흐음.. classification과 object-detection의 차이를 모르겠네..

아무튼 classification 쪽에 만들어진 model이 꽤 있으니 이걸 이용하면 될 듯.

 

[링크 : https://tfhub.dev/s?deployment-format=lite&module-type=image-classification]

[링크 : https://tfhub.dev/s?deployment-format=lite&module-type=image-object-detection]

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

tensorflow 객체 탐지 예제  (0) 2021.01.21
tensorflow lite on x86 / label_image  (0) 2021.01.20
tensorflow-lite minimal.cc 실행  (0) 2021.01.18
mobileNET/SSD  (0) 2021.01.14
caffe  (0) 2021.01.14
Posted by 구차니

+ 2021.01.19

./tensorflow/lite/tools/make/gen/rpi_armv7l/bin 에 빌드된 파일이 존재한다 -_-

왜 이 고생을 했지? ㅠㅠㅠㅠ

./tensorflow/lite/tools/make/build_rpi_lib.sh clean # clean object files
./tensorflow/lite/tools/make/build_rpi_lib.sh -j 16 # run with 16 jobs to leverage more CPU cores
./tensorflow/lite/tools/make/build_rpi_lib.sh label_image # # build label_image binary

[링크 : https://www.tensorflow.org/lite/guide/build_rpi]

 

---

어찌어찌 겨우겨우 libtensorflow-lite.a 를 생성해서

minimal.cc를 빌드해서 돌려보는데 감동...?까진 아니고

아무튼 텐서플로우 lite 자체는 static 한데, minimal.cc는 static 할 수 없는건 무슨 묘미인가...

$ ldd minimal
        linux-vdso.so.1 (0x7efc0000)
        /usr/lib/arm-linux-gnueabihf/libarmmem-${PLATFORM}.so => /usr/lib/arm-linux-gnueabihf/libarmmem-v7l.so (0x76eef000)
        libpthread.so.0 => /lib/arm-linux-gnueabihf/libpthread.so.0 (0x76ea4000)
        libdl.so.2 => /lib/arm-linux-gnueabihf/libdl.so.2 (0x76e91000)
        libstdc++.so.6 => /lib/arm-linux-gnueabihf/libstdc++.so.6 (0x76d4a000)
        libm.so.6 => /lib/arm-linux-gnueabihf/libm.so.6 (0x76cc8000)
        libgcc_s.so.1 => /lib/arm-linux-gnueabihf/libgcc_s.so.1 (0x76c9b000)
        libc.so.6 => /lib/arm-linux-gnueabihf/libc.so.6 (0x76b4d000)
        /lib/ld-linux-armhf.so.3 (0x76f04000)

 

~/src/tensorflow_src/tensorflow/lite/examples/minimal $ ./minimal 1.tflite
=== Pre-invoke Interpreter State ===
Interpreter has 184 tensors and 64 nodes
Inputs: 175
Outputs: 167 168 169 170

Tensor   0 BoxPredictor_0/BoxEncodingPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw       4332 bytes ( 0.0 MB)  1 19 19 12
Tensor   1 BoxPredictor_0/BoxEncodingPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo         48 bytes ( 0.0 MB)  12
Tensor   2 BoxPredictor_0/BoxEncodingPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       4608 bytes ( 0.0 MB)  12 1 1 384
Tensor   3 BoxPredictor_0/ClassPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw      98553 bytes ( 0.1 MB)  1 19 19 273
Tensor   4 BoxPredictor_0/ClassPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo       1092 bytes ( 0.0 MB)  273
Tensor   5 BoxPredictor_0/ClassPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     104832 bytes ( 0.1 MB)  273 1 1 384
Tensor   6 BoxPredictor_0/Reshape kTfLiteUInt8  kTfLiteArenaRw       4332 bytes ( 0.0 MB)  1 1083 1 4
Tensor   7 BoxPredictor_0/Reshape_1 kTfLiteUInt8  kTfLiteArenaRw      98553 bytes ( 0.1 MB)  1 1083 91
Tensor   8 BoxPredictor_0/stack kTfLiteInt32   kTfLiteMmapRo         16 bytes ( 0.0 MB)  4
Tensor   9 BoxPredictor_0/stack_1 kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor  10 BoxPredictor_1/BoxEncodingPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw       2400 bytes ( 0.0 MB)  1 10 10 24
Tensor  11 BoxPredictor_1/BoxEncodingPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor  12 BoxPredictor_1/BoxEncodingPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      18432 bytes ( 0.0 MB)  24 1 1 768
Tensor  13 BoxPredictor_1/ClassPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw      54600 bytes ( 0.1 MB)  1 10 10 546
Tensor  14 BoxPredictor_1/ClassPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo       2184 bytes ( 0.0 MB)  546
Tensor  15 BoxPredictor_1/ClassPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     419328 bytes ( 0.4 MB)  546 1 1 768
Tensor  16 BoxPredictor_1/Reshape kTfLiteUInt8  kTfLiteArenaRw       2400 bytes ( 0.0 MB)  1 600 1 4
Tensor  17 BoxPredictor_1/Reshape_1 kTfLiteUInt8  kTfLiteArenaRw      54600 bytes ( 0.1 MB)  1 600 91
Tensor  18 BoxPredictor_1/stack kTfLiteInt32   kTfLiteMmapRo         16 bytes ( 0.0 MB)  4
Tensor  19 BoxPredictor_1/stack_1 kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor  20 BoxPredictor_2/BoxEncodingPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw        600 bytes ( 0.0 MB)  1 5 5 24
Tensor  21 BoxPredictor_2/BoxEncodingPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor  22 BoxPredictor_2/BoxEncodingPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       9216 bytes ( 0.0 MB)  24 1 1 384
Tensor  23 BoxPredictor_2/ClassPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw      13650 bytes ( 0.0 MB)  1 5 5 546
Tensor  24 BoxPredictor_2/ClassPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo       2184 bytes ( 0.0 MB)  546
Tensor  25 BoxPredictor_2/ClassPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     209664 bytes ( 0.2 MB)  546 1 1 384
Tensor  26 BoxPredictor_2/Reshape kTfLiteUInt8  kTfLiteArenaRw        600 bytes ( 0.0 MB)  1 150 1 4
Tensor  27 BoxPredictor_2/Reshape_1 kTfLiteUInt8  kTfLiteArenaRw      13650 bytes ( 0.0 MB)  1 150 91
Tensor  28 BoxPredictor_2/stack kTfLiteInt32   kTfLiteMmapRo         16 bytes ( 0.0 MB)  4
Tensor  29 BoxPredictor_2/stack_1 kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor  30 BoxPredictor_3/BoxEncodingPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw        216 bytes ( 0.0 MB)  1 3 3 24
Tensor  31 BoxPredictor_3/BoxEncodingPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor  32 BoxPredictor_3/BoxEncodingPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       4608 bytes ( 0.0 MB)  24 1 1 192
Tensor  33 BoxPredictor_3/ClassPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw       4914 bytes ( 0.0 MB)  1 3 3 546
Tensor  34 BoxPredictor_3/ClassPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo       2184 bytes ( 0.0 MB)  546
Tensor  35 BoxPredictor_3/ClassPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     104832 bytes ( 0.1 MB)  546 1 1 192
Tensor  36 BoxPredictor_3/Reshape kTfLiteUInt8  kTfLiteArenaRw        216 bytes ( 0.0 MB)  1 54 1 4
Tensor  37 BoxPredictor_3/Reshape_1 kTfLiteUInt8  kTfLiteArenaRw       4914 bytes ( 0.0 MB)  1 54 91
Tensor  38 BoxPredictor_3/stack kTfLiteInt32   kTfLiteMmapRo         16 bytes ( 0.0 MB)  4
Tensor  39 BoxPredictor_3/stack_1 kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor  40 BoxPredictor_4/BoxEncodingPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw         96 bytes ( 0.0 MB)  1 2 2 24
Tensor  41 BoxPredictor_4/BoxEncodingPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor  42 BoxPredictor_4/BoxEncodingPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       4608 bytes ( 0.0 MB)  24 1 1 192
Tensor  43 BoxPredictor_4/ClassPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw       2184 bytes ( 0.0 MB)  1 2 2 546
Tensor  44 BoxPredictor_4/ClassPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo       2184 bytes ( 0.0 MB)  546
Tensor  45 BoxPredictor_4/ClassPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     104832 bytes ( 0.1 MB)  546 1 1 192
Tensor  46 BoxPredictor_4/Reshape kTfLiteUInt8  kTfLiteArenaRw         96 bytes ( 0.0 MB)  1 24 1 4
Tensor  47 BoxPredictor_4/Reshape_1 kTfLiteUInt8  kTfLiteArenaRw       2184 bytes ( 0.0 MB)  1 24 91
Tensor  48 BoxPredictor_4/stack kTfLiteInt32   kTfLiteMmapRo         16 bytes ( 0.0 MB)  4
Tensor  49 BoxPredictor_4/stack_1 kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor  50 BoxPredictor_5/BoxEncodingPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw         24 bytes ( 0.0 MB)  1 1 1 24
Tensor  51 BoxPredictor_5/BoxEncodingPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor  52 BoxPredictor_5/BoxEncodingPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       2304 bytes ( 0.0 MB)  24 1 1 96
Tensor  53 BoxPredictor_5/ClassPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw        546 bytes ( 0.0 MB)  1 1 1 546
Tensor  54 BoxPredictor_5/ClassPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo       2184 bytes ( 0.0 MB)  546
Tensor  55 BoxPredictor_5/ClassPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      52416 bytes ( 0.0 MB)  546 1 1 96
Tensor  56 BoxPredictor_5/Reshape kTfLiteUInt8  kTfLiteArenaRw         24 bytes ( 0.0 MB)  1 6 1 4
Tensor  57 BoxPredictor_5/Reshape_1 kTfLiteUInt8  kTfLiteArenaRw        546 bytes ( 0.0 MB)  1 6 91
Tensor  58 BoxPredictor_5/stack kTfLiteInt32   kTfLiteMmapRo         16 bytes ( 0.0 MB)  4
Tensor  59 BoxPredictor_5/stack_1 kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor  60 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_2_1x1_192/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor  61 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_2_1x1_192/Relu6 kTfLiteUInt8  kTfLiteArenaRw      19200 bytes ( 0.0 MB)  1 10 10 192
Tensor  62 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_2_1x1_192/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     147456 bytes ( 0.1 MB)  192 1 1 768
Tensor  63 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_3_1x1_96/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor  64 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_3_1x1_96/Relu6 kTfLiteUInt8  kTfLiteArenaRw       2400 bytes ( 0.0 MB)  1 5 5 96
Tensor  65 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_3_1x1_96/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      36864 bytes ( 0.0 MB)  96 1 1 384
Tensor  66 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_4_1x1_96/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor  67 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_4_1x1_96/Relu6 kTfLiteUInt8  kTfLiteArenaRw        864 bytes ( 0.0 MB)  1 3 3 96
Tensor  68 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_4_1x1_96/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      18432 bytes ( 0.0 MB)  96 1 1 192
Tensor  69 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_5_1x1_48/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        192 bytes ( 0.0 MB)  48
Tensor  70 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_5_1x1_48/Relu6 kTfLiteUInt8  kTfLiteArenaRw        192 bytes ( 0.0 MB)  1 2 2 48
Tensor  71 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_5_1x1_48/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       9216 bytes ( 0.0 MB)  48 1 1 192
Tensor  72 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_2_3x3_s2_384/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor  73 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_2_3x3_s2_384/Relu6 kTfLiteUInt8  kTfLiteArenaRw       9600 bytes ( 0.0 MB)  1 5 5 384
Tensor  74 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_2_3x3_s2_384/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     663552 bytes ( 0.6 MB)  384 3 3 192
Tensor  75 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_3_3x3_s2_192/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor  76 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_3_3x3_s2_192/Relu6 kTfLiteUInt8  kTfLiteArenaRw       1728 bytes ( 0.0 MB)  1 3 3 192
Tensor  77 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_3_3x3_s2_192/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     165888 bytes ( 0.2 MB)  192 3 3 96
Tensor  78 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_4_3x3_s2_192/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor  79 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_4_3x3_s2_192/Relu6 kTfLiteUInt8  kTfLiteArenaRw        768 bytes ( 0.0 MB)  1 2 2 192
Tensor  80 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_4_3x3_s2_192/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     165888 bytes ( 0.2 MB)  192 3 3 96
Tensor  81 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_5_3x3_s2_96/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor  82 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_5_3x3_s2_96/Relu6 kTfLiteUInt8  kTfLiteArenaRw         96 bytes ( 0.0 MB)  1 1 1 96
Tensor  83 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_5_3x3_s2_96/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      41472 bytes ( 0.0 MB)  96 3 3 48
Tensor  84 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_0/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor  85 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_0/Relu6 kTfLiteUInt8  kTfLiteArenaRw     540000 bytes ( 0.5 MB)  1 150 150 24
Tensor  86 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_0/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo        648 bytes ( 0.0 MB)  24 3 3 3
Tensor  87 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_10_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor  88 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_10_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor  89 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_10_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       3456 bytes ( 0.0 MB)  1 3 3 384
Tensor  90 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_10_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor  91 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_10_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor  92 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_10_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     147456 bytes ( 0.1 MB)  384 1 1 384
Tensor  93 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_11_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor  94 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_11_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor  95 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_11_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       3456 bytes ( 0.0 MB)  1 3 3 384
Tensor  96 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_11_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor  97 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_11_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor  98 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_11_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     147456 bytes ( 0.1 MB)  384 1 1 384
Tensor  99 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_12_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw      38400 bytes ( 0.0 MB)  1 10 10 384
Tensor 100 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_12_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 101 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_12_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       3456 bytes ( 0.0 MB)  1 3 3 384
Tensor 102 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_12_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       3072 bytes ( 0.0 MB)  768
Tensor 103 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_12_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw      76800 bytes ( 0.1 MB)  1 10 10 768
Tensor 104 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_12_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     294912 bytes ( 0.3 MB)  768 1 1 384
Tensor 105 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_13_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw      76800 bytes ( 0.1 MB)  1 10 10 768
Tensor 106 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_13_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       3072 bytes ( 0.0 MB)  768
Tensor 107 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_13_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       6912 bytes ( 0.0 MB)  1 3 3 768
Tensor 108 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_13_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       3072 bytes ( 0.0 MB)  768
Tensor 109 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_13_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw      76800 bytes ( 0.1 MB)  1 10 10 768
Tensor 110 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_13_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     589824 bytes ( 0.6 MB)  768 1 1 768
Tensor 111 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_1_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     540000 bytes ( 0.5 MB)  1 150 150 24
Tensor 112 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_1_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor 113 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_1_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo        216 bytes ( 0.0 MB)  1 3 3 24
Tensor 114 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_1_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        192 bytes ( 0.0 MB)  48
Tensor 115 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_1_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw    1080000 bytes ( 1.0 MB)  1 150 150 48
Tensor 116 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_1_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       1152 bytes ( 0.0 MB)  48 1 1 24
Tensor 117 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_2_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     270000 bytes ( 0.3 MB)  1 75 75 48
Tensor 118 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_2_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo        192 bytes ( 0.0 MB)  48
Tensor 119 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_2_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo        432 bytes ( 0.0 MB)  1 3 3 48
Tensor 120 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_2_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor 121 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_2_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     540000 bytes ( 0.5 MB)  1 75 75 96
Tensor 122 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_2_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       4608 bytes ( 0.0 MB)  96 1 1 48
Tensor 123 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_3_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     540000 bytes ( 0.5 MB)  1 75 75 96
Tensor 124 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_3_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor 125 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_3_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo        864 bytes ( 0.0 MB)  1 3 3 96
Tensor 126 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_3_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor 127 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_3_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     540000 bytes ( 0.5 MB)  1 75 75 96
Tensor 128 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_3_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       9216 bytes ( 0.0 MB)  96 1 1 96
Tensor 129 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_4_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 38 38 96
Tensor 130 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_4_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor 131 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_4_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo        864 bytes ( 0.0 MB)  1 3 3 96
Tensor 132 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_4_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor 133 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_4_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     277248 bytes ( 0.3 MB)  1 38 38 192
Tensor 134 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_4_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      18432 bytes ( 0.0 MB)  192 1 1 96
Tensor 135 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_5_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     277248 bytes ( 0.3 MB)  1 38 38 192
Tensor 136 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_5_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor 137 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_5_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       1728 bytes ( 0.0 MB)  1 3 3 192
Tensor 138 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_5_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor 139 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_5_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     277248 bytes ( 0.3 MB)  1 38 38 192
Tensor 140 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_5_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      36864 bytes ( 0.0 MB)  192 1 1 192
Tensor 141 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_6_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw      69312 bytes ( 0.1 MB)  1 19 19 192
Tensor 142 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_6_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor 143 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_6_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       1728 bytes ( 0.0 MB)  1 3 3 192
Tensor 144 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_6_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 145 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_6_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 146 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_6_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      73728 bytes ( 0.1 MB)  384 1 1 192
Tensor 147 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_7_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 148 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_7_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 149 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_7_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       3456 bytes ( 0.0 MB)  1 3 3 384
Tensor 150 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_7_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 151 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_7_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 152 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_7_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     147456 bytes ( 0.1 MB)  384 1 1 384
Tensor 153 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_8_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 154 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_8_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 155 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_8_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       3456 bytes ( 0.0 MB)  1 3 3 384
Tensor 156 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_8_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 157 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_8_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 158 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_8_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     147456 bytes ( 0.1 MB)  384 1 1 384
Tensor 159 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_9_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 160 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_9_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 161 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_9_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       3456 bytes ( 0.0 MB)  1 3 3 384
Tensor 162 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_9_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 163 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_9_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 164 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_9_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     147456 bytes ( 0.1 MB)  384 1 1 384
Tensor 165 Squeeze              kTfLiteUInt8  kTfLiteArenaRw       7668 bytes ( 0.0 MB)  1 1917 4
Tensor 166 Squeeze_shape        kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor 167 TFLite_Detection_PostProcess kTfLiteFloat32  kTfLiteArenaRw        160 bytes ( 0.0 MB)  1 10 4
Tensor 168 TFLite_Detection_PostProcess:1 kTfLiteFloat32  kTfLiteArenaRw         40 bytes ( 0.0 MB)  1 10
Tensor 169 TFLite_Detection_PostProcess:2 kTfLiteFloat32  kTfLiteArenaRw         40 bytes ( 0.0 MB)  1 10
Tensor 170 TFLite_Detection_PostProcess:3 kTfLiteFloat32  kTfLiteArenaRw          4 bytes ( 0.0 MB)  1
Tensor 171 anchors              kTfLiteUInt8   kTfLiteMmapRo       7668 bytes ( 0.0 MB)  1917 4
Tensor 172 concat               kTfLiteUInt8  kTfLiteArenaRw       7668 bytes ( 0.0 MB)  1 1917 1 4
Tensor 173 concat_1             kTfLiteUInt8  kTfLiteArenaRw     174447 bytes ( 0.2 MB)  1 1917 91
Tensor 174 convert_scores       kTfLiteUInt8  kTfLiteArenaRw     174447 bytes ( 0.2 MB)  1 1917 91
Tensor 175 normalized_input_image_tensor kTfLiteUInt8  kTfLiteArenaRw     270000 bytes ( 0.3 MB)  1 300 300 3
Tensor 176 (null)               kTfLiteFloat32  kTfLiteArenaRw      30672 bytes ( 0.0 MB)  1917 4
Tensor 177 (null)               kTfLiteFloat32  kTfLiteArenaRw     697788 bytes ( 0.7 MB)  1917 91
Tensor 178 (null)               kTfLiteUInt8  kTfLiteArenaRw       1917 bytes ( 0.0 MB)  1917
Tensor 179 (null)               kTfLiteUInt8  kTfLiteArenaRw     607500 bytes ( 0.6 MB)  1 150 150 27
Tensor 180 (null)               kTfLiteUInt8  kTfLiteArenaRw      43200 bytes ( 0.0 MB)  1 5 5 1728
Tensor 181 (null)               kTfLiteUInt8  kTfLiteArenaRw       7776 bytes ( 0.0 MB)  1 3 3 864
Tensor 182 (null)               kTfLiteUInt8  kTfLiteArenaRw       3456 bytes ( 0.0 MB)  1 2 2 864
Tensor 183 (null)               kTfLiteUInt8  kTfLiteArenaRw        432 bytes ( 0.0 MB)  1 1 1 432

Node   0 Operator Builtin Code   3 CONV_2D
  Inputs: 175 86 84
  Outputs: 85
  Temporaries: 179
Node   1 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 85 113 112
  Outputs: 111
Node   2 Operator Builtin Code   3 CONV_2D
  Inputs: 111 116 114
  Outputs: 115
Node   3 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 115 119 118
  Outputs: 117
Node   4 Operator Builtin Code   3 CONV_2D
  Inputs: 117 122 120
  Outputs: 121
Node   5 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 121 125 124
  Outputs: 123
Node   6 Operator Builtin Code   3 CONV_2D
  Inputs: 123 128 126
  Outputs: 127
Node   7 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 127 131 130
  Outputs: 129
Node   8 Operator Builtin Code   3 CONV_2D
  Inputs: 129 134 132
  Outputs: 133
Node   9 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 133 137 136
  Outputs: 135
Node  10 Operator Builtin Code   3 CONV_2D
  Inputs: 135 140 138
  Outputs: 139
Node  11 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 139 143 142
  Outputs: 141
Node  12 Operator Builtin Code   3 CONV_2D
  Inputs: 141 146 144
  Outputs: 145
Node  13 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 145 149 148
  Outputs: 147
Node  14 Operator Builtin Code   3 CONV_2D
  Inputs: 147 152 150
  Outputs: 151
Node  15 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 151 155 154
  Outputs: 153
Node  16 Operator Builtin Code   3 CONV_2D
  Inputs: 153 158 156
  Outputs: 157
Node  17 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 157 161 160
  Outputs: 159
Node  18 Operator Builtin Code   3 CONV_2D
  Inputs: 159 164 162
  Outputs: 163
Node  19 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 163 89 88
  Outputs: 87
Node  20 Operator Builtin Code   3 CONV_2D
  Inputs: 87 92 90
  Outputs: 91
Node  21 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 91 95 94
  Outputs: 93
Node  22 Operator Builtin Code   3 CONV_2D
  Inputs: 93 98 96
  Outputs: 97
Node  23 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 97 101 100
  Outputs: 99
Node  24 Operator Builtin Code   3 CONV_2D
  Inputs: 99 104 102
  Outputs: 103
Node  25 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 103 107 106
  Outputs: 105
Node  26 Operator Builtin Code   3 CONV_2D
  Inputs: 105 110 108
  Outputs: 109
Node  27 Operator Builtin Code   3 CONV_2D
  Inputs: 109 62 60
  Outputs: 61
Node  28 Operator Builtin Code   3 CONV_2D
  Inputs: 61 74 72
  Outputs: 73
  Temporaries: 180
Node  29 Operator Builtin Code   3 CONV_2D
  Inputs: 73 65 63
  Outputs: 64
Node  30 Operator Builtin Code   3 CONV_2D
  Inputs: 64 77 75
  Outputs: 76
  Temporaries: 181
Node  31 Operator Builtin Code   3 CONV_2D
  Inputs: 76 68 66
  Outputs: 67
Node  32 Operator Builtin Code   3 CONV_2D
  Inputs: 67 80 78
  Outputs: 79
  Temporaries: 182
Node  33 Operator Builtin Code   3 CONV_2D
  Inputs: 79 71 69
  Outputs: 70
Node  34 Operator Builtin Code   3 CONV_2D
  Inputs: 70 83 81
  Outputs: 82
  Temporaries: 183
Node  35 Operator Builtin Code   3 CONV_2D
  Inputs: 97 2 1
  Outputs: 0
Node  36 Operator Builtin Code  22 RESHAPE
  Inputs: 0 8
  Outputs: 6
Node  37 Operator Builtin Code   3 CONV_2D
  Inputs: 97 5 4
  Outputs: 3
Node  38 Operator Builtin Code  22 RESHAPE
  Inputs: 3 9
  Outputs: 7
Node  39 Operator Builtin Code   3 CONV_2D
  Inputs: 109 12 11
  Outputs: 10
Node  40 Operator Builtin Code  22 RESHAPE
  Inputs: 10 18
  Outputs: 16
Node  41 Operator Builtin Code   3 CONV_2D
  Inputs: 109 15 14
  Outputs: 13
Node  42 Operator Builtin Code  22 RESHAPE
  Inputs: 13 19
  Outputs: 17
Node  43 Operator Builtin Code   3 CONV_2D
  Inputs: 73 22 21
  Outputs: 20
Node  44 Operator Builtin Code  22 RESHAPE
  Inputs: 20 28
  Outputs: 26
Node  45 Operator Builtin Code   3 CONV_2D
  Inputs: 73 25 24
  Outputs: 23
Node  46 Operator Builtin Code  22 RESHAPE
  Inputs: 23 29
  Outputs: 27
Node  47 Operator Builtin Code   3 CONV_2D
  Inputs: 76 32 31
  Outputs: 30
Node  48 Operator Builtin Code  22 RESHAPE
  Inputs: 30 38
  Outputs: 36
Node  49 Operator Builtin Code   3 CONV_2D
  Inputs: 76 35 34
  Outputs: 33
Node  50 Operator Builtin Code  22 RESHAPE
  Inputs: 33 39
  Outputs: 37
Node  51 Operator Builtin Code   3 CONV_2D
  Inputs: 79 42 41
  Outputs: 40
Node  52 Operator Builtin Code  22 RESHAPE
  Inputs: 40 48
  Outputs: 46
Node  53 Operator Builtin Code   3 CONV_2D
  Inputs: 79 45 44
  Outputs: 43
Node  54 Operator Builtin Code  22 RESHAPE
  Inputs: 43 49
  Outputs: 47
Node  55 Operator Builtin Code   3 CONV_2D
  Inputs: 82 52 51
  Outputs: 50
Node  56 Operator Builtin Code  22 RESHAPE
  Inputs: 50 58
  Outputs: 56
Node  57 Operator Builtin Code   2 CONCATENATION
  Inputs: 6 16 26 36 46 56
  Outputs: 172
Node  58 Operator Builtin Code  22 RESHAPE
  Inputs: 172 166
  Outputs: 165
Node  59 Operator Builtin Code   3 CONV_2D
  Inputs: 82 55 54
  Outputs: 53
Node  60 Operator Builtin Code  22 RESHAPE
  Inputs: 53 59
  Outputs: 57
Node  61 Operator Builtin Code   2 CONCATENATION
  Inputs: 7 17 27 37 47 57
  Outputs: 173
Node  62 Operator Builtin Code  14 LOGISTIC
  Inputs: 173
  Outputs: 174
Node  63 Operator Custom Name TFLite_Detection_PostProcess
  Inputs: 165 174 171
  Outputs: 167 168 169 170
  Temporaries: 176 177 178


=== Post-invoke Interpreter State ===
Interpreter has 184 tensors and 64 nodes
Inputs: 175
Outputs: 167 168 169 170

Tensor   0 BoxPredictor_0/BoxEncodingPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw       4332 bytes ( 0.0 MB)  1 19 19 12
Tensor   1 BoxPredictor_0/BoxEncodingPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo         48 bytes ( 0.0 MB)  12
Tensor   2 BoxPredictor_0/BoxEncodingPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       4608 bytes ( 0.0 MB)  12 1 1 384
Tensor   3 BoxPredictor_0/ClassPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw      98553 bytes ( 0.1 MB)  1 19 19 273
Tensor   4 BoxPredictor_0/ClassPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo       1092 bytes ( 0.0 MB)  273
Tensor   5 BoxPredictor_0/ClassPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     104832 bytes ( 0.1 MB)  273 1 1 384
Tensor   6 BoxPredictor_0/Reshape kTfLiteUInt8  kTfLiteArenaRw       4332 bytes ( 0.0 MB)  1 1083 1 4
Tensor   7 BoxPredictor_0/Reshape_1 kTfLiteUInt8  kTfLiteArenaRw      98553 bytes ( 0.1 MB)  1 1083 91
Tensor   8 BoxPredictor_0/stack kTfLiteInt32   kTfLiteMmapRo         16 bytes ( 0.0 MB)  4
Tensor   9 BoxPredictor_0/stack_1 kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor  10 BoxPredictor_1/BoxEncodingPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw       2400 bytes ( 0.0 MB)  1 10 10 24
Tensor  11 BoxPredictor_1/BoxEncodingPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor  12 BoxPredictor_1/BoxEncodingPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      18432 bytes ( 0.0 MB)  24 1 1 768
Tensor  13 BoxPredictor_1/ClassPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw      54600 bytes ( 0.1 MB)  1 10 10 546
Tensor  14 BoxPredictor_1/ClassPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo       2184 bytes ( 0.0 MB)  546
Tensor  15 BoxPredictor_1/ClassPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     419328 bytes ( 0.4 MB)  546 1 1 768
Tensor  16 BoxPredictor_1/Reshape kTfLiteUInt8  kTfLiteArenaRw       2400 bytes ( 0.0 MB)  1 600 1 4
Tensor  17 BoxPredictor_1/Reshape_1 kTfLiteUInt8  kTfLiteArenaRw      54600 bytes ( 0.1 MB)  1 600 91
Tensor  18 BoxPredictor_1/stack kTfLiteInt32   kTfLiteMmapRo         16 bytes ( 0.0 MB)  4
Tensor  19 BoxPredictor_1/stack_1 kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor  20 BoxPredictor_2/BoxEncodingPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw        600 bytes ( 0.0 MB)  1 5 5 24
Tensor  21 BoxPredictor_2/BoxEncodingPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor  22 BoxPredictor_2/BoxEncodingPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       9216 bytes ( 0.0 MB)  24 1 1 384
Tensor  23 BoxPredictor_2/ClassPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw      13650 bytes ( 0.0 MB)  1 5 5 546
Tensor  24 BoxPredictor_2/ClassPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo       2184 bytes ( 0.0 MB)  546
Tensor  25 BoxPredictor_2/ClassPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     209664 bytes ( 0.2 MB)  546 1 1 384
Tensor  26 BoxPredictor_2/Reshape kTfLiteUInt8  kTfLiteArenaRw        600 bytes ( 0.0 MB)  1 150 1 4
Tensor  27 BoxPredictor_2/Reshape_1 kTfLiteUInt8  kTfLiteArenaRw      13650 bytes ( 0.0 MB)  1 150 91
Tensor  28 BoxPredictor_2/stack kTfLiteInt32   kTfLiteMmapRo         16 bytes ( 0.0 MB)  4
Tensor  29 BoxPredictor_2/stack_1 kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor  30 BoxPredictor_3/BoxEncodingPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw        216 bytes ( 0.0 MB)  1 3 3 24
Tensor  31 BoxPredictor_3/BoxEncodingPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor  32 BoxPredictor_3/BoxEncodingPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       4608 bytes ( 0.0 MB)  24 1 1 192
Tensor  33 BoxPredictor_3/ClassPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw       4914 bytes ( 0.0 MB)  1 3 3 546
Tensor  34 BoxPredictor_3/ClassPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo       2184 bytes ( 0.0 MB)  546
Tensor  35 BoxPredictor_3/ClassPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     104832 bytes ( 0.1 MB)  546 1 1 192
Tensor  36 BoxPredictor_3/Reshape kTfLiteUInt8  kTfLiteArenaRw        216 bytes ( 0.0 MB)  1 54 1 4
Tensor  37 BoxPredictor_3/Reshape_1 kTfLiteUInt8  kTfLiteArenaRw       4914 bytes ( 0.0 MB)  1 54 91
Tensor  38 BoxPredictor_3/stack kTfLiteInt32   kTfLiteMmapRo         16 bytes ( 0.0 MB)  4
Tensor  39 BoxPredictor_3/stack_1 kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor  40 BoxPredictor_4/BoxEncodingPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw         96 bytes ( 0.0 MB)  1 2 2 24
Tensor  41 BoxPredictor_4/BoxEncodingPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor  42 BoxPredictor_4/BoxEncodingPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       4608 bytes ( 0.0 MB)  24 1 1 192
Tensor  43 BoxPredictor_4/ClassPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw       2184 bytes ( 0.0 MB)  1 2 2 546
Tensor  44 BoxPredictor_4/ClassPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo       2184 bytes ( 0.0 MB)  546
Tensor  45 BoxPredictor_4/ClassPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     104832 bytes ( 0.1 MB)  546 1 1 192
Tensor  46 BoxPredictor_4/Reshape kTfLiteUInt8  kTfLiteArenaRw         96 bytes ( 0.0 MB)  1 24 1 4
Tensor  47 BoxPredictor_4/Reshape_1 kTfLiteUInt8  kTfLiteArenaRw       2184 bytes ( 0.0 MB)  1 24 91
Tensor  48 BoxPredictor_4/stack kTfLiteInt32   kTfLiteMmapRo         16 bytes ( 0.0 MB)  4
Tensor  49 BoxPredictor_4/stack_1 kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor  50 BoxPredictor_5/BoxEncodingPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw         24 bytes ( 0.0 MB)  1 1 1 24
Tensor  51 BoxPredictor_5/BoxEncodingPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor  52 BoxPredictor_5/BoxEncodingPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       2304 bytes ( 0.0 MB)  24 1 1 96
Tensor  53 BoxPredictor_5/ClassPredictor/BiasAdd kTfLiteUInt8  kTfLiteArenaRw        546 bytes ( 0.0 MB)  1 1 1 546
Tensor  54 BoxPredictor_5/ClassPredictor/Conv2D_bias kTfLiteInt32   kTfLiteMmapRo       2184 bytes ( 0.0 MB)  546
Tensor  55 BoxPredictor_5/ClassPredictor/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      52416 bytes ( 0.0 MB)  546 1 1 96
Tensor  56 BoxPredictor_5/Reshape kTfLiteUInt8  kTfLiteArenaRw         24 bytes ( 0.0 MB)  1 6 1 4
Tensor  57 BoxPredictor_5/Reshape_1 kTfLiteUInt8  kTfLiteArenaRw        546 bytes ( 0.0 MB)  1 6 91
Tensor  58 BoxPredictor_5/stack kTfLiteInt32   kTfLiteMmapRo         16 bytes ( 0.0 MB)  4
Tensor  59 BoxPredictor_5/stack_1 kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor  60 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_2_1x1_192/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor  61 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_2_1x1_192/Relu6 kTfLiteUInt8  kTfLiteArenaRw      19200 bytes ( 0.0 MB)  1 10 10 192
Tensor  62 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_2_1x1_192/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     147456 bytes ( 0.1 MB)  192 1 1 768
Tensor  63 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_3_1x1_96/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor  64 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_3_1x1_96/Relu6 kTfLiteUInt8  kTfLiteArenaRw       2400 bytes ( 0.0 MB)  1 5 5 96
Tensor  65 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_3_1x1_96/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      36864 bytes ( 0.0 MB)  96 1 1 384
Tensor  66 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_4_1x1_96/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor  67 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_4_1x1_96/Relu6 kTfLiteUInt8  kTfLiteArenaRw        864 bytes ( 0.0 MB)  1 3 3 96
Tensor  68 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_4_1x1_96/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      18432 bytes ( 0.0 MB)  96 1 1 192
Tensor  69 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_5_1x1_48/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        192 bytes ( 0.0 MB)  48
Tensor  70 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_5_1x1_48/Relu6 kTfLiteUInt8  kTfLiteArenaRw        192 bytes ( 0.0 MB)  1 2 2 48
Tensor  71 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_1_Conv2d_5_1x1_48/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       9216 bytes ( 0.0 MB)  48 1 1 192
Tensor  72 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_2_3x3_s2_384/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor  73 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_2_3x3_s2_384/Relu6 kTfLiteUInt8  kTfLiteArenaRw       9600 bytes ( 0.0 MB)  1 5 5 384
Tensor  74 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_2_3x3_s2_384/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     663552 bytes ( 0.6 MB)  384 3 3 192
Tensor  75 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_3_3x3_s2_192/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor  76 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_3_3x3_s2_192/Relu6 kTfLiteUInt8  kTfLiteArenaRw       1728 bytes ( 0.0 MB)  1 3 3 192
Tensor  77 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_3_3x3_s2_192/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     165888 bytes ( 0.2 MB)  192 3 3 96
Tensor  78 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_4_3x3_s2_192/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor  79 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_4_3x3_s2_192/Relu6 kTfLiteUInt8  kTfLiteArenaRw        768 bytes ( 0.0 MB)  1 2 2 192
Tensor  80 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_4_3x3_s2_192/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     165888 bytes ( 0.2 MB)  192 3 3 96
Tensor  81 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_5_3x3_s2_96/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor  82 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_5_3x3_s2_96/Relu6 kTfLiteUInt8  kTfLiteArenaRw         96 bytes ( 0.0 MB)  1 1 1 96
Tensor  83 FeatureExtractor/MobilenetV1/Conv2d_13_pointwise_2_Conv2d_5_3x3_s2_96/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      41472 bytes ( 0.0 MB)  96 3 3 48
Tensor  84 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_0/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor  85 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_0/Relu6 kTfLiteUInt8  kTfLiteArenaRw     540000 bytes ( 0.5 MB)  1 150 150 24
Tensor  86 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_0/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo        648 bytes ( 0.0 MB)  24 3 3 3
Tensor  87 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_10_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor  88 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_10_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor  89 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_10_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       3456 bytes ( 0.0 MB)  1 3 3 384
Tensor  90 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_10_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor  91 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_10_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor  92 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_10_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     147456 bytes ( 0.1 MB)  384 1 1 384
Tensor  93 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_11_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor  94 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_11_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor  95 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_11_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       3456 bytes ( 0.0 MB)  1 3 3 384
Tensor  96 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_11_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor  97 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_11_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor  98 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_11_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     147456 bytes ( 0.1 MB)  384 1 1 384
Tensor  99 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_12_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw      38400 bytes ( 0.0 MB)  1 10 10 384
Tensor 100 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_12_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 101 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_12_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       3456 bytes ( 0.0 MB)  1 3 3 384
Tensor 102 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_12_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       3072 bytes ( 0.0 MB)  768
Tensor 103 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_12_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw      76800 bytes ( 0.1 MB)  1 10 10 768
Tensor 104 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_12_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     294912 bytes ( 0.3 MB)  768 1 1 384
Tensor 105 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_13_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw      76800 bytes ( 0.1 MB)  1 10 10 768
Tensor 106 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_13_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       3072 bytes ( 0.0 MB)  768
Tensor 107 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_13_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       6912 bytes ( 0.0 MB)  1 3 3 768
Tensor 108 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_13_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       3072 bytes ( 0.0 MB)  768
Tensor 109 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_13_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw      76800 bytes ( 0.1 MB)  1 10 10 768
Tensor 110 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_13_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     589824 bytes ( 0.6 MB)  768 1 1 768
Tensor 111 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_1_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     540000 bytes ( 0.5 MB)  1 150 150 24
Tensor 112 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_1_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo         96 bytes ( 0.0 MB)  24
Tensor 113 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_1_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo        216 bytes ( 0.0 MB)  1 3 3 24
Tensor 114 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_1_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        192 bytes ( 0.0 MB)  48
Tensor 115 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_1_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw    1080000 bytes ( 1.0 MB)  1 150 150 48
Tensor 116 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_1_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       1152 bytes ( 0.0 MB)  48 1 1 24
Tensor 117 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_2_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     270000 bytes ( 0.3 MB)  1 75 75 48
Tensor 118 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_2_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo        192 bytes ( 0.0 MB)  48
Tensor 119 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_2_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo        432 bytes ( 0.0 MB)  1 3 3 48
Tensor 120 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_2_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor 121 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_2_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     540000 bytes ( 0.5 MB)  1 75 75 96
Tensor 122 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_2_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       4608 bytes ( 0.0 MB)  96 1 1 48
Tensor 123 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_3_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     540000 bytes ( 0.5 MB)  1 75 75 96
Tensor 124 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_3_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor 125 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_3_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo        864 bytes ( 0.0 MB)  1 3 3 96
Tensor 126 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_3_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor 127 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_3_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     540000 bytes ( 0.5 MB)  1 75 75 96
Tensor 128 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_3_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       9216 bytes ( 0.0 MB)  96 1 1 96
Tensor 129 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_4_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 38 38 96
Tensor 130 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_4_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo        384 bytes ( 0.0 MB)  96
Tensor 131 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_4_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo        864 bytes ( 0.0 MB)  1 3 3 96
Tensor 132 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_4_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor 133 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_4_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     277248 bytes ( 0.3 MB)  1 38 38 192
Tensor 134 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_4_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      18432 bytes ( 0.0 MB)  192 1 1 96
Tensor 135 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_5_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     277248 bytes ( 0.3 MB)  1 38 38 192
Tensor 136 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_5_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor 137 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_5_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       1728 bytes ( 0.0 MB)  1 3 3 192
Tensor 138 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_5_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor 139 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_5_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     277248 bytes ( 0.3 MB)  1 38 38 192
Tensor 140 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_5_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      36864 bytes ( 0.0 MB)  192 1 1 192
Tensor 141 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_6_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw      69312 bytes ( 0.1 MB)  1 19 19 192
Tensor 142 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_6_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo        768 bytes ( 0.0 MB)  192
Tensor 143 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_6_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       1728 bytes ( 0.0 MB)  1 3 3 192
Tensor 144 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_6_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 145 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_6_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 146 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_6_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo      73728 bytes ( 0.1 MB)  384 1 1 192
Tensor 147 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_7_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 148 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_7_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 149 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_7_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       3456 bytes ( 0.0 MB)  1 3 3 384
Tensor 150 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_7_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 151 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_7_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 152 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_7_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     147456 bytes ( 0.1 MB)  384 1 1 384
Tensor 153 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_8_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 154 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_8_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 155 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_8_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       3456 bytes ( 0.0 MB)  1 3 3 384
Tensor 156 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_8_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 157 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_8_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 158 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_8_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     147456 bytes ( 0.1 MB)  384 1 1 384
Tensor 159 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_9_depthwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 160 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_9_depthwise/depthwise_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 161 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_9_depthwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo       3456 bytes ( 0.0 MB)  1 3 3 384
Tensor 162 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_9_pointwise/Conv2D_Fold_bias kTfLiteInt32   kTfLiteMmapRo       1536 bytes ( 0.0 MB)  384
Tensor 163 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_9_pointwise/Relu6 kTfLiteUInt8  kTfLiteArenaRw     138624 bytes ( 0.1 MB)  1 19 19 384
Tensor 164 FeatureExtractor/MobilenetV1/MobilenetV1/Conv2d_9_pointwise/weights_quant/FakeQuantWithMinMaxVars kTfLiteUInt8   kTfLiteMmapRo     147456 bytes ( 0.1 MB)  384 1 1 384
Tensor 165 Squeeze              kTfLiteUInt8  kTfLiteArenaRw       7668 bytes ( 0.0 MB)  1 1917 4
Tensor 166 Squeeze_shape        kTfLiteInt32   kTfLiteMmapRo         12 bytes ( 0.0 MB)  3
Tensor 167 TFLite_Detection_PostProcess kTfLiteFloat32  kTfLiteArenaRw        160 bytes ( 0.0 MB)  1 10 4
Tensor 168 TFLite_Detection_PostProcess:1 kTfLiteFloat32  kTfLiteArenaRw         40 bytes ( 0.0 MB)  1 10
Tensor 169 TFLite_Detection_PostProcess:2 kTfLiteFloat32  kTfLiteArenaRw         40 bytes ( 0.0 MB)  1 10
Tensor 170 TFLite_Detection_PostProcess:3 kTfLiteFloat32  kTfLiteArenaRw          4 bytes ( 0.0 MB)  1
Tensor 171 anchors              kTfLiteUInt8   kTfLiteMmapRo       7668 bytes ( 0.0 MB)  1917 4
Tensor 172 concat               kTfLiteUInt8  kTfLiteArenaRw       7668 bytes ( 0.0 MB)  1 1917 1 4
Tensor 173 concat_1             kTfLiteUInt8  kTfLiteArenaRw     174447 bytes ( 0.2 MB)  1 1917 91
Tensor 174 convert_scores       kTfLiteUInt8  kTfLiteArenaRw     174447 bytes ( 0.2 MB)  1 1917 91
Tensor 175 normalized_input_image_tensor kTfLiteUInt8  kTfLiteArenaRw     270000 bytes ( 0.3 MB)  1 300 300 3
Tensor 176 (null)               kTfLiteFloat32  kTfLiteArenaRw      30672 bytes ( 0.0 MB)  1917 4
Tensor 177 (null)               kTfLiteFloat32  kTfLiteArenaRw     697788 bytes ( 0.7 MB)  1917 91
Tensor 178 (null)               kTfLiteUInt8  kTfLiteArenaRw       1917 bytes ( 0.0 MB)  1917
Tensor 179 (null)               kTfLiteUInt8  kTfLiteArenaRw     607500 bytes ( 0.6 MB)  1 150 150 27
Tensor 180 (null)               kTfLiteUInt8  kTfLiteArenaRw      43200 bytes ( 0.0 MB)  1 5 5 1728
Tensor 181 (null)               kTfLiteUInt8  kTfLiteArenaRw       7776 bytes ( 0.0 MB)  1 3 3 864
Tensor 182 (null)               kTfLiteUInt8  kTfLiteArenaRw       3456 bytes ( 0.0 MB)  1 2 2 864
Tensor 183 (null)               kTfLiteUInt8  kTfLiteArenaRw        432 bytes ( 0.0 MB)  1 1 1 432

Node   0 Operator Builtin Code   3 CONV_2D
  Inputs: 175 86 84
  Outputs: 85
  Temporaries: 179
Node   1 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 85 113 112
  Outputs: 111
Node   2 Operator Builtin Code   3 CONV_2D
  Inputs: 111 116 114
  Outputs: 115
Node   3 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 115 119 118
  Outputs: 117
Node   4 Operator Builtin Code   3 CONV_2D
  Inputs: 117 122 120
  Outputs: 121
Node   5 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 121 125 124
  Outputs: 123
Node   6 Operator Builtin Code   3 CONV_2D
  Inputs: 123 128 126
  Outputs: 127
Node   7 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 127 131 130
  Outputs: 129
Node   8 Operator Builtin Code   3 CONV_2D
  Inputs: 129 134 132
  Outputs: 133
Node   9 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 133 137 136
  Outputs: 135
Node  10 Operator Builtin Code   3 CONV_2D
  Inputs: 135 140 138
  Outputs: 139
Node  11 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 139 143 142
  Outputs: 141
Node  12 Operator Builtin Code   3 CONV_2D
  Inputs: 141 146 144
  Outputs: 145
Node  13 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 145 149 148
  Outputs: 147
Node  14 Operator Builtin Code   3 CONV_2D
  Inputs: 147 152 150
  Outputs: 151
Node  15 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 151 155 154
  Outputs: 153
Node  16 Operator Builtin Code   3 CONV_2D
  Inputs: 153 158 156
  Outputs: 157
Node  17 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 157 161 160
  Outputs: 159
Node  18 Operator Builtin Code   3 CONV_2D
  Inputs: 159 164 162
  Outputs: 163
Node  19 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 163 89 88
  Outputs: 87
Node  20 Operator Builtin Code   3 CONV_2D
  Inputs: 87 92 90
  Outputs: 91
Node  21 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 91 95 94
  Outputs: 93
Node  22 Operator Builtin Code   3 CONV_2D
  Inputs: 93 98 96
  Outputs: 97
Node  23 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 97 101 100
  Outputs: 99
Node  24 Operator Builtin Code   3 CONV_2D
  Inputs: 99 104 102
  Outputs: 103
Node  25 Operator Builtin Code   4 DEPTHWISE_CONV_2D
  Inputs: 103 107 106
  Outputs: 105
Node  26 Operator Builtin Code   3 CONV_2D
  Inputs: 105 110 108
  Outputs: 109
Node  27 Operator Builtin Code   3 CONV_2D
  Inputs: 109 62 60
  Outputs: 61
Node  28 Operator Builtin Code   3 CONV_2D
  Inputs: 61 74 72
  Outputs: 73
  Temporaries: 180
Node  29 Operator Builtin Code   3 CONV_2D
  Inputs: 73 65 63
  Outputs: 64
Node  30 Operator Builtin Code   3 CONV_2D
  Inputs: 64 77 75
  Outputs: 76
  Temporaries: 181
Node  31 Operator Builtin Code   3 CONV_2D
  Inputs: 76 68 66
  Outputs: 67
Node  32 Operator Builtin Code   3 CONV_2D
  Inputs: 67 80 78
  Outputs: 79
  Temporaries: 182
Node  33 Operator Builtin Code   3 CONV_2D
  Inputs: 79 71 69
  Outputs: 70
Node  34 Operator Builtin Code   3 CONV_2D
  Inputs: 70 83 81
  Outputs: 82
  Temporaries: 183
Node  35 Operator Builtin Code   3 CONV_2D
  Inputs: 97 2 1
  Outputs: 0
Node  36 Operator Builtin Code  22 RESHAPE
  Inputs: 0 8
  Outputs: 6
Node  37 Operator Builtin Code   3 CONV_2D
  Inputs: 97 5 4
  Outputs: 3
Node  38 Operator Builtin Code  22 RESHAPE
  Inputs: 3 9
  Outputs: 7
Node  39 Operator Builtin Code   3 CONV_2D
  Inputs: 109 12 11
  Outputs: 10
Node  40 Operator Builtin Code  22 RESHAPE
  Inputs: 10 18
  Outputs: 16
Node  41 Operator Builtin Code   3 CONV_2D
  Inputs: 109 15 14
  Outputs: 13
Node  42 Operator Builtin Code  22 RESHAPE
  Inputs: 13 19
  Outputs: 17
Node  43 Operator Builtin Code   3 CONV_2D
  Inputs: 73 22 21
  Outputs: 20
Node  44 Operator Builtin Code  22 RESHAPE
  Inputs: 20 28
  Outputs: 26
Node  45 Operator Builtin Code   3 CONV_2D
  Inputs: 73 25 24
  Outputs: 23
Node  46 Operator Builtin Code  22 RESHAPE
  Inputs: 23 29
  Outputs: 27
Node  47 Operator Builtin Code   3 CONV_2D
  Inputs: 76 32 31
  Outputs: 30
Node  48 Operator Builtin Code  22 RESHAPE
  Inputs: 30 38
  Outputs: 36
Node  49 Operator Builtin Code   3 CONV_2D
  Inputs: 76 35 34
  Outputs: 33
Node  50 Operator Builtin Code  22 RESHAPE
  Inputs: 33 39
  Outputs: 37
Node  51 Operator Builtin Code   3 CONV_2D
  Inputs: 79 42 41
  Outputs: 40
Node  52 Operator Builtin Code  22 RESHAPE
  Inputs: 40 48
  Outputs: 46
Node  53 Operator Builtin Code   3 CONV_2D
  Inputs: 79 45 44
  Outputs: 43
Node  54 Operator Builtin Code  22 RESHAPE
  Inputs: 43 49
  Outputs: 47
Node  55 Operator Builtin Code   3 CONV_2D
  Inputs: 82 52 51
  Outputs: 50
Node  56 Operator Builtin Code  22 RESHAPE
  Inputs: 50 58
  Outputs: 56
Node  57 Operator Builtin Code   2 CONCATENATION
  Inputs: 6 16 26 36 46 56
  Outputs: 172
Node  58 Operator Builtin Code  22 RESHAPE
  Inputs: 172 166
  Outputs: 165
Node  59 Operator Builtin Code   3 CONV_2D
  Inputs: 82 55 54
  Outputs: 53
Node  60 Operator Builtin Code  22 RESHAPE
  Inputs: 53 59
  Outputs: 57
Node  61 Operator Builtin Code   2 CONCATENATION
  Inputs: 7 17 27 37 47 57
  Outputs: 173
Node  62 Operator Builtin Code  14 LOGISTIC
  Inputs: 173
  Outputs: 174
Node  63 Operator Custom Name TFLite_Detection_PostProcess
  Inputs: 165 174 171
  Outputs: 167 168 169 170
  Temporaries: 176 177 178

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

tensorflow lite on x86 / label_image  (0) 2021.01.20
tensorflow-lite 모델  (0) 2021.01.18
mobileNET/SSD  (0) 2021.01.14
caffe  (0) 2021.01.14
darknet detect  (0) 2021.01.11
Posted by 구차니
embeded/raspberry pi2021. 1. 18. 15:25

아따.. tensorflow 하기 힘들다 -_-

[링크 : https://github.com/koenvervloesem/bazel-on-arm]

 

nano scripts/bootstrap/compile.sh
run "${JAVAC}" -classpath "${classpath}" -sourcepath "${sourcepath}" \
      -d "${output}/classes" -source "$JAVA_VERSION" -target "$JAVA_VERSION" \
      -encoding UTF-8 "@${paramfile}" -J-Xmx500M

[링크 : https://gist.github.com/EKami/9869ae6347f68c592c5b5cd181a3b205#3-build-bazel]

 

해도 안되네

The system is out of resources.
Consult the following stack trace for details.
java.lang.OutOfMemoryError: Java heap space

 

 

+

-Xms 가 그나마 Heap 전체에 대해서 설정이 가능한 옵션이 맞긴하나보네..

[링크 : https://www.samsungsds.com/kr/insights/1232761_4627.html]

 

 

+

구글에서 쓰니.. 텐서플로우도 구글꺼니 이래저래 물리는건가?

bazel 생성하는데 javac를 쓰다니 먼가 사악한(?) 놈이다 ㅠㅠ

[링크 : https://bazel.build/]

'embeded > raspberry pi' 카테고리의 다른 글

fbset geometry, timing 수치  (0) 2021.02.19
rpi pico  (0) 2021.01.31
라즈베리에 GUI 환경 설치하기  (0) 2021.01.11
rpi youtube  (0) 2020.12.24
mpicc on rpi  (0) 2020.12.15
Posted by 구차니

집에서 kbs1,2만 겨우 잡히고

sbs, mbc는 채널은 잡히는데 수신은 안되고

ebs는 잡히지 조차 않는데

 

공청 케이블 DTV랑 비교해보는데 화질이 동일하다

대충 보기에는 1080 FHD 같아서 편성표를 보는데

kbs는 하루에 두 세개 정도의 프로그램만 UHD프로 편성이 되어있다.

mbc도 보니 두 세개.

 

평일에는 5개 정도 보이긴하네..

 

도대체 UHD를 왜 사나 싶어지네

 

[schedule.kbs.co.kr/#popup-close]

[schedule.imbc.com/]

Posted by 구차니
프로그램 사용/rtl-sdr2021. 1. 16. 16:57

아는 분에게 USB DVD/CD-ROM 빌려서 드라이버랑 프로그램 CD에서 추출해서 설치하니

되긴 한데 안된다고 해야하나...?

 

아무튼 안테나도 거의 3M 이상 길게 해야지 겨우 2채널 잡혔는데

프로그램 쓰는법을 몰라서 수동으로 채널을 잡는법을 모르겠다.

 

시작 스플래시 이미지

 

TV 버튼을 누르면 DVB-T / DAB / FM이 뜬다.

처음에는 TV만 되는줄 알고 헉 했네..

스캔 1차 시도 -_-

 

안테나 연장해서 다시 하니 겨우 2개 잡힌다.

잡힌 2개 채널 89.1MHz 랑

 

89.7MHz 두개 인듯.

 

채널을 편집할수가 없네..

RTL-SDR 보다는 확실히 CPU를 조금 먹는것 같긴한데

RTL-SDR을 설치하면 윈도우에서 또 돌리기 귀찮으니 고민중..

 

해당 어플리케이션은 G4600 에서 1.4% 정도 먹는데

RTL-SDR도 대충 10% 는 먹었던것 같은데 Gain 건드릴 수 있으면 좀 더 나아지려나?

 

+

안테나 선길이가 점점 길어지는구나 ㅋㅋㅋ

대충 재보니 380cm ㅋㅋㅋㅋ 회로에 쓰는 얇은 선으로 안테나 처럼 만들었는데 고작 5개라니 ㅠㅠㅠ

gqrx에서 확인해보니

89.1 -9 dBFS

89.7 -14 dBFS

98.1 -7 dBFS

93.1 -11 dBFS

93.6 -12 dBFS(삐~ 소리, 대남방송 재밍)

95.1 -13 dBFS

---

91.9 -16 dBFS

93.9 -18 dBFS

 

dBFS가 무슨 단위인지 모르겠지만 RTL-SDR이 잡음이 생겨도 들을순 있다면

RTL Driver를 이용한 FM은 노이즈가 안들릴정도로 깔끔해야만 잡힌다.

'프로그램 사용 > rtl-sdr' 카테고리의 다른 글

수신강도가 너무 약한건가?  (0) 2021.11.25
rtp sdr noaa  (0) 2021.11.24
rtl sdr driver libusb  (0) 2021.01.05
sdr# PAL tv plugin  (0) 2021.01.05
blazeVideo HDTV Player (DVB-T)  (0) 2021.01.04
Posted by 구차니

집에 욕조가 없어서 아쉬웠는데 인터넷에 뜨는 광고보고 아내 알려줬더니

 

"아니 사라고 보여준거 아냐?"

라고 버럭 하더니 질렀다 ㅋㅋㅋㅋ

 

그것도 2개!

 

 

아무튼 도착해서 잘쓰고 있는데

문제는 내 크리스마스 + 생일 선물로 산 

LG V50s를 아내가 어느샌가 자연스럽게 가지고 쓴다 ㅠㅠ

 

그냥 마음을 비우고... 갤럭시 폴드 1이나 찾아보는데 60만원.. 내가 살 수 있는 금액대가 아냐...

'개소리 왈왈 > 육아관련 주저리' 카테고리의 다른 글

오늘자 득템  (0) 2021.01.18
UHD 프로그램이 별로 없네  (0) 2021.01.17
세탁기가 얼다니!!  (0) 2021.01.09
오늘 최고기온 -10도  (4) 2021.01.07
새가 졸리면 하는 행동?  (0) 2021.01.02
Posted by 구차니

문자를 잘 안쓰다 보니 일년도 넘어서야 알게 되었네.

그래서 저번에 MMS를 보냈는데 MMS가 안갔나 보구나 ㅠㅠ

 

증상 : 일반 문자/전화 수신 발신 가능,

         MMS를 발신 에러가 발생하진 않으나 상대방이 수신할 수 없음

 

 

 

휴대폰에 다른 통신사 USIM이 장착되어 있을 경우 발생할 수 있습니다.

[링크 : https://www.samsungsvc.co.kr/...?node_Id=NODE0000152151&kb_Id=KNOW0000042269]

 

모바일 네트워크 - 액세스 포인트 이름 - 설정 초기화

한 2분 정도 걸린것 같은데 APN 2개 있던거 삭제하고 1개가 새롭게 추가되고 나서는

MMS 가 정상적으로 보내진다.

[링크 : https://comterman.tistory.com/2273]

[링크 : https://openart.tistory.com/2349]

'개소리 왈왈 > 모바일 생활' 카테고리의 다른 글

hifi quad dac?  (0) 2021.01.25
LG 핸드폰 접는건가?!  (0) 2021.01.20
(이제야) 갤럭시 폴드 동영상을 보니  (2) 2021.01.15
lg 듀얼 스크린 확장모드  (0) 2021.01.09
액정 보호 필름 도착!!  (0) 2021.01.06
Posted by 구차니

lg v50s와는 "컨셉이 다르다" 라는게 내가 내린 결론

 

갤럭시 폴드는 전면의 작은 화면과 펼치면 큰 화면으로 전환되는 (즉, 두 개 화면을 동시에 쓸 수 없음)

"단일 모니터, 두가지 해상도"라는 컨셉이라면

 

lg v50s dual screen은

컴퓨터로 치면 듀얼 모니터를 싱글 / 듀얼 / eyefinity(AMD) or Surround(NVidia) 모드로 쓰는 컨셉이라고 보면 된다.

"듀얼 모니터, 3가지 해상도, 2가지 모드" 컨셉으로 요약!

 

그렇기에 삼성의 폴드는 사용자가 앱별로 추가 설정할 것이 없고

접어서는 작은 해상도로 나오면 되고, 펼치면 큰 해상도로 나오면 되니 시나리오가 단순해지는데

 

LG dual screen은

평소 모드 / 듀얼 스크린 모드 / 듀얼 확장 모드

3가지를 사용자가 선택을 해야만 한다.

 

그렇기에 사용자 경험이나 편의성 측면에서 좋은 점수를 주기 힘들수 밖에 없고

트리플 모니터를 쓰던 나로서도 핸드폰에까지

그리고 매 앱마다 그런걸 신경써서 써야 하다 보니 매우 귀찮고 짜증이 난다.

 

차라리 LG도 열면 듀얼 확장모드

뒤로 접으면 싱글 모드 식으로 작동하게 했으면 더 편하지 않았을까 생각이 된다.

그리고 사용자 설정에 따라서 듀얼 확장 모드 <-> 듀얼 스크린 모드로 상단 메뉴를 끌어내려 회전 on/off 하듯 했더라면

오히려 사용자 시나리오도 더 간단해지지 않았을까?

 

그리고 dual screen은

뒤로 가도 꺼지지 않게 할 수 있다 보니 시나리오가 이래저래 꼬이는 듯.

 

1. v50s 본체

2. v50s + dual screen 펼치기 + on

3. v50s + dual screen 펼치기 + off

4. v50s + dual screen 뒤집기 + on

5. v50s + dual screen 뒤집기 + off

사용자의 선택권을 주는 것도 좋지만 선택지를 줄여서 사용자 편의를 올리는 것도 방법이 아닐까 싶다.

 

추가로 가로 고정 모드도 있으면 위아래로 놓고 굴러다니면서 침대에서 쓰기 더 편할텐데.

LG 애들은 바른생활 사나이라 죄다 앉아서 공손하게 핸드폰을 만지는 듯 -_-

 

 

아무튼 포캣몬고 두 개 돌리면 두개 스크린에 독립적으로 동시에 터치가 되는건 강점이자 장점이고

(두 사용자가 각각 화면을 쓰고 2인 플레이 가능하도록 구성도 할 수 있으니)

확장모드시 2픽셀정도 중복 출력되는 것은 아쉬움으로 혹은 버그로 생각된다.

그리고 힌지가 일자가 아닌 테트리스 ㄹ 모양 같이 펴지는건 단점이자 약점이다.

 

무게는 폴더블, 폰더블 모두 치명적임 단점이자 한계일지도?

 

 

아무튼 정리 하자면

1. 가로모드 고정이 있으면 좋겠다.

2. 기본이 확장모드로 API 지원 없는 앱이라도 확장 모드로 다 쓸 수 있음 좋겠다.

3. 선택적으로 듀얼 모드를 했으면 좋겠다.

4. 뒤집기 모드 시에는 듀얼 모드로 가거나 off를 선택하면 오히려 시나리오가 깔끔해질듯?

   (그런데 뒤집기 모드를 머하는데 쓸까..)

Posted by 구차니
Programming/openCV2021. 1. 14. 17:23

오랫만에 빌드하니 다 까먹었네 -_ㅠ

 

g++ myprog.cpp -lopencv_core -lopencv_imgproc -lopencv_videoio -lopencv_highgui -lopencv_objdetect -o myprog

[링크 : https://answers.opencv.org/question/165673]

 

 

아놔.. 에러 뿜뿜 -_-

/usr/bin/ld: warning: libopencv_imgproc.so.3.4, needed by //usr/local/lib/libopencv_dnn.so, may conflict with libopencv_imgproc.so.3.2
/usr/bin/ld: warning: libopencv_core.so.3.4, needed by //usr/local/lib/libopencv_dnn.so, may conflict with libopencv_core.so.3.2
/usr/bin/ld: /tmp/cc64nD60.o: undefined reference to symbol '_ZNK2cv3Mat5emptyEv'
/usr/bin/ld: //usr/local/lib/libopencv_core.so.3.4: error adding symbols: DSO missing from command line

 

+

아무튼 라즈베리 3B 에서 하는데

3.2.0이 설치되어 있고, 필요에 의해서 3.4.0을 빌드해서 설치했더니 3.2.0과 섞여서 난리가 났다.

그래서 -L 옵션을 통해 3.4.0이 설치된 곳을 우선적으로 보도록 해주니 문제없이 해결!

$ whereis libopencv_imgproc
libopencv_imgproc: /usr/lib/arm-linux-gnueabihf/libopencv_imgproc.so /usr/lib/arm-linux-gnueabihf/libopencv_imgproc.a /usr/local/lib/libopencv_imgproc.so

$ ls -al /usr/local/lib/libopencv_imgproc.so*
lrwxrwxrwx 1 root root      24 Jan 14 08:04 /usr/local/lib/libopencv_imgproc.so -> libopencv_imgproc.so.3.4
lrwxrwxrwx 1 root root      27 Jan 14 08:04 /usr/local/lib/libopencv_imgproc.so.3.4 -> libopencv_imgproc.so.3.4.13
-rw-r--r-- 1 root root 3292320 Jan 14 05:01 /usr/local/lib/libopencv_imgproc.so.3.4.13

$ ls -al /usr/lib/arm-linux-gnueabihf/libopencv_imgproc.so
lrwxrwxrwx 1 root root 24 Feb 12  2019 /usr/lib/arm-linux-gnueabihf/libopencv_imgproc.so -> libopencv_imgproc.so.3.2

$ g++ opencv_video.cpp -lopencv_core -lopencv_dnn -lopencv_imgproc -lopencv_videoio -lopencv_highgui -lopencv_objdetect -lopencv_imgcodecs -L/usr/local/lib

 

 

 

 

MobileNet-SSD의 ssd_detect.cpp 빌드하려다가

아 몰라 대충 흑화중.. 후...

sudo apt-get install libcaffe-cpu-dev libcaffe-cpu1 libboost-all-dev libgflags-dev libgoogle-glog-dev libprotobuf-dev libopenblas-dev

$ g++ ssd_detect.cpp -lopencv_core -lopencv_dnn -lopencv_imgproc -lopencv_videoio -lopencv_highgui -lopencv_objdetect -lopencv_imgcodecs -L/usr/local/lib -o ssd_detect
ssd_detect.cpp:15:10: fatal error: caffe/caffe.hpp: No such file or directory
 #include <caffe/caffe.hpp>
          ^~~~~~~~~~~~~~~~~
compilation terminated.
pi@raspberrypi:~/src/MobileNet-SSD $ g++ ssd_detect.cpp -lopencv_core -lopencv_dnn -lopencv_imgproc -lopencv_videoio -lopencv_highgui -lopencv_objdetect -lopencv_imgcodecs -L/usr/local/lib -o ssd_detect
In file included from /usr/include/caffe/blob.hpp:8,
                 from /usr/include/caffe/caffe.hpp:7,
                 from ssd_detect.cpp:15:
/usr/include/caffe/common.hpp:4:10: fatal error: boost/shared_ptr.hpp: No such file or directory
 #include <boost/shared_ptr.hpp>
          ^~~~~~~~~~~~~~~~~~~~~~
compilation terminated.

pi@raspberrypi:~/src/MobileNet-SSD $ g++ ssd_detect.cpp -lopencv_core -lopencv_dnn -lopencv_imgproc -lopencv_videoio -lopencv_highgui -lopencv_objdetect -lopencv_imgcodecs -L/usr/local/lib -o ssd_detect
In file included from /usr/include/caffe/blob.hpp:8,
                 from /usr/include/caffe/caffe.hpp:7,
                 from ssd_detect.cpp:15:
/usr/include/caffe/common.hpp:5:10: fatal error: gflags/gflags.h: No such file or directory
 #include <gflags/gflags.h>
          ^~~~~~~~~~~~~~~~~
compilation terminated.
$ g++ ssd_detect.cpp -lopencv_core -lopencv_dnn -lopencv_imgproc -lopencv_videoio -lopencv_highgui -lopencv_objdetect -lopencv_imgcodecs -L/usr/local/lib -o ssd_detect
In file included from /usr/include/caffe/common.hpp:19,
                 from /usr/include/caffe/blob.hpp:8,
                 from /usr/include/caffe/caffe.hpp:7,
                 from ssd_detect.cpp:15:
/usr/include/caffe/util/device_alternate.hpp:34:10: fatal error: cublas_v2.h: No such file or directory
 #include <cublas_v2.h>
          ^~~~~~~~~~~~~
compilation terminated.

-DCPU_ONLY


 $ g++ ssd_detect.cpp -lopencv_core -lopencv_dnn -lopencv_imgproc -lopencv_videoio -lopencv_highgui -lopencv_objdetect -lopencv_imgcodecs -L/usr/local/lib -o ssd_detect -DCPU_ONLY
In file included from /usr/include/caffe/util/math_functions.hpp:11,
                 from /usr/include/caffe/filler.hpp:13,
                 from /usr/include/caffe/caffe.hpp:9,
                 from ssd_detect.cpp:15:
/usr/include/caffe/util/mkl_alternate.hpp:14:10: fatal error: cblas.h: No such file or directory
 #include <cblas.h>
          ^~~~~~~~~
compilation terminated.

$ g++ ssd_detect.cpp -lopencv_core -lopencv_dnn -lopencv_imgproc -lopencv_videoio -lopencv_highgui -lopencv_objdetect -lopencv_imgcodecs -L/usr/local/lib -o ssd_detect -DCPU_ONLY -lcaffe
/usr/bin/ld: /tmp/ccDNrn3i.o: undefined reference to symbol '_ZN6google4base21CheckOpMessageBuilder7ForVar2Ev'
/usr/bin/ld: //lib/arm-linux-gnueabihf/libglog.so.0: error adding symbols: DSO missing from command line
collect2: error: ld returned 1 exit status


$ g++ ssd_detect.cpp -lopencv_core -lopencv_dnn -lopencv_imgproc -lopencv_videoio -lopencv_highgui -lopencv_objdetect -lopencv_imgcodecs -L/usr/local/lib -o ssd_detect -DCPU_ONLY -lcaffe -lglog
/usr/bin/ld: /tmp/ccO8VjqT.o: undefined reference to symbol '_ZN6google26ShowUsageWithFlagsRestrictEPKcS1_'
/usr/bin/ld: //lib/arm-linux-gnueabihf/libgflags.so.2.2: error adding symbols: DSO missing from command line
collect2: error: ld returned 1 exit status




$ g++ ssd_detect.cpp -lopencv_core -lopencv_dnn -lopencv_imgproc -lopencv_videoio -lopencv_highgui -lopencv_objdetect -lopencv_imgcodecs -L/usr/local/lib -o ssd_detect -DCPU_ONLY -lcaffe -lgflags -lglog
/usr/bin/ld: warning: libopencv_imgcodecs.so.3.2, needed by /usr/lib/gcc/arm-linux-gnueabihf/8/../../../arm-linux-gnueabihf/libcaffe.so, may conflict with libopencv_imgcodecs.so.3.4
/usr/bin/ld: warning: libopencv_imgproc.so.3.2, needed by /usr/lib/gcc/arm-linux-gnueabihf/8/../../../arm-linux-gnueabihf/libcaffe.so, may conflict with libopencv_imgproc.so.3.4
/usr/bin/ld: warning: libopencv_core.so.3.2, needed by /usr/lib/gcc/arm-linux-gnueabihf/8/../../../arm-linux-gnueabihf/libcaffe.so, may conflict with libopencv_core.so.3.4

 

 

caffe 빌드

위에 lib들 설치하고 아래것 추가로 설치

sudo apt-get install protobuf-compiler libhdf5-dev liblmdb-dev libleveldb-dev libatlas-base-dev python-numpy

 

sudo apt-get install protobuf-compiler

[링크 : https://stackoverflow.com/questions/46698260/could-not-find-protobuf-compiler]

 

sudo apt-get install libhdf5-dev

[링크 : https://github.com/jcjohnson/torch-rnn/issues/121]

 

sudo apt-get install liblmdb-dev

[링크 : https://caffe.berkeleyvision.org/install_apt.html]

 

Could NOT find LevelDB

sudo apt-get install libleveldb-dev

 

Could NOT find Atlas

sudo apt-get install libatlas-base-dev

 

Could NOT find NumPy

sudo apt-get install python-numpy

[링크 : https://github.com/CMU-Perceptual-Computing-Lab/openpose/issues/306]

 

'Programming > openCV' 카테고리의 다른 글

cv2.imshow cv2.waitKey  (0) 2022.03.14
virtual mouse  (0) 2022.01.25
opencv face detect  (0) 2019.05.10
vscode python3 opencv lint  (0) 2019.05.10
opencv cannyedge  (0) 2019.01.16
Posted by 구차니

python 안쓰고 c++ 만으로 해보려는데

tensorflow / rosrun 등이 필요함

mobilenet ssd
tensorflow /c++/python
catkin_make

[링크 : https://github.com/haosen9527/mobileNet-ssd]

 

빌드는 독립적으로 되는데 caffe를 먼저 빌드는 해야 함..

caffe를 빌드하는데 openCV-3.4.0 이상을 요구하고 있음..

Also it may help the beginners to build a project using cmake. You don't need to build this within the caffe root. But remember you should compile caffe first.

[링크 : https://github.com/MediosZ/MobileNet-SSD]

 

+

python으로 된 가장 원본 파일?

[링크 : https://github.com/chuanqi305/MobileNet-SSD]

 

+

tensorflow lite 버전과 openCV-4.1.0 이상 버전 요구?

[링크 : https://github.com/finnickniu/tensorflow_object_detection_tflite]

 

+

rpi 2b에서 텐서가 되는지 모르겠네?

[링크 : https://www.tensorflow.org/install/source_rpi?hl=ko]

'프로그램 사용 > yolo_tensorflow' 카테고리의 다른 글

tensorflow-lite 모델  (0) 2021.01.18
tensorflow-lite minimal.cc 실행  (0) 2021.01.18
caffe  (0) 2021.01.14
darknet detect  (0) 2021.01.11
CNN convolution과 maxpool  (0) 2021.01.10
Posted by 구차니