 262144, 1(fp32), ( 0, 0.000000) conv2d_30 Tensor[252] 262144, 1(fp32), ( 0, 0.000000) add_29 Tensor[253] 262144, 1(fp32), ( 0, 0.000000) activation_30 Tensor[254] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_30/Kernel Tensor[255] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_30/Bias Tensor[256] 262144, 1(fp32), ( 0, 0.000000) depthwise_conv2d_30 Tensor[257] 262144, 1(fp32), ( 0, 0.000000) conv2d_31/Kernel Tensor[258] 1024, 1(fp32), ( 0, 0.000000) conv2d_31/Bias Tensor[259] 262144, 1(fp32), ( 0, 0.000000) conv2d_31 Tensor[260] 262144, 1(fp32), ( 0, 0.000000) add_30 Tensor[261] 262144, 1(fp32), ( 0, 0.000000) activation_31 Tensor[262] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_31/Kernel Tensor[263] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_31/Bias Tensor[264] 65536, 1(fp32), ( 0, 0.000000) depthwise_conv2d_31 Tensor[265] 262144, 1(fp32), ( 0, 0.000000) conv2d_32/Kernel Tensor[266] 1024, 1(fp32), ( 0, 0.000000) conv2d_32/Bias Tensor[267] 65536, 1(fp32), ( 0, 0.000000) conv2d_32 Tensor[268] 65536, 1(fp32), ( 0, 0.000000) max_pooling2d_3 Tensor[269] 65536, 1(fp32), ( 0, 0.000000) add_31 Tensor[270] 65536, 1(fp32), ( 0, 0.000000) activation_32 Tensor[271] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_32/Kernel Tensor[272] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_32/Bias Tensor[273] 65536, 1(fp32), ( 0, 0.000000) depthwise_conv2d_32 Tensor[274] 262144, 1(fp32), ( 0, 0.000000) conv2d_33/Kernel Tensor[275] 1024, 1(fp32), ( 0, 0.000000) conv2d_33/Bias Tensor[276] 65536, 1(fp32), ( 0, 0.000000) conv2d_33 Tensor[277] 65536, 1(fp32), ( 0, 0.000000) add_32 Tensor[278] 65536, 1(fp32), ( 0, 0.000000) activation_33 Tensor[279] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_33/Kernel Tensor[280] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_33/Bias Tensor[281] 65536, 1(fp32), ( 0, 0.000000) depthwise_conv2d_33 Tensor[282] 262144, 1(fp32), ( 0, 0.000000) conv2d_34/Kernel Tensor[283] 1024, 1(fp32), ( 0, 0.000000) conv2d_34/Bias Tensor[284] 65536, 1(fp32), ( 0, 0.000000) conv2d_34 Tensor[285] 65536, 1(fp32), ( 0, 0.000000) add_33 Tensor[286] 65536, 1(fp32), ( 0, 0.000000) activation_34 Tensor[287] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_34/Kernel Tensor[288] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_34/Bias Tensor[289] 65536, 1(fp32), ( 0, 0.000000) depthwise_conv2d_34 Tensor[290] 262144, 1(fp32), ( 0, 0.000000) conv2d_35/Kernel Tensor[291] 1024, 1(fp32), ( 0, 0.000000) conv2d_35/Bias Tensor[292] 65536, 1(fp32), ( 0, 0.000000) conv2d_35 Tensor[293] 65536, 1(fp32), ( 0, 0.000000) add_34 Tensor[294] 65536, 1(fp32), ( 0, 0.000000) activation_35 Tensor[295] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_35/Kernel Tensor[296] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_35/Bias Tensor[297] 65536, 1(fp32), ( 0, 0.000000) depthwise_conv2d_35 Tensor[298] 262144, 1(fp32), ( 0, 0.000000) conv2d_36/Kernel Tensor[299] 1024, 1(fp32), ( 0, 0.000000) conv2d_36/Bias Tensor[300] 65536, 1(fp32), ( 0, 0.000000) conv2d_36 Tensor[301] 65536, 1(fp32), ( 0, 0.000000) add_35 Tensor[302] 65536, 1(fp32), ( 0, 0.000000) activation_36 Tensor[303] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_36/Kernel Tensor[304] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_36/Bias Tensor[305] 65536, 1(fp32), ( 0, 0.000000) depthwise_conv2d_36 Tensor[306] 262144, 1(fp32), ( 0, 0.000000) conv2d_37/Kernel Tensor[307] 1024, 1(fp32), ( 0, 0.000000) conv2d_37/Bias Tensor[308] 65536, 1(fp32), ( 0, 0.000000) conv2d_37 Tensor[309] 65536, 1(fp32), ( 0, 0.000000) add_36 Tensor[310] 65536, 1(fp32), ( 0, 0.000000) activation_37 Tensor[311] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_37/Kernel Tensor[312] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_37/Bias Tensor[313] 65536, 1(fp32), ( 0, 0.000000) depthwise_conv2d_37 Tensor[314] 262144, 1(fp32), ( 0, 0.000000) conv2d_38/Kernel Tensor[315] 1024, 1(fp32), ( 0, 0.000000) conv2d_38/Bias Tensor[316] 65536, 1(fp32), ( 0, 0.000000) conv2d_38 Tensor[317] 65536, 1(fp32), ( 0, 0.000000) add_37 Tensor[318] 65536, 1(fp32), ( 0, 0.000000) activation_38 Tensor[319] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_38/Kernel Tensor[320] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_38/Bias Tensor[321] 65536, 1(fp32), ( 0, 0.000000) depthwise_conv2d_38 Tensor[322] 262144, 1(fp32), ( 0, 0.000000) conv2d_39/Kernel Tensor[323] 1024, 1(fp32), ( 0, 0.000000) conv2d_39/Bias Tensor[324] 65536, 1(fp32), ( 0, 0.000000) conv2d_39 Tensor[325] 65536, 1(fp32), ( 0, 0.000000) add_38 Tensor[326] 65536, 1(fp32), ( 0, 0.000000) activation_39 Tensor[327] 1048576, 1(fp32), ( 0, 0.000000) conv2d_transpose/Kernel Tensor[328] 1024, 1(fp32), ( 0, 0.000000) conv2d_transpose/Bias Tensor[329] 262144, 1(fp32), ( 0, 0.000000) conv2d_transpose Tensor[330] 262144, 1(fp32), ( 0, 0.000000) activation_40 Tensor[331] 262144, 1(fp32), ( 0, 0.000000) add_39 Tensor[332] 9216, 1(fp32), ( 0, 0.000000) depthwise_conv2d_39/Kernel Tensor[333] 1024, 1(fp32), ( 0, 0.000000) depthwise_conv2d_39/Bias Tensor[334] 262144, 1(fp32), ( 0, 0.000000) depthwise_conv2d_39 Tensor[335] 262144, 1(fp32), ( 0, 0.000000) conv2d_40/Kernel Tensor[336] 1024, 1(fp32), ( 0, 0.000000) conv2d_40/Bias Tensor[337] 262144, 1(fp32), ( 0, 0.000000) conv2d_40 Tensor[338] 262144, 1(fp32), ( 0, 0.000000) add_40 Tensor[339] 262144, 1(fp32), ( 0, 0.000000) activation_41 Tensor[340] 524288, 1(fp32), ( 0, 0.000000) conv2d_transpose_1/Kernel Tensor[341] 512, 1(fp32), ( 0, 0.000000) conv2d_transpose_1/Bias Tensor[342] 524288, 1(fp32), ( 0, 0.000000) conv2d_transpose_1 Tensor[343] 524288, 1(fp32), ( 0, 0.000000) activation_42 Tensor[344] 524288, 1(fp32), ( 0, 0.000000) add_41 Tensor[345] 4608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_40/Kernel Tensor[346] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_40/Bias Tensor[347] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_40 Tensor[348] 65536, 1(fp32), ( 0, 0.000000) conv2d_41/Kernel Tensor[349] 512, 1(fp32), ( 0, 0.000000) conv2d_41/Bias Tensor[350] 524288, 1(fp32), ( 0, 0.000000) conv2d_41 Tensor[351] 524288, 1(fp32), ( 0, 0.000000) add_42 Tensor[352] 524288, 1(fp32), ( 0, 0.000000) activation_43 Tensor[353] 1024, 1(fp32), ( 0, 0.000000) classificator_8/Kernel Tensor[354] 8, 1(fp32), ( 0, 0.000000) classificator_8/Bias Tensor[355] 8192, 1(fp32), ( 0, 0.000000) classificator_8 Tensor[356] 2048, 1(fp32), ( 0, 0.000000) classificator_16/Kernel Tensor[357] 8, 1(fp32), ( 0, 0.000000) classificator_16/Bias Tensor[358] 2048, 1(fp32), ( 0, 0.000000) classificator_16 Tensor[359] 6144, 1(fp32), ( 0, 0.000000) classificator_32/Kernel Tensor[360] 24, 1(fp32), ( 0, 0.000000) classificator_32/Bias Tensor[361] 1536, 1(fp32), ( 0, 0.000000) classificator_32 Tensor[362] 18432, 1(fp32), ( 0, 0.000000) regressor_8/Kernel Tensor[363] 144, 1(fp32), ( 0, 0.000000) regressor_8/Bias Tensor[364] 147456, 1(fp32), ( 0, 0.000000) regressor_8 Tensor[365] 36864, 1(fp32), ( 0, 0.000000) regressor_16/Kernel Tensor[366] 144, 1(fp32), ( 0, 0.000000) regressor_16/Bias Tensor[367] 36864, 1(fp32), ( 0, 0.000000) regressor_16 Tensor[368] 110592, 1(fp32), ( 0, 0.000000) regressor_32/Kernel Tensor[369] 432, 1(fp32), ( 0, 0.000000) regressor_32/Bias Tensor[370] 27648, 1(fp32), ( 0, 0.000000) regressor_32 Tensor[371] 8192, 1(fp32), ( 0, 0.000000) reshape Tensor[372] 2048, 1(fp32), ( 0, 0.000000) reshape_2 Tensor[373] 1536, 1(fp32), ( 0, 0.000000) reshape_4 Tensor[374] 147456, 1(fp32), ( 0, 0.000000) reshape_1 Tensor[375] 36864, 1(fp32), ( 0, 0.000000) reshape_3 Tensor[376] 27648, 1(fp32), ( 0, 0.000000) reshape_5 Tensor[377] 11776, 1(fp32), ( 0, 0.000000) classificators Tensor[378] 211968, 1(fp32), ( 0, 0.000000) regressors Tensor[379] 1769472, 1(fp32), ( 0, 0.000000) (null) [1x256x256x3]
[32x3x3x3]
[32]
[1x128x128x32]
[1x128x128x32]
[1x3x3x32]
[32]
[1x128x128x32]
[32x1x1x32]
[32]
[1x128x128x32]
[1x128x128x32]
[1x128x128x32]
[1x3x3x32]
[32]
[1x128x128x32]
[32x1x1x32]
[32]
[1x128x128x32]
[1x128x128x32]
[1x128x128x32]
[1x3x3x32]
[32]
[1x128x128x32]
[32x1x1x32]
[32]
[1x128x128x32]
[1x128x128x32]
[1x128x128x32]
[1x3x3x32]
[32]
[1x128x128x32]
[32x1x1x32]
[32]
[1x128x128x32]
[1x128x128x32]
[1x128x128x32]
[1x3x3x32]
[32]
[1x128x128x32]
[32x1x1x32]
[32]
[1x128x128x32]
[1x128x128x32]
[1x128x128x32]
[1x3x3x32]
[32]
[1x128x128x32]
[32x1x1x32]
[32]
[1x128x128x32]
[1x128x128x32]
[1x128x128x32]
[1x3x3x32]
[32]
[1x128x128x32]
[32x1x1x32]
[32]
[1x128x128x32]
[1x128x128x32]
[1x128x128x32]
[1x3x3x32]
[32]
[1x64x64x32]
[1x64x64x32]
[64x1x1x32]
[64]
[1x64x64x64]
[4x2]
[1x64x64x64]
[1x64x64x64]
[1x64x64x64]
[1x3x3x64]
[64]
[1x64x64x64]
[64x1x1x64]
[64]
[1x64x64x64]
[1x64x64x64]
[1x64x64x64]
[1x3x3x64]
[64]
[1x64x64x64]
[64x1x1x64]
[64]
[1x64x64x64]
[1x64x64x64]
[1x64x64x64]
[1x3x3x64]
[64]
[1x64x64x64]
[64x1x1x64]
[64]
[1x64x64x64]
[1x64x64x64]
[1x64x64x64]
[1x3x3x64]
[64]
[1x64x64x64]
[64x1x1x64]
[64]
[1x64x64x64]
[1x64x64x64]
[1x64x64x64]
[1x3x3x64]
[64]
[1x64x64x64]
[64x1x1x64]
[64]
[1x64x64x64]
[1x64x64x64]
[1x64x64x64]
[1x3x3x64]
[64]
[1x64x64x64]
[64x1x1x64]
[64]
[1x64x64x64]
[1x64x64x64]
[1x64x64x64]
[1x3x3x64]
[64]
[1x64x64x64]
[64x1x1x64]
[64]
[1x64x64x64]
[1x64x64x64]
[1x64x64x64]
[1x3x3x64]
[64]
[1x32x32x64]
[1x32x32x64]
[128x1x1x64]
[128]
[1x32x32x128]
[4x2]
[1x32x32x128]
[1x32x32x128]
[1x32x32x128]
[1x3x3x128]
[128]
[1x32x32x128]
[128x1x1x128]
[128]
[1x32x32x128]
[1x32x32x128]
[1x32x32x128]
[1x3x3x128]
[128]
[1x32x32x128]
[128x1x1x128]
[128]
[1x32x32x128]
[1x32x32x128]
[1x32x32x128]
[1x3x3x128]
[128]
[1x32x32x128]
[128x1x1x128]
[128]
[1x32x32x128]
[1x32x32x128]
[1x32x32x128]
[1x3x3x128]
[128]
[1x32x32x128]
[128x1x1x128]
[128]
[1x32x32x128]
[1x32x32x128]
[1x32x32x128]
[1x3x3x128]
[128]
[1x32x32x128]
[128x1x1x128]
[128]
[1x32x32x128]
[1x32x32x128]
[1x32x32x128]
[1x3x3x128]
[128]
[1x32x32x128]
[128x1x1x128]
[128]
[1x32x32x128]
[1x32x32x128]
[1x32x32x128]
[1x3x3x128]
[128]
[1x32x32x128]
[128x1x1x128]
[128]
[1x32x32x128]
[1x32x32x128]
[1x32x32x128]
[1x3x3x128]
[128]
[1x16x16x128]
[1x16x16x128]
[256x1x1x128]
[256]
[1x16x16x256]
[4x2]
[1x16x16x256]
[1x16x16x256]
[1x16x16x256]
[1x3x3x256]
[256]
[1x16x16x256]
[256x1x1x256]
[256]
[1x16x16x256]
[1x16x16x256]
[1x16x16x256]
[1x3x3x256]
[256]
[1x16x16x256]
[256x1x1x256]
[256]
[1x16x16x256]
[1x16x16x256]
[1x16x16x256]
[1x3x3x256]
[256]
[1x16x16x256]
[256x1x1x256]
[256]
[1x16x16x256]
[1x16x16x256]
[1x16x16x256]
[1x3x3x256]
[256]
[1x16x16x256]
[256x1x1x256]
[256]
[1x16x16x256]
[1x16x16x256]
[1x16x16x256]
[1x3x3x256]
[256]
[1x16x16x256]
[256x1x1x256]
[256]
[1x16x16x256]
[1x16x16x256]
[1x16x16x256]
[1x3x3x256]
[256]
[1x16x16x256]
[256x1x1x256]
[256]
[1x16x16x256]
[1x16x16x256]
[1x16x16x256]
[1x3x3x256]
[256]
[1x16x16x256]
[256x1x1x256]
[256]
[1x16x16x256]
[1x16x16x256]
[1x16x16x256]
[1x3x3x256]
[256]
[1x8x8x256]
[256x1x1x256]
[256]
[1x8x8x256]
[1x8x8x256]
[1x8x8x256]
[1x8x8x256]
[1x3x3x256]
[256]
[1x8x8x256]
[256x1x1x256]
[256]
[1x8x8x256]
[1x8x8x256]
[1x8x8x256]
[1x3x3x256]
[256]
[1x8x8x256]
[256x1x1x256]
[256]
[1x8x8x256]
[1x8x8x256]
[1x8x8x256]
[1x3x3x256]
[256]
[1x8x8x256]
[256x1x1x256]
[256]
[1x8x8x256]
[1x8x8x256]
[1x8x8x256]
[1x3x3x256]
[256]
[1x8x8x256]
[256x1x1x256]
[256]
[1x8x8x256]
[1x8x8x256]
[1x8x8x256]
[1x3x3x256]
[256]
[1x8x8x256]
[256x1x1x256]
[256]
[1x8x8x256]
[1x8x8x256]
[1x8x8x256]
[1x3x3x256]
[256]
[1x8x8x256]
[256x1x1x256]
[256]
[1x8x8x256]
[1x8x8x256]
[1x8x8x256]
[1x3x3x256]
[256]
[1x8x8x256]
[256x1x1x256]
[256]
[1x8x8x256]
[1x8x8x256]
[1x8x8x256]
[256x2x2x256]
[256]
[1x16x16x256]
[1x16x16x256]
[1x16x16x256]
[1x3x3x256]
[256]
[1x16x16x256]
[256x1x1x256]
[256]
[1x16x16x256]
[1x16x16x256]
[1x16x16x256]
[128x2x2x256]
[128]
[1x32x32x128]
[1x32x32x128]
[1x32x32x128]
[1x3x3x128]
[128]
[1x32x32x128]
[128x1x1x128]
[128]
[1x32x32x128]
[1x32x32x128]
[1x32x32x128]
[2x1x1x128]
[2]
[1x32x32x2]
[2x1x1x256]
[2]
[1x16x16x2]
[6x1x1x256]
[6]
[1x8x8x6]
[36x1x1x128]
[36]
[1x32x32x36]
[36x1x1x256]
[36]
[1x16x16x36]
[108x1x1x256]
[108]
[1x8x8x108]
[1x2048x1]
[1x512x1]
[1x384x1]
[1x2048x18]
[1x512x18]
[1x384x18]
[1x2944x1]
[1x2944x18]
[1x128x128x27]

Input Tensor Dimension

Tensor[ 0] 786432, 1(fp32), ( 0, 0.000000) input [1x256x256x3]

Output Tensor Dimension

Tensor[378] 211968, 1(fp32), ( 0, 0.000000) regressors [1x2944x18]
Tensor[377] 11776, 1(fp32), ( 0, 0.000000) classificators [1x2944x1]

ERROR: Can not open OpenCL library on this device - libOpenCL.so: cannot open shared object file: No such file or directory
ERROR: Falling back to OpenGL
INFO: Created 1 GPU delegate kernels.

   T E N S O R S

tensors size : 390
nodes size : 197
number of inputs : 1
number of outputs: 2

                 name                     bytes  type  scale   zero_point

Tensor[ 0] 786432, 1(fp32), ( 0, 0.000000) input_1 [1x256x256x3]
Tensor[ 1] 3456, 1(fp32), ( 0, 0.000000) conv2d/Kernel [32x3x3x3]
Tensor[ 2] 128, 1(fp32), ( 0, 0.000000) conv2d/Bias [32]
Tensor[ 3] 2097152, 1(fp32), ( 0, 0.000000) conv2d [1x128x128x32]
Tensor[ 4] 128, 1(fp32), ( 0, 0.000000) p_re_lu/Alpha [1x1x32]
Tensor[ 5] 2097152, 1(fp32), ( 0, 0.000000) p_re_lu [1x128x128x32]
Tensor[ 6] 3200, 1(fp32), ( 0, 0.000000) depthwise_conv2d/Kernel [1x5x5x32]
Tensor[ 7] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d/Bias [32]
Tensor[ 8] 2097152, 1(fp32), ( 0, 0.000000) depthwise_conv2d [1x128x128x32]
Tensor[ 9] 4096, 1(fp32), ( 0, 0.000000) conv2d_1/Kernel [32x1x1x32]
Tensor[ 10] 128, 1(fp32), ( 0, 0.000000) conv2d_1/Bias [32]
Tensor[ 11] 2097152, 1(fp32), ( 0, 0.000000) conv2d_1 [1x128x128x32]
Tensor[ 12] 2097152, 1(fp32), ( 0, 0.000000) add [1x128x128x32]
Tensor[ 13] 2097152, 1(fp32), ( 0, 0.000000) activation [1x128x128x32]
Tensor[ 14] 3200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_1/Kernel [1x5x5x32]
Tensor[ 15] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d_1/Bias [32]
Tensor[ 16] 2097152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_1 [1x128x128x32]
Tensor[ 17] 4096, 1(fp32), ( 0, 0.000000) conv2d_2/Kernel [32x1x1x32]
Tensor[ 18] 128, 1(fp32), ( 0, 0.000000) conv2d_2/Bias [32]
Tensor[ 19] 2097152, 1(fp32), ( 0, 0.000000) conv2d_2 [1x128x128x32]
Tensor[ 20] 2097152, 1(fp32), ( 0, 0.000000) add_1 [1x128x128x32]
Tensor[ 21] 2097152, 1(fp32), ( 0, 0.000000) activation_1 [1x128x128x32]
Tensor[ 22] 3200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_2/Kernel [1x5x5x32]
Tensor[ 23] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d_2/Bias [32]
Tensor[ 24] 2097152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_2 [1x128x128x32]
Tensor[ 25] 4096, 1(fp32), ( 0, 0.000000) conv2d_3/Kernel [32x1x1x32]
Tensor[ 26] 128, 1(fp32), ( 0, 0.000000) conv2d_3/Bias [32]
Tensor[ 27] 2097152, 1(fp32), ( 0, 0.000000) conv2d_3 [1x128x128x32]
Tensor[ 28] 2097152, 1(fp32), ( 0, 0.000000) add_2 [1x128x128x32]
Tensor[ 29] 2097152, 1(fp32), ( 0, 0.000000) activation_2 [1x128x128x32]
Tensor[ 30] 3200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_3/Kernel [1x5x5x32]
Tensor[ 31] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d_3/Bias [32]
Tensor[ 32] 2097152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_3 [1x128x128x32]
Tensor[ 33] 4096, 1(fp32), ( 0, 0.000000) conv2d_4/Kernel [32x1x1x32]
Tensor[ 34] 128, 1(fp32), ( 0, 0.000000) conv2d_4/Bias [32]
Tensor[ 35] 2097152, 1(fp32), ( 0, 0.000000) conv2d_4 [1x128x128x32]
Tensor[ 36] 2097152, 1(fp32), ( 0, 0.000000) add_3 [1x128x128x32]
Tensor[ 37] 2097152, 1(fp32), ( 0, 0.000000) activation_3 [1x128x128x32]
Tensor[ 38] 3200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_4/Kernel [1x5x5x32]
Tensor[ 39] 128, 1(fp32), ( 0, 0.000000) depthwise_conv2d_4/Bias [32]
Tensor[ 40] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_4 [1x64x64x32]
Tensor[ 41] 524288, 1(fp32), ( 0, 0.000000) max_pooling2d [1x64x64x32]
Tensor[ 42] 8192, 1(fp32), ( 0, 0.000000) conv2d_5/Kernel [64x1x1x32]
Tensor[ 43] 256, 1(fp32), ( 0, 0.000000) conv2d_5/Bias [64]
Tensor[ 44] 1048576, 1(fp32), ( 0, 0.000000) conv2d_5 [1x64x64x64]
Tensor[ 45] 32, 2( i32), ( 0, 0.000000) channel_padding/Paddings [4x2]
Tensor[ 46] 1048576, 1(fp32), ( 0, 0.000000) channel_padding [1x64x64x64]
Tensor[ 47] 1048576, 1(fp32), ( 0, 0.000000) add_4 [1x64x64x64]
Tensor[ 48] 1048576, 1(fp32), ( 0, 0.000000) activation_4 [1x64x64x64]
Tensor[ 49] 6400, 1(fp32), ( 0, 0.000000) depthwise_conv2d_5/Kernel [1x5x5x64]
Tensor[ 50] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_5/Bias [64]
Tensor[ 51] 1048576, 1(fp32), ( 0, 0.000000) depthwise_conv2d_5 [1x64x64x64]
Tensor[ 52] 16384, 1(fp32), ( 0, 0.000000) conv2d_6/Kernel [64x1x1x64]
Tensor[ 53] 256, 1(fp32), ( 0, 0.000000) conv2d_6/Bias [64]
Tensor[ 54] 1048576, 1(fp32), ( 0, 0.000000) conv2d_6 [1x64x64x64]
Tensor[ 55] 1048576, 1(fp32), ( 0, 0.000000) add_5 [1x64x64x64]
Tensor[ 56] 1048576, 1(fp32), ( 0, 0.000000) activation_5 [1x64x64x64]
Tensor[ 57] 6400, 1(fp32), ( 0, 0.000000) depthwise_conv2d_6/Kernel [1x5x5x64]
Tensor[ 58] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_6/Bias [64]
Tensor[ 59] 1048576, 1(fp32), ( 0, 0.000000) depthwise_conv2d_6 [1x64x64x64]
Tensor[ 60] 16384, 1(fp32), ( 0, 0.000000) conv2d_7/Kernel [64x1x1x64]
Tensor[ 61] 256, 1(fp32), ( 0, 0.000000) conv2d_7/Bias [64]
Tensor[ 62] 1048576, 1(fp32), ( 0, 0.000000) conv2d_7 [1x64x64x64]
Tensor[ 63] 1048576, 1(fp32), ( 0, 0.000000) add_6 [1x64x64x64]
Tensor[ 64] 1048576, 1(fp32), ( 0, 0.000000) activation_6 [1x64x64x64]
Tensor[ 65] 6400, 1(fp32), ( 0, 0.000000) depthwise_conv2d_7/Kernel [1x5x5x64]
Tensor[ 66] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_7/Bias [64]
Tensor[ 67] 1048576, 1(fp32), ( 0, 0.000000) depthwise_conv2d_7 [1x64x64x64]
Tensor[ 68] 16384, 1(fp32), ( 0, 0.000000) conv2d_8/Kernel [64x1x1x64]
Tensor[ 69] 256, 1(fp32), ( 0, 0.000000) conv2d_8/Bias [64]
Tensor[ 70] 1048576, 1(fp32), ( 0, 0.000000) conv2d_8 [1x64x64x64]
Tensor[ 71] 1048576, 1(fp32), ( 0, 0.000000) add_7 [1x64x64x64]
Tensor[ 72] 1048576, 1(fp32), ( 0, 0.000000) activation_7 [1x64x64x64]
Tensor[ 73] 6400, 1(fp32), ( 0, 0.000000) depthwise_conv2d_8/Kernel [1x5x5x64]
Tensor[ 74] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_8/Bias [64]
Tensor[ 75] 1048576, 1(fp32), ( 0, 0.000000) depthwise_conv2d_8 [1x64x64x64]
Tensor[ 76] 16384, 1(fp32), ( 0, 0.000000) conv2d_9/Kernel [64x1x1x64]
Tensor[ 77] 256, 1(fp32), ( 0, 0.000000) conv2d_9/Bias [64]
Tensor[ 78] 1048576, 1(fp32), ( 0, 0.000000) conv2d_9 [1x64x64x64]
Tensor[ 79] 1048576, 1(fp32), ( 0, 0.000000) add_8 [1x64x64x64]
Tensor[ 80] 1048576, 1(fp32), ( 0, 0.000000) activation_8 [1x64x64x64]
Tensor[ 81] 6400, 1(fp32), ( 0, 0.000000) depthwise_conv2d_9/Kernel [1x5x5x64]
Tensor[ 82] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_9/Bias [64]
Tensor[ 83] 1048576, 1(fp32), ( 0, 0.000000) depthwise_conv2d_9 [1x64x64x64]
Tensor[ 84] 16384, 1(fp32), ( 0, 0.000000) conv2d_10/Kernel [64x1x1x64]
Tensor[ 85] 256, 1(fp32), ( 0, 0.000000) conv2d_10/Bias [64]
Tensor[ 86] 1048576, 1(fp32), ( 0, 0.000000) conv2d_10 [1x64x64x64]
Tensor[ 87] 1048576, 1(fp32), ( 0, 0.000000) add_9 [1x64x64x64]
Tensor[ 88] 1048576, 1(fp32), ( 0, 0.000000) activation_9 [1x64x64x64]
Tensor[ 89] 6400, 1(fp32), ( 0, 0.000000) depthwise_conv2d_10/Kernel [1x5x5x64]
Tensor[ 90] 256, 1(fp32), ( 0, 0.000000) depthwise_conv2d_10/Bias [64]
Tensor[ 91] 262144, 1(fp32), ( 0, 0.000000) depthwise_conv2d_10 [1x32x32x64]
Tensor[ 92] 262144, 1(fp32), ( 0, 0.000000) max_pooling2d_1 [1x32x32x64]
Tensor[ 93] 32768, 1(fp32), ( 0, 0.000000) conv2d_11/Kernel [128x1x1x64]
Tensor[ 94] 512, 1(fp32), ( 0, 0.000000) conv2d_11/Bias [128]
Tensor[ 95] 524288, 1(fp32), ( 0, 0.000000) conv2d_11 [1x32x32x128]
Tensor[ 96] 32, 2( i32), ( 0, 0.000000) channel_padding_1/Paddings [4x2]
Tensor[ 97] 524288, 1(fp32), ( 0, 0.000000) channel_padding_1 [1x32x32x128]
Tensor[ 98] 524288, 1(fp32), ( 0, 0.000000) add_10 [1x32x32x128]
Tensor[ 99] 524288, 1(fp32), ( 0, 0.000000) activation_10 [1x32x32x128]
Tensor[100] 12800, 1(fp32), ( 0, 0.000000) depthwise_conv2d_11/Kernel [1x5x5x128]
Tensor[101] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_11/Bias [128]
Tensor[102] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_11 [1x32x32x128]
Tensor[103] 65536, 1(fp32), ( 0, 0.000000) conv2d_12/Kernel [128x1x1x128]
Tensor[104] 512, 1(fp32), ( 0, 0.000000) conv2d_12/Bias [128]
Tensor[105] 524288, 1(fp32), ( 0, 0.000000) conv2d_12 [1x32x32x128]
Tensor[106] 524288, 1(fp32), ( 0, 0.000000) add_11 [1x32x32x128]
Tensor[107] 524288, 1(fp32), ( 0, 0.000000) activation_11 [1x32x32x128]
Tensor[108] 12800, 1(fp32), ( 0, 0.000000) depthwise_conv2d_12/Kernel [1x5x5x128]
Tensor[109] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_12/Bias [128]
Tensor[110] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_12 [1x32x32x128]
Tensor[111] 65536, 1(fp32), ( 0, 0.000000) conv2d_13/Kernel [128x1x1x128]
Tensor[112] 512, 1(fp32), ( 0, 0.000000) conv2d_13/Bias [128]
Tensor[113] 524288, 1(fp32), ( 0, 0.000000) conv2d_13 [1x32x32x128]
Tensor[114] 524288, 1(fp32), ( 0, 0.000000) add_12 [1x32x32x128]
Tensor[115] 524288, 1(fp32), ( 0, 0.000000) activation_12 [1x32x32x128]
Tensor[116] 12800, 1(fp32), ( 0, 0.000000) depthwise_conv2d_13/Kernel [1x5x5x128]
Tensor[117] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_13/Bias [128]
Tensor[118] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_13 [1x32x32x128]
Tensor[119] 65536, 1(fp32), ( 0, 0.000000) conv2d_14/Kernel [128x1x1x128]
Tensor[120] 512, 1(fp32), ( 0, 0.000000) conv2d_14/Bias [128]
Tensor[121] 524288, 1(fp32), ( 0, 0.000000) conv2d_14 [1x32x32x128]
Tensor[122] 524288, 1(fp32), ( 0, 0.000000) add_13 [1x32x32x128]
Tensor[123] 524288, 1(fp32), ( 0, 0.000000) activation_13 [1x32x32x128]
Tensor[124] 12800, 1(fp32), ( 0, 0.000000) depthwise_conv2d_14/Kernel [1x5x5x128]
Tensor[125] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_14/Bias [128]
Tensor[126] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_14 [1x32x32x128]
Tensor[127] 65536, 1(fp32), ( 0, 0.000000) conv2d_15/Kernel [128x1x1x128]
Tensor[128] 512, 1(fp32), ( 0, 0.000000) conv2d_15/Bias [128]
Tensor[129] 524288, 1(fp32), ( 0, 0.000000) conv2d_15 [1x32x32x128]
Tensor[130] 524288, 1(fp32), ( 0, 0.000000) add_14 [1x32x32x128]
Tensor[131] 524288, 1(fp32), ( 0, 0.000000) activation_14 [1x32x32x128]
Tensor[132] 12800, 1(fp32), ( 0, 0.000000) depthwise_conv2d_15/Kernel [1x5x5x128]
Tensor[133] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_15/Bias [128]
Tensor[134] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_15 [1x32x32x128]
Tensor[135] 65536, 1(fp32), ( 0, 0.000000) conv2d_16/Kernel [128x1x1x128]
Tensor[136] 512, 1(fp32), ( 0, 0.000000) conv2d_16/Bias [128]
Tensor[137] 524288, 1(fp32), ( 0, 0.000000) conv2d_16 [1x32x32x128]
Tensor[138] 524288, 1(fp32), ( 0, 0.000000) add_15 [1x32x32x128]
Tensor[139] 524288, 1(fp32), ( 0, 0.000000) activation_15 [1x32x32x128]
Tensor[140] 12800, 1(fp32), ( 0, 0.000000) depthwise_conv2d_16/Kernel [1x5x5x128]
Tensor[141] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_16/Bias [128]
Tensor[142] 524288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_16 [1x32x32x128]
Tensor[143] 65536, 1(fp32), ( 0, 0.000000) conv2d_17/Kernel [128x1x1x128]
Tensor[144] 512, 1(fp32), ( 0, 0.000000) conv2d_17/Bias [128]
Tensor[145] 524288, 1(fp32), ( 0, 0.000000) conv2d_17 [1x32x32x128]
Tensor[146] 524288, 1(fp32), ( 0, 0.000000) add_16 [1x32x32x128]
Tensor[147] 524288, 1(fp32), ( 0, 0.000000) activation_16 [1x32x32x128]
Tensor[148] 12800, 1(fp32), ( 0, 0.000000) depthwise_conv2d_17/Kernel [1x5x5x128]
Tensor[149] 512, 1(fp32), ( 0, 0.000000) depthwise_conv2d_17/Bias [128]
Tensor[150] 131072, 1(fp32), ( 0, 0.000000) depthwise_conv2d_17 [1x16x16x128]
Tensor[151] 131072, 1(fp32), ( 0, 0.000000) max_pooling2d_2 [1x16x16x128]
Tensor[152] 98304, 1(fp32), ( 0, 0.000000) conv2d_18/Kernel [192x1x1x128]
Tensor[153] 768, 1(fp32), ( 0, 0.000000) conv2d_18/Bias [192]
Tensor[154] 196608, 1(fp32), ( 0, 0.000000) conv2d_18 [1x16x16x192]
Tensor[155] 32, 2( i32), ( 0, 0.000000) channel_padding_2/Paddings [4x2]
Tensor[156] 196608, 1(fp32), ( 0, 0.000000) channel_padding_2 [1x16x16x192]
Tensor[157] 196608, 1(fp32), ( 0, 0.000000) add_17 [1x16x16x192]
Tensor[158] 196608, 1(fp32), ( 0, 0.000000) activation_17 [1x16x16x192]
Tensor[159] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_18/Kernel [1x5x5x192]
Tensor[160] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_18/Bias [192]
Tensor[161] 196608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_18 [1x16x16x192]
Tensor[162] 147456, 1(fp32), ( 0, 0.000000) conv2d_19/Kernel [192x1x1x192]
Tensor[163] 768, 1(fp32), ( 0, 0.000000) conv2d_19/Bias [192]
Tensor[164] 196608, 1(fp32), ( 0, 0.000000) conv2d_19 [1x16x16x192]
Tensor[165] 196608, 1(fp32), ( 0, 0.000000) add_18 [1x16x16x192]
Tensor[166] 196608, 1(fp32), ( 0, 0.000000) activation_18 [1x16x16x192]
Tensor[167] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_19/Kernel [1x5x5x192]
Tensor[168] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_19/Bias [192]
Tensor[169] 196608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_19 [1x16x16x192]
Tensor[170] 147456, 1(fp32), ( 0, 0.000000) conv2d_20/Kernel [192x1x1x192]
Tensor[171] 768, 1(fp32), ( 0, 0.000000) conv2d_20/Bias [192]
Tensor[172] 196608, 1(fp32), ( 0, 0.000000) conv2d_20 [1x16x16x192]
Tensor[173] 196608, 1(fp32), ( 0, 0.000000) add_19 [1x16x16x192]
Tensor[174] 196608, 1(fp32), ( 0, 0.000000) activation_19 [1x16x16x192]
Tensor[175] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_20/Kernel [1x5x5x192]
Tensor[176] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_20/Bias [192]
Tensor[177] 196608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_20 [1x16x16x192]
Tensor[178] 147456, 1(fp32), ( 0, 0.000000) conv2d_21/Kernel [192x1x1x192]
Tensor[179] 768, 1(fp32), ( 0, 0.000000) conv2d_21/Bias [192]
Tensor[180] 196608, 1(fp32), ( 0, 0.000000) conv2d_21 [1x16x16x192]
Tensor[181] 196608, 1(fp32), ( 0, 0.000000) add_20 [1x16x16x192]
Tensor[182] 196608, 1(fp32), ( 0, 0.000000) activation_20 [1x16x16x192]
Tensor[183] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_21/Kernel [1x5x5x192]
Tensor[184] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_21/Bias [192]
Tensor[185] 196608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_21 [1x16x16x192]
Tensor[186] 147456, 1(fp32), ( 0, 0.000000) conv2d_22/Kernel [192x1x1x192]
Tensor[187] 768, 1(fp32), ( 0, 0.000000) conv2d_22/Bias [192]
Tensor[188] 196608, 1(fp32), ( 0, 0.000000) conv2d_22 [1x16x16x192]
Tensor[189] 196608, 1(fp32), ( 0, 0.000000) add_21 [1x16x16x192]
Tensor[190] 196608, 1(fp32), ( 0, 0.000000) activation_21 [1x16x16x192]
Tensor[191] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_22/Kernel [1x5x5x192]
Tensor[192] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_22/Bias [192]
Tensor[193] 196608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_22 [1x16x16x192]
Tensor[194] 147456, 1(fp32), ( 0, 0.000000) conv2d_23/Kernel [192x1x1x192]
Tensor[195] 768, 1(fp32), ( 0, 0.000000) conv2d_23/Bias [192]
Tensor[196] 196608, 1(fp32), ( 0, 0.000000) conv2d_23 [1x16x16x192]
Tensor[197] 196608, 1(fp32), ( 0, 0.000000) add_22 [1x16x16x192]
Tensor[198] 196608, 1(fp32), ( 0, 0.000000) activation_22 [1x16x16x192]
Tensor[199] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_23/Kernel [1x5x5x192]
Tensor[200] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_23/Bias [192]
Tensor[201] 196608, 1(fp32), ( 0, 0.000000) depthwise_conv2d_23 [1x16x16x192]
Tensor[202] 147456, 1(fp32), ( 0, 0.000000) conv2d_24/Kernel [192x1x1x192]
Tensor[203] 768, 1(fp32), ( 0, 0.000000) conv2d_24/Bias [192]
Tensor[204] 196608, 1(fp32), ( 0, 0.000000) conv2d_24 [1x16x16x192]
Tensor[205] 196608, 1(fp32), ( 0, 0.000000) add_23 [1x16x16x192]
Tensor[206] 196608, 1(fp32), ( 0, 0.000000) activation_23 [1x16x16x192]
Tensor[207] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_24/Kernel [1x5x5x192]
Tensor[208] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_24/Bias [192]
Tensor[209] 49152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_24 [1x8x8x192]
Tensor[210] 147456, 1(fp32), ( 0, 0.000000) conv2d_25/Kernel [192x1x1x192]
Tensor[211] 768, 1(fp32), ( 0, 0.000000) conv2d_25/Bias [192]
Tensor[212] 49152, 1(fp32), ( 0, 0.000000) conv2d_25 [1x8x8x192]
Tensor[213] 49152, 1(fp32), ( 0, 0.000000) max_pooling2d_3 [1x8x8x192]
Tensor[214] 49152, 1(fp32), ( 0, 0.000000) add_24 [1x8x8x192]
Tensor[215] 49152, 1(fp32), ( 0, 0.000000) activation_24 [1x8x8x192]
Tensor[216] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_25/Kernel [1x5x5x192]
Tensor[217] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_25/Bias [192]
Tensor[218] 49152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_25 [1x8x8x192]
Tensor[219] 147456, 1(fp32), ( 0, 0.000000) conv2d_26/Kernel [192x1x1x192]
Tensor[220] 768, 1(fp32), ( 0, 0.000000) conv2d_26/Bias [192]
Tensor[221] 49152, 1(fp32), ( 0, 0.000000) conv2d_26 [1x8x8x192]
Tensor[222] 49152, 1(fp32), ( 0, 0.000000) add_25 [1x8x8x192]
Tensor[223] 49152, 1(fp32), ( 0, 0.000000) activation_25 [1x8x8x192]
Tensor[224] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_26/Kernel [1x5x5x192]
Tensor[225] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_26/Bias [192]
Tensor[226] 49152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_26 [1x8x8x192]
Tensor[227] 147456, 1(fp32), ( 0, 0.000000) conv2d_27/Kernel [192x1x1x192]
Tensor[228] 768, 1(fp32), ( 0, 0.000000) conv2d_27/Bias [192]
Tensor[229] 49152, 1(fp32), ( 0, 0.000000) conv2d_27 [1x8x8x192]
Tensor[230] 49152, 1(fp32), ( 0, 0.000000) add_26 [1x8x8x192]
Tensor[231] 49152, 1(fp32), ( 0, 0.000000) activation_26 [1x8x8x192]
Tensor[232] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_27/Kernel [1x5x5x192]
Tensor[233] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_27/Bias [192]
Tensor[234] 49152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_27 [1x8x8x192]
Tensor[235] 147456, 1(fp32), ( 0, 0.000000) conv2d_28/Kernel [192x1x1x192]
Tensor[236] 768, 1(fp32), ( 0, 0.000000) conv2d_28/Bias [192]
Tensor[237] 49152, 1(fp32), ( 0, 0.000000) conv2d_28 [1x8x8x192]
Tensor[238] 49152, 1(fp32), ( 0, 0.000000) add_27 [1x8x8x192]
Tensor[239] 49152, 1(fp32), ( 0, 0.000000) activation_27 [1x8x8x192]
Tensor[240] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_28/Kernel [1x5x5x192]
Tensor[241] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_28/Bias [192]
Tensor[242] 49152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_28 [1x8x8x192]
Tensor[243] 147456, 1(fp32), ( 0, 0.000000) conv2d_29/Kernel [192x1x1x192]
Tensor[244] 768, 1(fp32), ( 0, 0.000000) conv2d_29/Bias [192]
Tensor[245] 49152, 1(fp32), ( 0, 0.000000) conv2d_29 [1x8x8x192]
Tensor[246] 49152, 1(fp32), ( 0, 0.000000) add_28 [1x8x8x192]
Tensor[247] 49152, 1(fp32), ( 0, 0.000000) activation_28 [1x8x8x192]
Tensor[248] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_29/Kernel [1x5x5x192]
Tensor[249] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_29/Bias [192]
Tensor[250] 49152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_29 [1x8x8x192]
Tensor[251] 147456, 1(fp32), ( 0, 0.000000) conv2d_30/Kernel [192x1x1x192]
Tensor[252] 768, 1(fp32), ( 0, 0.000000) conv2d_30/Bias [192]
Tensor[253] 49152, 1(fp32), ( 0, 0.000000) conv2d_30 [1x8x8x192]
Tensor[254] 49152, 1(fp32), ( 0, 0.000000) add_29 [1x8x8x192]
Tensor[255] 49152, 1(fp32), ( 0, 0.000000) activation_29 [1x8x8x192]
Tensor[256] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_30/Kernel [1x5x5x192]
Tensor[257] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_30/Bias [192]
Tensor[258] 49152, 1(fp32), ( 0, 0.000000) depthwise_conv2d_30 [1x8x8x192]
Tensor[259] 147456, 1(fp32), ( 0, 0.000000) conv2d_31/Kernel [192x1x1x192]
Tensor[260] 768, 1(fp32), ( 0, 0.000000) conv2d_31/Bias [192]
Tensor[261] 49152, 1(fp32), ( 0, 0.000000) conv2d_31 [1x8x8x192]
Tensor[262] 49152, 1(fp32), ( 0, 0.000000) add_30 [1x8x8x192]
Tensor[263] 49152, 1(fp32), ( 0, 0.000000) activation_30 [1x8x8x192]
Tensor[264] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_31/Kernel [1x5x5x192]
Tensor[265] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_31/Bias [192]
Tensor[266] 12288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_31 [1x4x4x192]
Tensor[267] 147456, 1(fp32), ( 0, 0.000000) conv2d_32/Kernel [192x1x1x192]
Tensor[268] 768, 1(fp32), ( 0, 0.000000) conv2d_32/Bias [192]
Tensor[269] 12288, 1(fp32), ( 0, 0.000000) conv2d_32 [1x4x4x192]
Tensor[270] 12288, 1(fp32), ( 0, 0.000000) max_pooling2d_4 [1x4x4x192]
Tensor[271] 12288, 1(fp32), ( 0, 0.000000) add_31 [1x4x4x192]
Tensor[272] 12288, 1(fp32), ( 0, 0.000000) activation_31 [1x4x4x192]
Tensor[273] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_32/Kernel [1x5x5x192]
Tensor[274] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_32/Bias [192]
Tensor[275] 12288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_32 [1x4x4x192]
Tensor[276] 147456, 1(fp32), ( 0, 0.000000) conv2d_33/Kernel [192x1x1x192]
Tensor[277] 768, 1(fp32), ( 0, 0.000000) conv2d_33/Bias [192]
Tensor[278] 12288, 1(fp32), ( 0, 0.000000) conv2d_33 [1x4x4x192]
Tensor[279] 12288, 1(fp32), ( 0, 0.000000) add_32 [1x4x4x192]
Tensor[280] 12288, 1(fp32), ( 0, 0.000000) activation_32 [1x4x4x192]
Tensor[281] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_33/Kernel [1x5x5x192]
Tensor[282] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_33/Bias [192]
Tensor[283] 12288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_33 [1x4x4x192]
Tensor[284] 147456, 1(fp32), ( 0, 0.000000) conv2d_34/Kernel [192x1x1x192]
Tensor[285] 768, 1(fp32), ( 0, 0.000000) conv2d_34/Bias [192]
Tensor[286] 12288, 1(fp32), ( 0, 0.000000) conv2d_34 [1x4x4x192]
Tensor[287] 12288, 1(fp32), ( 0, 0.000000) add_33 [1x4x4x192]
Tensor[288] 12288, 1(fp32), ( 0, 0.000000) activation_33 [1x4x4x192]
Tensor[289] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_34/Kernel [1x5x5x192]
Tensor[290] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_34/Bias [192]
Tensor[291] 12288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_34 [1x4x4x192]
Tensor[292] 147456, 1(fp32), ( 0, 0.000000) conv2d_35/Kernel [192x1x1x192]
Tensor[293] 768, 1(fp32), ( 0, 0.000000) conv2d_35/Bias [192]
Tensor[294] 12288, 1(fp32), ( 0, 0.000000) conv2d_35 [1x4x4x192]
Tensor[295] 12288, 1(fp32), ( 0, 0.000000) add_34 [1x4x4x192]
Tensor[296] 12288, 1(fp32), ( 0, 0.000000) activation_34 [1x4x4x192]
Tensor[297] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_35/Kernel [1x5x5x192]
Tensor[298] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_35/Bias [192]
Tensor[299] 12288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_35 [1x4x4x192]
Tensor[300] 147456, 1(fp32), ( 0, 0.000000) conv2d_36/Kernel [192x1x1x192]
Tensor[301] 768, 1(fp32), ( 0, 0.000000) conv2d_36/Bias [192]
Tensor[302] 12288, 1(fp32), ( 0, 0.000000) conv2d_36 [1x4x4x192]
Tensor[303] 12288, 1(fp32), ( 0, 0.000000) add_35 [1x4x4x192]
Tensor[304] 12288, 1(fp32), ( 0, 0.000000) activation_35 [1x4x4x192]
Tensor[305] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_36/Kernel [1x5x5x192]
Tensor[306] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_36/Bias [192]
Tensor[307] 12288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_36 [1x4x4x192]
Tensor[308] 147456, 1(fp32), ( 0, 0.000000) conv2d_37/Kernel [192x1x1x192]
Tensor[309] 768, 1(fp32), ( 0, 0.000000) conv2d_37/Bias [192]
Tensor[310] 12288, 1(fp32), ( 0, 0.000000) conv2d_37 [1x4x4x192]
Tensor[311] 12288, 1(fp32), ( 0, 0.000000) add_36 [1x4x4x192]
Tensor[312] 12288, 1(fp32), ( 0, 0.000000) activation_36 [1x4x4x192]
Tensor[313] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_37/Kernel [1x5x5x192]
Tensor[314] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_37/Bias [192]
Tensor[315] 12288, 1(fp32), ( 0, 0.000000) depthwise_conv2d_37 [1x4x4x192]
Tensor[316] 147456, 1(fp32), ( 0, 0.000000) conv2d_38/Kernel [192x1x1x192]
Tensor[317] 768, 1(fp32), ( 0, 0.000000) conv2d_38/Bias [192]
Tensor[318] 12288, 1(fp32), ( 0, 0.000000) conv2d_38 [1x4x4x192]
Tensor[319] 12288, 1(fp32), ( 0, 0.000000) add_37 [1x4x4x192]
Tensor[320] 12288, 1(fp32), ( 0, 0.000000) activation_37 [1x4x4x192]
Tensor[321] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_38/Kernel [1x5x5x192]
Tensor[322] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_38/Bias [192]
Tensor[323] 3072, 1(fp32), ( 0, 0.000000) depthwise_conv2d_38 [1x2x2x192]
Tensor[324] 147456, 1(fp32), ( 0, 0.000000) conv2d_39/Kernel [192x1x1x192]
Tensor[325] 768, 1(fp32), ( 0, 0.000000) conv2d_39/Bias [192]
Tensor[326] 3072, 1(fp32), ( 0, 0.000000) conv2d_39 [1x2x2x192]
Tensor[327] 3072, 1(fp32), ( 0, 0.000000) max_pooling2d_5 [1x2x2x192]
Tensor[328] 3072, 1(fp32), ( 0, 0.000000) add_38 [1x2x2x192]
Tensor[329] 3072, 1(fp32), ( 0, 0.000000) activation_38 [1x2x2x192]
Tensor[330] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_39/Kernel [1x5x5x192]
Tensor[331] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_39/Bias [192]
Tensor[332] 3072, 1(fp32), ( 0, 0.000000) depthwise_conv2d_39 [1x2x2x192]
Tensor[333] 147456, 1(fp32), ( 0, 0.000000) conv2d_40/Kernel [192x1x1x192]
Tensor[334] 768, 1(fp32), ( 0, 0.000000) conv2d_40/Bias [192]
Tensor[335] 3072, 1(fp32), ( 0, 0.000000) conv2d_40 [1x2x2x192]
Tensor[336] 3072, 1(fp32), ( 0, 0.000000) add_39 [1x2x2x192]
Tensor[337] 3072, 1(fp32), ( 0, 0.000000) activation_39 [1x2x2x192]
Tensor[338] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_40/Kernel [1x5x5x192]
Tensor[339] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_40/Bias [192]
Tensor[340] 3072, 1(fp32), ( 0, 0.000000) depthwise_conv2d_40 [1x2x2x192]
Tensor[341] 147456, 1(fp32), ( 0, 0.000000) conv2d_41/Kernel [192x1x1x192]
Tensor[342] 768, 1(fp32), ( 0, 0.000000) conv2d_41/Bias [192]
Tensor[343] 3072, 1(fp32), ( 0, 0.000000) conv2d_41 [1x2x2x192]
Tensor[344] 3072, 1(fp32), ( 0, 0.000000) add_40 [1x2x2x192]
Tensor[345] 3072, 1(fp32), ( 0, 0.000000) activation_40 [1x2x2x192]
Tensor[346] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_41/Kernel [1x5x5x192]
Tensor[347] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_41/Bias [192]
Tensor[348] 3072, 1(fp32), ( 0, 0.000000) depthwise_conv2d_41 [1x2x2x192]
Tensor[349] 147456, 1(fp32), ( 0, 0.000000) conv2d_42/Kernel [192x1x1x192]
Tensor[350] 768, 1(fp32), ( 0, 0.000000) conv2d_42/Bias [192]
Tensor[351] 3072, 1(fp32), ( 0, 0.000000) conv2d_42 [1x2x2x192]
Tensor[352] 3072, 1(fp32), ( 0, 0.000000) add_41 [1x2x2x192]
Tensor[353] 3072, 1(fp32), ( 0, 0.000000) activation_41 [1x2x2x192]
Tensor[354] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_42/Kernel [1x5x5x192]
Tensor[355] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_42/Bias [192]
Tensor[356] 3072, 1(fp32), ( 0, 0.000000) depthwise_conv2d_42 [1x2x2x192]
Tensor[357] 147456, 1(fp32), ( 0, 0.000000) conv2d_43/Kernel [192x1x1x192]
Tensor[358] 768, 1(fp32), ( 0, 0.000000) conv2d_43/Bias [192]
Tensor[359] 3072, 1(fp32), ( 0, 0.000000) conv2d_43 [1x2x2x192]
Tensor[360] 3072, 1(fp32), ( 0, 0.000000) add_42 [1x2x2x192]
Tensor[361] 3072, 1(fp32), ( 0, 0.000000) activation_42 [1x2x2x192]
Tensor[362] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_43/Kernel [1x5x5x192]
Tensor[363] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_43/Bias [192]
Tensor[364] 3072, 1(fp32), ( 0, 0.000000) depthwise_conv2d_43 [1x2x2x192]
Tensor[365] 147456, 1(fp32), ( 0, 0.000000) conv2d_44/Kernel [192x1x1x192]
Tensor[366] 768, 1(fp32), ( 0, 0.000000) conv2d_44/Bias [192]
Tensor[367] 3072, 1(fp32), ( 0, 0.000000) conv2d_44 [1x2x2x192]
Tensor[368] 3072, 1(fp32), ( 0, 0.000000) add_43 [1x2x2x192]
Tensor[369] 3072, 1(fp32), ( 0, 0.000000) activation_43 [1x2x2x192]
Tensor[370] 19200, 1(fp32), ( 0, 0.000000) depthwise_conv2d_44/Kernel [1x5x5x192]
Tensor[371] 768, 1(fp32), ( 0, 0.000000) depthwise_conv2d_44/Bias [192]
Tensor[372] 3072, 1(fp32), ( 0, 0.000000) depthwise_conv2d_44 [1x2x2x192]
Tensor[373] 147456, 1(fp32), ( 0, 0.000000) conv2d_45/Kernel [192x1x1x192]
Tensor[374] 768, 1(fp32), ( 0, 0.000000) conv2d_45/Bias [192]
Tensor[375] 3072, 1(fp32), ( 0, 0.000000) conv2d_45 [1x2x2x192]
Tensor[376] 3072, 1(fp32), ( 0, 0.000000) add_44 [1x2x2x192]
Tensor[377] 3072, 1(fp32), ( 0, 0.000000) activation_44 [1x2x2x192]
Tensor[378] 3072, 1(fp32), ( 0, 0.000000) conv_handflag/Kernel [1x2x2x192]
Tensor[379] 4, 1(fp32), ( 0, 0.000000) conv_handflag/Bias [1]
Tensor[380] 4, 1(fp32), ( 0, 0.000000) conv_handflag [1x1x1x1]
Tensor[381] 193536, 1(fp32), ( 0, 0.000000) convld_21_3d/Kernel [63x2x2x192]
Tensor[382] 252, 1(fp32), ( 0, 0.000000) convld_21_3d/Bias [63]
Tensor[383] 252, 1(fp32), ( 0, 0.000000) convld_21_3d [1x1x1x63]
Tensor[384] 4, 1(fp32), ( 0, 0.000000) activation_handflag [1x1x1x1]
Tensor[385] 252, 1(fp32), ( 0, 0.000000) ld_21_3d [1x63]
Tensor[386] 4, 1(fp32), ( 0, 0.000000) output_handflag [1x1]
Tensor[387] 1769472, 1(fp32), ( 0, 0.000000) (null) [1x128x128x27]
Tensor[388] 3072, 1(fp32), ( 0, 0.000000) (null) [1x1x1x768]
Tensor[389] 3072, 1(fp32), ( 0, 0.000000) (null) [1x1x1x768]

Input Tensor Dimension

Tensor[ 0] 786432, 1(fp32), ( 0, 0.000000) input_1 [1x256x256x3]

Output Tensor Dimension

Tensor[385] 252, 1(fp32), ( 0, 0.000000) ld_21_3d [1x63]
Tensor[386] 4, 1(fp32), ( 0, 0.000000) output_handflag [1x1]

Answer 13 · 2020-10-10T17:55:53.000Z

is that I have compiled and installed tflite on the target directly (jetson nano) and no cross compile

@palcode
Can share how you install tflite on Jetson Nano directly?

Answer 14 · 2020-10-11T05:29:58.000Z

@terryky

I managed to get it working.
I could have compiled the tflite for host-pc (x86) and uploaded to Jetson Nano. That is why I gotten a black screen.

I re-pulled your github repo and re-installed it all over. And it works. Awesome !!!

Thanks for the prompt support. Appreciate it.

Answer 15 · 2020-10-11T05:37:19.000Z

That's great.
Thanks for informing me.

Answer 16 · 2022-07-29T12:16:31.000Z

The issue seems to have been solved.