Colab_OpenVINO_mobilenet_v1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Colab_OpenVINO_mobilenet-v1.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PUynAV9J1eRi",
        "colab_type": "text"
      },
      "source": [
        "# **Colab安裝OpenVINO並下載執行mobilenet-v1影像分類範例**\n",
        "\n",
        "範例提供：OmniXRI Jack 2020.9.10\n",
        "\n",
        "主要工作項目：\n",
        "\n",
        "1.   檢查系統\n",
        "2.   安裝OpenVINO工具包\n",
        "3.   模型下載\n",
        "4.   模型優化\n",
        "5.   執行推論\n",
        "\n",
        "參考範例來源：https://colab.research.google.com/drive/1-ETOSlDftRCkAW4FoBHGLgslfaRMRqBU"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_diohn2ldAJc",
        "colab_type": "text"
      },
      "source": [
        "# **0. 檢查系統**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W2Evd4UedTyh",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 453
        },
        "outputId": "27176447-04b5-494d-d8dd-75a5f04c36e1"
      },
      "source": [
        "# 檢查CPU狀態\n",
        "# 預設會分配到 Intel Xeon CPU x64 2.2GHz 二顆\n",
        "# OpenVINO只支援虛擬機的Intel CPU Xeon，不支援虛擬機的Nvidia GPU或Google TPU\n",
        "!lscpu"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Architecture:        x86_64\n",
            "CPU op-mode(s):      32-bit, 64-bit\n",
            "Byte Order:          Little Endian\n",
            "CPU(s):              2\n",
            "On-line CPU(s) list: 0,1\n",
            "Thread(s) per core:  2\n",
            "Core(s) per socket:  1\n",
            "Socket(s):           1\n",
            "NUMA node(s):        1\n",
            "Vendor ID:           GenuineIntel\n",
            "CPU family:          6\n",
            "Model:               63\n",
            "Model name:          Intel(R) Xeon(R) CPU @ 2.30GHz\n",
            "Stepping:            0\n",
            "CPU MHz:             2300.000\n",
            "BogoMIPS:            4600.00\n",
            "Hypervisor vendor:   KVM\n",
            "Virtualization type: full\n",
            "L1d cache:           32K\n",
            "L1i cache:           32K\n",
            "L2 cache:            256K\n",
            "L3 cache:            46080K\n",
            "NUMA node0 CPU(s):   0,1\n",
            "Flags:               fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm invpcid_single ssbd ibrs ibpb stibp fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid xsaveopt arat md_clear arch_capabilities\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7gc99_HVdV-I",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 225
        },
        "outputId": "721cf2c2-093e-4ecc-8b00-d097474e42c4"
      },
      "source": [
        "# 檢查作業系統狀態\n",
        "# 預設會分配到 Ubuntu 18.04.3 LTS版本\n",
        "!cat /etc/os-release"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "NAME=\"Ubuntu\"\n",
            "VERSION=\"18.04.3 LTS (Bionic Beaver)\"\n",
            "ID=ubuntu\n",
            "ID_LIKE=debian\n",
            "PRETTY_NAME=\"Ubuntu 18.04.3 LTS\"\n",
            "VERSION_ID=\"18.04\"\n",
            "HOME_URL=\"https://www.ubuntu.com/\"\n",
            "SUPPORT_URL=\"https://help.ubuntu.com/\"\n",
            "BUG_REPORT_URL=\"https://bugs.launchpad.net/ubuntu/\"\n",
            "PRIVACY_POLICY_URL=\"https://www.ubuntu.com/legal/terms-and-policies/privacy-policy\"\n",
            "VERSION_CODENAME=bionic\n",
            "UBUNTU_CODENAME=bionic\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x-iMA0iIXMYt",
        "colab_type": "text"
      },
      "source": [
        "# **1. 安裝Intel OpenVINO工具包**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3BPyuF_tMR5I",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 208
        },
        "outputId": "2e59e2ac-043f-4486-825a-63c18dd58967"
      },
      "source": [
        "# 取得OpenVINO2020公開金錀\n",
        "!wget https://apt.repos.intel.com/openvino/2020/GPG-PUB-KEY-INTEL-OPENVINO-2020"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2020-09-10 02:40:24--  https://apt.repos.intel.com/openvino/2020/GPG-PUB-KEY-INTEL-OPENVINO-2020\n",
            "Resolving apt.repos.intel.com (apt.repos.intel.com)... 23.62.218.27, 2600:1402:3800:2ab::4b23, 2600:1402:3800:2aa::4b23\n",
            "Connecting to apt.repos.intel.com (apt.repos.intel.com)|23.62.218.27|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 939 [binary/octet-stream]\n",
            "Saving to: ‘GPG-PUB-KEY-INTEL-OPENVINO-2020’\n",
            "\n",
            "GPG-PUB-KEY-INTEL-O 100%[===================>]     939  --.-KB/s    in 0s      \n",
            "\n",
            "2020-09-10 02:40:25 (24.1 MB/s) - ‘GPG-PUB-KEY-INTEL-OPENVINO-2020’ saved [939/939]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vqQ1evnXLWuq",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "bbbe5f3d-3e36-4bfa-b51c-d94058098c15"
      },
      "source": [
        "# 加入OpenVINO公開金錀到系統金錀群中\n",
        "!apt-key add GPG-PUB-KEY-INTEL-OPENVINO-2020"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "OK\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8YTK6yZYMjtJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# 建立更新安裝清單檔案\n",
        "!touch /etc/apt/sources.list.d/intel-openvino-2020.list"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sjB2tPE5Mjv7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# 將下載指令加入安裝清單中\n",
        "!echo \"deb https://apt.repos.intel.com/openvino/2020 all main\" >> /etc/apt/sources.list.d/intel-openvino-2020.list"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p_Y0njq-MsoP",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 589
        },
        "outputId": "733caea3-c567-4a96-ab58-2de0d6d9515e"
      },
      "source": [
        "# 更新系統\n",
        "!apt update"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[33m\r0% [Working]\u001b[0m\r            \rIgn:1 https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64  InRelease\n",
            "\u001b[33m\r0% [Connecting to archive.ubuntu.com (91.189.88.152)] [Waiting for headers] [Wa\u001b[0m\r                                                                               \rGet:2 https://cloud.r-project.org/bin/linux/ubuntu bionic-cran35/ InRelease [3,626 B]\n",
            "\u001b[33m\r0% [Connecting to archive.ubuntu.com (91.189.88.152)] [Waiting for headers] [2 \u001b[0m\u001b[33m\r0% [Connecting to archive.ubuntu.com (91.189.88.152)] [Waiting for headers] [Wa\u001b[0m\u001b[33m\r0% [2 InRelease gpgv 3,626 B] [Connecting to archive.ubuntu.com (91.189.88.152)\u001b[0m\r                                                                               \rGet:3 http://security.ubuntu.com/ubuntu bionic-security InRelease [88.7 kB]\n",
            "\u001b[33m\r0% [2 InRelease gpgv 3,626 B] [Connecting to archive.ubuntu.com (91.189.88.152)\u001b[0m\r                                                                               \rGet:4 http://ppa.launchpad.net/graphics-drivers/ppa/ubuntu bionic InRelease [21.3 kB]\n",
            "\u001b[33m\r0% [2 InRelease gpgv 3,626 B] [Connecting to archive.ubuntu.com (91.189.88.152)\u001b[0m\r                                                                               \rIgn:5 https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64  InRelease\n",
            "\u001b[33m\r0% [2 InRelease gpgv 3,626 B] [Waiting for headers] [3 InRelease 14.2 kB/88.7 k\u001b[0m\r                                                                               \rHit:6 https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64  Release\n",
            "\u001b[33m\r0% [2 InRelease gpgv 3,626 B] [Waiting for headers] [3 InRelease 14.2 kB/88.7 k\u001b[0m\r                                                                               \rGet:7 https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64  Release [564 B]\n",
            "Get:8 https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64  Release.gpg [833 B]\n",
            "Hit:9 http://archive.ubuntu.com/ubuntu bionic InRelease\n",
            "Get:10 https://apt.repos.intel.com/openvino/2020 all InRelease [5,659 B]\n",
            "Get:11 http://archive.ubuntu.com/ubuntu bionic-updates InRelease [88.7 kB]\n",
            "Get:12 https://cloud.r-project.org/bin/linux/ubuntu bionic-cran35/ Packages [95.7 kB]\n",
            "Get:13 http://ppa.launchpad.net/marutter/c2d4u3.5/ubuntu bionic InRelease [15.4 kB]\n",
            "Get:15 http://archive.ubuntu.com/ubuntu bionic-backports InRelease [74.6 kB]\n",
            "Get:16 https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64  Packages [47.5 kB]\n",
            "Get:17 http://ppa.launchpad.net/graphics-drivers/ppa/ubuntu bionic/main amd64 Packages [43.0 kB]\n",
            "Get:18 http://security.ubuntu.com/ubuntu bionic-security/universe amd64 Packages [897 kB]\n",
            "Get:19 http://ppa.launchpad.net/marutter/c2d4u3.5/ubuntu bionic/main Sources [1,864 kB]\n",
            "Get:20 https://apt.repos.intel.com/openvino/2020 all/main amd64 Packages [17.8 kB]\n",
            "Get:21 http://security.ubuntu.com/ubuntu bionic-security/restricted amd64 Packages [116 kB]\n",
            "Get:22 http://archive.ubuntu.com/ubuntu bionic-updates/multiverse amd64 Packages [27.7 kB]\n",
            "Get:23 http://security.ubuntu.com/ubuntu bionic-security/multiverse amd64 Packages [10.1 kB]\n",
            "Get:24 http://security.ubuntu.com/ubuntu bionic-security/main amd64 Packages [1,089 kB]\n",
            "Get:25 http://archive.ubuntu.com/ubuntu bionic-updates/universe amd64 Packages [1,425 kB]\n",
            "Get:26 http://archive.ubuntu.com/ubuntu bionic-updates/main amd64 Packages [1,384 kB]\n",
            "Get:27 http://archive.ubuntu.com/ubuntu bionic-updates/restricted amd64 Packages [132 kB]\n",
            "Get:28 https://apt.repos.intel.com/openvino/2020 all/main all Packages [2,522 B]\n",
            "Get:29 http://ppa.launchpad.net/marutter/c2d4u3.5/ubuntu bionic/main amd64 Packages [900 kB]\n",
            "Fetched 8,352 kB in 3s (2,401 kB/s)\n",
            "Reading package lists... Done\n",
            "Building dependency tree       \n",
            "Reading state information... Done\n",
            "74 packages can be upgraded. Run 'apt list --upgradable' to see them.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IJLw0IbBh8hj",
        "colab_type": "text"
      },
      "source": [
        "可利用 !apt-cache search openvino 指令搜尋可安裝的版本系統包，包含Runtime及Developer Packages"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6TdMkjTMMypN",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "99ef083f-ae94-4948-b77a-fcc1bbab3f85"
      },
      "source": [
        "# 安裝OpenVINO到虛擬機系統中\n",
        "!apt install intel-openvino-dev-ubuntu18-2020.4.287"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Reading package lists... Done\n",
            "Building dependency tree       \n",
            "Reading state information... Done\n",
            "The following package was automatically installed and is no longer required:\n",
            "  libnvidia-common-440\n",
            "Use 'apt autoremove' to remove it.\n",
            "The following additional packages will be installed:\n",
            "  intel-openvino-dl-workbench-2020.4.287 intel-openvino-docs-2020.4.287\n",
            "  intel-openvino-eula-2020.4.287\n",
            "  intel-openvino-gstreamer-rt-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-gva-dev-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-gva-rt-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-gva-sdk-2020.4.287\n",
            "  intel-openvino-ie-bin-python-tools-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-rt-2020.4.287\n",
            "  intel-openvino-ie-rt-core-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-rt-cpu-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-rt-gna-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-rt-gpu-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-rt-hddl-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-rt-vpu-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-samples-2020.4.287\n",
            "  intel-openvino-ie-sdk-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-model-optimizer-2020.4.287 intel-openvino-omz-dev-2020.4.287\n",
            "  intel-openvino-omz-tools-2020.4.287 intel-openvino-opencv-etc-2020.4.287\n",
            "  intel-openvino-opencv-generic-2020.4.287\n",
            "  intel-openvino-opencv-lib-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-pot-2020.4.287 intel-openvino-setupvars-2020.4.287\n",
            "The following NEW packages will be installed:\n",
            "  intel-openvino-dev-ubuntu18-2020.4.287\n",
            "  intel-openvino-dl-workbench-2020.4.287 intel-openvino-docs-2020.4.287\n",
            "  intel-openvino-eula-2020.4.287\n",
            "  intel-openvino-gstreamer-rt-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-gva-dev-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-gva-rt-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-gva-sdk-2020.4.287\n",
            "  intel-openvino-ie-bin-python-tools-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-rt-2020.4.287\n",
            "  intel-openvino-ie-rt-core-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-rt-cpu-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-rt-gna-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-rt-gpu-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-rt-hddl-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-rt-vpu-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-ie-samples-2020.4.287\n",
            "  intel-openvino-ie-sdk-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-model-optimizer-2020.4.287 intel-openvino-omz-dev-2020.4.287\n",
            "  intel-openvino-omz-tools-2020.4.287 intel-openvino-opencv-etc-2020.4.287\n",
            "  intel-openvino-opencv-generic-2020.4.287\n",
            "  intel-openvino-opencv-lib-ubuntu-bionic-2020.4.287\n",
            "  intel-openvino-pot-2020.4.287 intel-openvino-setupvars-2020.4.287\n",
            "0 upgraded, 26 newly installed, 0 to remove and 74 not upgraded.\n",
            "Need to get 178 MB of archives.\n",
            "After this operation, 0 B of additional disk space will be used.\n",
            "Get:1 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-eula-2020.4.287 all 2020.4-287 [99.9 kB]\n",
            "Get:2 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-setupvars-2020.4.287 all 2020.4-287 [6,870 B]\n",
            "Get:3 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-docs-2020.4.287 all 2020.4-287 [13.4 kB]\n",
            "Get:4 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-ie-samples-2020.4.287 amd64 2020.4-287 [2,475 kB]\n",
            "Get:5 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-model-optimizer-2020.4.287 amd64 2020.4-287 [437 kB]\n",
            "Get:6 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-ie-bin-python-tools-ubuntu-bionic-2020.4.287 amd64 2020.4-287 [19.4 kB]\n",
            "Get:7 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-ie-sdk-ubuntu-bionic-2020.4.287 amd64 2020.4-287 [682 kB]\n",
            "Get:8 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-ie-rt-2020.4.287 amd64 2020.4-287 [637 kB]\n",
            "Get:9 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-ie-rt-core-ubuntu-bionic-2020.4.287 amd64 2020.4-287 [6,621 kB]\n",
            "Get:10 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-ie-rt-cpu-ubuntu-bionic-2020.4.287 amd64 2020.4-287 [4,192 kB]\n",
            "Get:11 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-ie-rt-gpu-ubuntu-bionic-2020.4.287 amd64 2020.4-287 [2,201 kB]\n",
            "Get:12 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-ie-rt-gna-ubuntu-bionic-2020.4.287 amd64 2020.4-287 [1,034 kB]\n",
            "Get:13 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-ie-rt-vpu-ubuntu-bionic-2020.4.287 amd64 2020.4-287 [20.4 MB]\n",
            "Get:14 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-ie-rt-hddl-ubuntu-bionic-2020.4.287 amd64 2020.4-287 [3,076 kB]\n",
            "Get:15 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-opencv-generic-2020.4.287 all 2020.4-287 [710 kB]\n",
            "Get:16 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-opencv-lib-ubuntu-bionic-2020.4.287 amd64 2020.4-287 [16.2 MB]\n",
            "Get:17 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-opencv-etc-2020.4.287 all 2020.4-287 [1,209 kB]\n",
            "Get:18 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-omz-dev-2020.4.287 amd64 2020.4-287 [32.6 MB]\n",
            "Get:19 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-omz-tools-2020.4.287 amd64 2020.4-287 [952 kB]\n",
            "Get:20 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-dl-workbench-2020.4.287 amd64 2020.4-287 [1,044 kB]\n",
            "Get:21 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-gva-sdk-2020.4.287 amd64 2020.4-287 [18.7 kB]\n",
            "Get:22 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-gva-dev-ubuntu-bionic-2020.4.287 amd64 2020.4-287 [5,446 kB]\n",
            "Get:23 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-gstreamer-rt-ubuntu-bionic-2020.4.287 amd64 2020.4-287 [76.4 MB]\n",
            "Get:24 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-gva-rt-ubuntu-bionic-2020.4.287 amd64 2020.4-287 [832 kB]\n",
            "Get:25 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-pot-2020.4.287 amd64 2020.4-287 [429 kB]\n",
            "Get:26 https://apt.repos.intel.com/openvino/2020 all/main amd64 intel-openvino-dev-ubuntu18-2020.4.287 amd64 2020.4-287 [1,830 B]\n",
            "Fetched 178 MB in 55s (3,218 kB/s)\n",
            "Selecting previously unselected package intel-openvino-eula-2020.4.287.\n",
            "(Reading database ... 144579 files and directories currently installed.)\n",
            "Preparing to unpack .../00-intel-openvino-eula-2020.4.287_2020.4-287_all.deb ...\n",
            "Unpacking intel-openvino-eula-2020.4.287 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-setupvars-2020.4.287.\n",
            "Preparing to unpack .../01-intel-openvino-setupvars-2020.4.287_2020.4-287_all.deb ...\n",
            "Unpacking intel-openvino-setupvars-2020.4.287 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-docs-2020.4.287.\n",
            "Preparing to unpack .../02-intel-openvino-docs-2020.4.287_2020.4-287_all.deb ...\n",
            "Unpacking intel-openvino-docs-2020.4.287 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-ie-samples-2020.4.287:amd64.\n",
            "Preparing to unpack .../03-intel-openvino-ie-samples-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-ie-samples-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-model-optimizer-2020.4.287:amd64.\n",
            "Preparing to unpack .../04-intel-openvino-model-optimizer-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-model-optimizer-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-ie-bin-python-tools-ubuntu-bionic-2020.4.287:amd64.\n",
            "Preparing to unpack .../05-intel-openvino-ie-bin-python-tools-ubuntu-bionic-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-ie-bin-python-tools-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-ie-sdk-ubuntu-bionic-2020.4.287:amd64.\n",
            "Preparing to unpack .../06-intel-openvino-ie-sdk-ubuntu-bionic-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-ie-sdk-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-ie-rt-2020.4.287:amd64.\n",
            "Preparing to unpack .../07-intel-openvino-ie-rt-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-ie-rt-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-ie-rt-core-ubuntu-bionic-2020.4.287:amd64.\n",
            "Preparing to unpack .../08-intel-openvino-ie-rt-core-ubuntu-bionic-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-ie-rt-core-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-ie-rt-cpu-ubuntu-bionic-2020.4.287:amd64.\n",
            "Preparing to unpack .../09-intel-openvino-ie-rt-cpu-ubuntu-bionic-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-ie-rt-cpu-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-ie-rt-gpu-ubuntu-bionic-2020.4.287:amd64.\n",
            "Preparing to unpack .../10-intel-openvino-ie-rt-gpu-ubuntu-bionic-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-ie-rt-gpu-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-ie-rt-gna-ubuntu-bionic-2020.4.287:amd64.\n",
            "Preparing to unpack .../11-intel-openvino-ie-rt-gna-ubuntu-bionic-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-ie-rt-gna-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-ie-rt-vpu-ubuntu-bionic-2020.4.287:amd64.\n",
            "Preparing to unpack .../12-intel-openvino-ie-rt-vpu-ubuntu-bionic-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-ie-rt-vpu-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-ie-rt-hddl-ubuntu-bionic-2020.4.287:amd64.\n",
            "Preparing to unpack .../13-intel-openvino-ie-rt-hddl-ubuntu-bionic-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-ie-rt-hddl-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-opencv-generic-2020.4.287.\n",
            "Preparing to unpack .../14-intel-openvino-opencv-generic-2020.4.287_2020.4-287_all.deb ...\n",
            "Unpacking intel-openvino-opencv-generic-2020.4.287 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-opencv-lib-ubuntu-bionic-2020.4.287:amd64.\n",
            "Preparing to unpack .../15-intel-openvino-opencv-lib-ubuntu-bionic-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-opencv-lib-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-opencv-etc-2020.4.287.\n",
            "Preparing to unpack .../16-intel-openvino-opencv-etc-2020.4.287_2020.4-287_all.deb ...\n",
            "Unpacking intel-openvino-opencv-etc-2020.4.287 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-omz-dev-2020.4.287:amd64.\n",
            "Preparing to unpack .../17-intel-openvino-omz-dev-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-omz-dev-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-omz-tools-2020.4.287:amd64.\n",
            "Preparing to unpack .../18-intel-openvino-omz-tools-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-omz-tools-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-dl-workbench-2020.4.287:amd64.\n",
            "Preparing to unpack .../19-intel-openvino-dl-workbench-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-dl-workbench-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-gva-sdk-2020.4.287:amd64.\n",
            "Preparing to unpack .../20-intel-openvino-gva-sdk-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-gva-sdk-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-gva-dev-ubuntu-bionic-2020.4.287:amd64.\n",
            "Preparing to unpack .../21-intel-openvino-gva-dev-ubuntu-bionic-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-gva-dev-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-gstreamer-rt-ubuntu-bionic-2020.4.287:amd64.\n",
            "Preparing to unpack .../22-intel-openvino-gstreamer-rt-ubuntu-bionic-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-gstreamer-rt-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-gva-rt-ubuntu-bionic-2020.4.287:amd64.\n",
            "Preparing to unpack .../23-intel-openvino-gva-rt-ubuntu-bionic-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-gva-rt-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-pot-2020.4.287:amd64.\n",
            "Preparing to unpack .../24-intel-openvino-pot-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-pot-2020.4.287:amd64 (2020.4-287) ...\n",
            "Selecting previously unselected package intel-openvino-dev-ubuntu18-2020.4.287:amd64.\n",
            "Preparing to unpack .../25-intel-openvino-dev-ubuntu18-2020.4.287_2020.4-287_amd64.deb ...\n",
            "Unpacking intel-openvino-dev-ubuntu18-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-pot-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-setupvars-2020.4.287 (2020.4-287) ...\n",
            "Setting up intel-openvino-opencv-generic-2020.4.287 (2020.4-287) ...\n",
            "Setting up intel-openvino-gva-sdk-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-omz-tools-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-dl-workbench-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-opencv-etc-2020.4.287 (2020.4-287) ...\n",
            "Setting up intel-openvino-eula-2020.4.287 (2020.4-287) ...\n",
            "Setting up intel-openvino-docs-2020.4.287 (2020.4-287) ...\n",
            "Setting up intel-openvino-gva-rt-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-ie-rt-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-ie-samples-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-gstreamer-rt-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-ie-bin-python-tools-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-ie-sdk-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-gva-dev-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-omz-dev-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-model-optimizer-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-opencv-lib-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-ie-rt-core-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-ie-rt-gna-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-ie-rt-cpu-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-ie-rt-gpu-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-ie-rt-hddl-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-ie-rt-vpu-ubuntu-bionic-2020.4.287:amd64 (2020.4-287) ...\n",
            "Setting up intel-openvino-dev-ubuntu18-2020.4.287:amd64 (2020.4-287) ...\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Y_wMT5onL-qY",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "6ad95056-bba9-4c1d-9dfb-3f0d3f7b8441"
      },
      "source": [
        "# 列出安裝路徑下內容\n",
        "!ls /opt/intel"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "openvino  openvino_2020.4.287\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OVt2eeaeRZ98",
        "colab_type": "text"
      },
      "source": [
        "# **2. 模型下載**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0CB7jHlWhJ_c",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 208
        },
        "outputId": "03b38ebb-1793-42e8-fc24-996c7618f8c2"
      },
      "source": [
        "# 執行環境設定批次檔並以模型下載器取得mobilenet-v1-1.0-224\n",
        "# 可以用 --name 取得OpenVINO提供之任意指定模型名稱\n",
        "!source /opt/intel/openvino/bin/setupvars.sh && \\\n",
        "python /opt/intel/openvino/deployment_tools/tools/model_downloader/downloader.py \\\n",
        "--name mobilenet-v1-1.0-224"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[setupvars.sh] OpenVINO environment initialized\n",
            "################|| Downloading models ||################\n",
            "\n",
            "========== Downloading /content/public/mobilenet-v1-1.0-224/mobilenet-v1-1.0-224.prototxt\n",
            "... 100%, 27 KB, 83011 KB/s, 0 seconds passed\n",
            "\n",
            "========== Downloading /content/public/mobilenet-v1-1.0-224/mobilenet-v1-1.0-224.caffemodel\n",
            "... 100%, 16627 KB, 24777 KB/s, 0 seconds passed\n",
            "\n",
            "################|| Post-processing ||################\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hptfdelFy65V",
        "colab_type": "text"
      },
      "source": [
        "# **3. 模型優化**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ui7toGLHz9Zk",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 124
        },
        "outputId": "9fe0e56c-6e8e-4d33-c9f5-acd2ac7f486c"
      },
      "source": [
        "# 下載及安裝test-generator 方便檢查程式運行錯誤\n",
        "!pip install test-generator==0.1.1"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting test-generator==0.1.1\n",
            "  Downloading https://files.pythonhosted.org/packages/4a/52/e5eec4d926eb466844eaeeaac84af5372e946dd520fb2b6adf3388e620b0/test_generator-0.1.1-py2.py3-none-any.whl\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.6/dist-packages (from test-generator==0.1.1) (1.15.0)\n",
            "Installing collected packages: test-generator\n",
            "Successfully installed test-generator-0.1.1\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "imkMcaQYz-Tb",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "24843bf9-6fae-4c99-f6eb-11f206a31737"
      },
      "source": [
        "# 執行環境設定批次檔並將下載到的mobilenet-v1-1.0-224模型檔進行優化轉換產生IR(xml & bin)檔\n",
        "# 可以用 --name 指定下載到模型名稱\n",
        "!source /opt/intel/openvino/bin/setupvars.sh && \\\n",
        "python /opt/intel/openvino/deployment_tools/tools/model_downloader/converter.py \\\n",
        "--name mobilenet-v1-1.0-224"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[setupvars.sh] OpenVINO environment initialized\n",
            "========== Converting mobilenet-v1-1.0-224 to IR (FP16)\n",
            "Conversion command: /usr/bin/python3 -- /opt/intel/openvino_2020.4.287/deployment_tools/model_optimizer/mo.py --framework=caffe --data_type=FP16 --output_dir=/content/public/mobilenet-v1-1.0-224/FP16 --model_name=mobilenet-v1-1.0-224 '--input_shape=[1,3,224,224]' --input=data '--mean_values=data[103.94,116.78,123.68]' '--scale_values=data[58.8235294117647]' --output=prob --input_model=/content/public/mobilenet-v1-1.0-224/mobilenet-v1-1.0-224.caffemodel --input_proto=/content/public/mobilenet-v1-1.0-224/mobilenet-v1-1.0-224.prototxt\n",
            "\n",
            "Model Optimizer arguments:\n",
            "Common parameters:\n",
            "\t- Path to the Input Model: \t/content/public/mobilenet-v1-1.0-224/mobilenet-v1-1.0-224.caffemodel\n",
            "\t- Path for generated IR: \t/content/public/mobilenet-v1-1.0-224/FP16\n",
            "\t- IR output name: \tmobilenet-v1-1.0-224\n",
            "\t- Log level: \tERROR\n",
            "\t- Batch: \tNot specified, inherited from the model\n",
            "\t- Input layers: \tdata\n",
            "\t- Output layers: \tprob\n",
            "\t- Input shapes: \t[1,3,224,224]\n",
            "\t- Mean values: \tdata[103.94,116.78,123.68]\n",
            "\t- Scale values: \tdata[58.8235294117647]\n",
            "\t- Scale factor: \tNot specified\n",
            "\t- Precision of IR: \tFP16\n",
            "\t- Enable fusing: \tTrue\n",
            "\t- Enable grouped convolutions fusing: \tTrue\n",
            "\t- Move mean values to preprocess section: \tFalse\n",
            "\t- Reverse input channels: \tFalse\n",
            "Caffe specific parameters:\n",
            "\t- Path to Python Caffe* parser generated from caffe.proto: \t/opt/intel/openvino_2020.4.287/deployment_tools/model_optimizer/mo/front/caffe/proto\n",
            "\t- Enable resnet optimization: \tTrue\n",
            "\t- Path to the Input prototxt: \t/content/public/mobilenet-v1-1.0-224/mobilenet-v1-1.0-224.prototxt\n",
            "\t- Path to CustomLayersMapping.xml: \tDefault\n",
            "\t- Path to a mean file: \tNot specified\n",
            "\t- Offsets for a mean file: \tNot specified\n",
            "Model Optimizer version: \t\n",
            "\n",
            "[ SUCCESS ] Generated IR version 10 model.\n",
            "[ SUCCESS ] XML file: /content/public/mobilenet-v1-1.0-224/FP16/mobilenet-v1-1.0-224.xml\n",
            "[ SUCCESS ] BIN file: /content/public/mobilenet-v1-1.0-224/FP16/mobilenet-v1-1.0-224.bin\n",
            "[ SUCCESS ] Total execution time: 9.76 seconds. \n",
            "[ SUCCESS ] Memory consumed: 177 MB. \n",
            "\n",
            "========== Converting mobilenet-v1-1.0-224 to IR (FP32)\n",
            "Conversion command: /usr/bin/python3 -- /opt/intel/openvino_2020.4.287/deployment_tools/model_optimizer/mo.py --framework=caffe --data_type=FP32 --output_dir=/content/public/mobilenet-v1-1.0-224/FP32 --model_name=mobilenet-v1-1.0-224 '--input_shape=[1,3,224,224]' --input=data '--mean_values=data[103.94,116.78,123.68]' '--scale_values=data[58.8235294117647]' --output=prob --input_model=/content/public/mobilenet-v1-1.0-224/mobilenet-v1-1.0-224.caffemodel --input_proto=/content/public/mobilenet-v1-1.0-224/mobilenet-v1-1.0-224.prototxt\n",
            "\n",
            "Model Optimizer arguments:\n",
            "Common parameters:\n",
            "\t- Path to the Input Model: \t/content/public/mobilenet-v1-1.0-224/mobilenet-v1-1.0-224.caffemodel\n",
            "\t- Path for generated IR: \t/content/public/mobilenet-v1-1.0-224/FP32\n",
            "\t- IR output name: \tmobilenet-v1-1.0-224\n",
            "\t- Log level: \tERROR\n",
            "\t- Batch: \tNot specified, inherited from the model\n",
            "\t- Input layers: \tdata\n",
            "\t- Output layers: \tprob\n",
            "\t- Input shapes: \t[1,3,224,224]\n",
            "\t- Mean values: \tdata[103.94,116.78,123.68]\n",
            "\t- Scale values: \tdata[58.8235294117647]\n",
            "\t- Scale factor: \tNot specified\n",
            "\t- Precision of IR: \tFP32\n",
            "\t- Enable fusing: \tTrue\n",
            "\t- Enable grouped convolutions fusing: \tTrue\n",
            "\t- Move mean values to preprocess section: \tFalse\n",
            "\t- Reverse input channels: \tFalse\n",
            "Caffe specific parameters:\n",
            "\t- Path to Python Caffe* parser generated from caffe.proto: \t/opt/intel/openvino_2020.4.287/deployment_tools/model_optimizer/mo/front/caffe/proto\n",
            "\t- Enable resnet optimization: \tTrue\n",
            "\t- Path to the Input prototxt: \t/content/public/mobilenet-v1-1.0-224/mobilenet-v1-1.0-224.prototxt\n",
            "\t- Path to CustomLayersMapping.xml: \tDefault\n",
            "\t- Path to a mean file: \tNot specified\n",
            "\t- Offsets for a mean file: \tNot specified\n",
            "Model Optimizer version: \t\n",
            "\n",
            "[ SUCCESS ] Generated IR version 10 model.\n",
            "[ SUCCESS ] XML file: /content/public/mobilenet-v1-1.0-224/FP32/mobilenet-v1-1.0-224.xml\n",
            "[ SUCCESS ] BIN file: /content/public/mobilenet-v1-1.0-224/FP32/mobilenet-v1-1.0-224.bin\n",
            "[ SUCCESS ] Total execution time: 9.30 seconds. \n",
            "[ SUCCESS ] Memory consumed: 175 MB. \n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "izJpllcrzh1F",
        "colab_type": "text"
      },
      "source": [
        "# **4. 執行推論**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "s-y9DTUj0M07",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 179
        },
        "outputId": "2b1c41d2-92a1-44c4-8a10-319359dfffbc"
      },
      "source": [
        "# 以OpenCV檢視輸入影像\n",
        "import cv2\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "img = cv2.imread('/opt/intel/openvino/deployment_tools/demo/car.png') # 讀入指定彩色影像\n",
        "rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) # 將影像從BGR格式轉到RGB格式，才能讓plt.imshow()正確顯示\n",
        "plt.figure() # 準備一顯示影像空間\n",
        "plt.imshow(rgb) # 顯示影像"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fe300d73b38>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mlv7Hhqk0UZL",
        "colab_type": "text"
      },
      "source": [
        "執行環境設定批次檔並執行mobilenet影像分類\n",
        "\n",
        "環境設定批次檔 /opt/intel/openvino/bin/setupvars.sh\n",
        "\n",
        "影像分類範例 /opt/intel/openvino/inference_engine/samples/python/classification_sample/classification_sample.py\n",
        "\n",
        "參數：\n",
        "\n",
        "-m 模型名稱 (*.xml)\n",
        "\n",
        "-i 輸入影像名稱 (*.png, *.jpg ...)\n",
        "\n",
        "--labels ImageNet 1000分類名稱標籤檔"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KwFnVDvZ0N3O",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "a542c3b3-c644-4dd8-e42b-6493416972fe"
      },
      "source": [
        "# 檢視範例程式碼（可略過不執行）\n",
        "!cat /opt/intel/openvino/inference_engine/samples/python/classification_sample/classification_sample.py"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "#!/usr/bin/env python\n",
            "\"\"\"\n",
            " Copyright (C) 2018-2020 Intel Corporation\n",
            "\n",
            " Licensed under the Apache License, Version 2.0 (the \"License\");\n",
            " you may not use this file except in compliance with the License.\n",
            " You may obtain a copy of the License at\n",
            "\n",
            "      http://www.apache.org/licenses/LICENSE-2.0\n",
            "\n",
            " Unless required by applicable law or agreed to in writing, software\n",
            " distributed under the License is distributed on an \"AS IS\" BASIS,\n",
            " WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
            " See the License for the specific language governing permissions and\n",
            " limitations under the License.\n",
            "\"\"\"\n",
            "from __future__ import print_function\n",
            "import sys\n",
            "import os\n",
            "from argparse import ArgumentParser, SUPPRESS\n",
            "import cv2\n",
            "import numpy as np\n",
            "import logging as log\n",
            "from openvino.inference_engine import IECore\n",
            "\n",
            "\n",
            "def build_argparser():\n",
            "    parser = ArgumentParser(add_help=False)\n",
            "    args = parser.add_argument_group('Options')\n",
            "    args.add_argument('-h', '--help', action='help', default=SUPPRESS, help='Show this help message and exit.')\n",
            "    args.add_argument(\"-m\", \"--model\", help=\"Required. Path to an .xml file with a trained model.\", required=True,\n",
            "                      type=str)\n",
            "    args.add_argument(\"-i\", \"--input\", help=\"Required. Path to a folder with images or path to an image files\",\n",
            "                      required=True,\n",
            "                      type=str, nargs=\"+\")\n",
            "    args.add_argument(\"-l\", \"--cpu_extension\",\n",
            "                      help=\"Optional. Required for CPU custom layers. \"\n",
            "                           \"MKLDNN (CPU)-targeted custom layers. Absolute path to a shared library with the\"\n",
            "                           \" kernels implementations.\", type=str, default=None)\n",
            "    args.add_argument(\"-d\", \"--device\",\n",
            "                      help=\"Optional. Specify the target device to infer on; CPU, GPU, FPGA, HDDL, MYRIAD or HETERO: is \"\n",
            "                           \"acceptable. The sample will look for a suitable plugin for device specified. Default \"\n",
            "                           \"value is CPU\",\n",
            "                      default=\"CPU\", type=str)\n",
            "    args.add_argument(\"--labels\", help=\"Optional. Path to a labels mapping file\", default=None, type=str)\n",
            "    args.add_argument(\"-nt\", \"--number_top\", help=\"Optional. Number of top results\", default=10, type=int)\n",
            "\n",
            "    return parser\n",
            "\n",
            "\n",
            "def main():\n",
            "    log.basicConfig(format=\"[ %(levelname)s ] %(message)s\", level=log.INFO, stream=sys.stdout)\n",
            "    args = build_argparser().parse_args()\n",
            "    model_xml = args.model\n",
            "    model_bin = os.path.splitext(model_xml)[0] + \".bin\"\n",
            "\n",
            "    # Plugin initialization for specified device and load extensions library if specified\n",
            "    log.info(\"Creating Inference Engine\")\n",
            "    ie = IECore()\n",
            "    if args.cpu_extension and 'CPU' in args.device:\n",
            "        ie.add_extension(args.cpu_extension, \"CPU\")\n",
            "    # Read IR\n",
            "    log.info(\"Loading network files:\\n\\t{}\\n\\t{}\".format(model_xml, model_bin))\n",
            "    net = ie.read_network(model=model_xml, weights=model_bin)\n",
            "\n",
            "    if \"CPU\" in args.device:\n",
            "        supported_layers = ie.query_network(net, \"CPU\")\n",
            "        not_supported_layers = [l for l in net.layers.keys() if l not in supported_layers]\n",
            "        if len(not_supported_layers) != 0:\n",
            "            log.error(\"Following layers are not supported by the plugin for specified device {}:\\n {}\".\n",
            "                      format(args.device, ', '.join(not_supported_layers)))\n",
            "            log.error(\"Please try to specify cpu extensions library path in sample's command line parameters using -l \"\n",
            "                      \"or --cpu_extension command line argument\")\n",
            "            sys.exit(1)\n",
            "\n",
            "    assert len(net.input_info.keys()) == 1, \"Sample supports only single input topologies\"\n",
            "    assert len(net.outputs) == 1, \"Sample supports only single output topologies\"\n",
            "\n",
            "    log.info(\"Preparing input blobs\")\n",
            "    input_blob = next(iter(net.input_info))\n",
            "    out_blob = next(iter(net.outputs))\n",
            "    net.batch_size = len(args.input)\n",
            "\n",
            "    # Read and pre-process input images\n",
            "    n, c, h, w = net.input_info[input_blob].input_data.shape\n",
            "    images = np.ndarray(shape=(n, c, h, w))\n",
            "    for i in range(n):\n",
            "        image = cv2.imread(args.input[i])\n",
            "        if image.shape[:-1] != (h, w):\n",
            "            log.warning(\"Image {} is resized from {} to {}\".format(args.input[i], image.shape[:-1], (h, w)))\n",
            "            image = cv2.resize(image, (w, h))\n",
            "        image = image.transpose((2, 0, 1))  # Change data layout from HWC to CHW\n",
            "        images[i] = image\n",
            "    log.info(\"Batch size is {}\".format(n))\n",
            "\n",
            "    # Loading model to the plugin\n",
            "    log.info(\"Loading model to the plugin\")\n",
            "    exec_net = ie.load_network(network=net, device_name=args.device)\n",
            "\n",
            "    # Start sync inference\n",
            "    log.info(\"Starting inference in synchronous mode\")\n",
            "    res = exec_net.infer(inputs={input_blob: images})\n",
            "\n",
            "    # Processing output blob\n",
            "    log.info(\"Processing output blob\")\n",
            "    res = res[out_blob]\n",
            "    log.info(\"Top {} results: \".format(args.number_top))\n",
            "    if args.labels:\n",
            "        with open(args.labels, 'r') as f:\n",
            "            labels_map = [x.split(sep=' ', maxsplit=1)[-1].strip() for x in f]\n",
            "    else:\n",
            "        labels_map = None\n",
            "    classid_str = \"classid\"\n",
            "    probability_str = \"probability\"\n",
            "    for i, probs in enumerate(res):\n",
            "        probs = np.squeeze(probs)\n",
            "        top_ind = np.argsort(probs)[-args.number_top:][::-1]\n",
            "        print(\"Image {}\\n\".format(args.input[i]))\n",
            "        print(classid_str, probability_str)\n",
            "        print(\"{} {}\".format('-' * len(classid_str), '-' * len(probability_str)))\n",
            "        for id in top_ind:\n",
            "            det_label = labels_map[id] if labels_map else \"{}\".format(id)\n",
            "            label_length = len(det_label)\n",
            "            space_num_before = (len(classid_str) - label_length) // 2\n",
            "            space_num_after = len(classid_str) - (space_num_before + label_length) + 2\n",
            "            space_num_before_prob = (len(probability_str) - len(str(probs[id]))) // 2\n",
            "            print(\"{}{}{}{}{:.7f}\".format(' ' * space_num_before, det_label,\n",
            "                                          ' ' * space_num_after, ' ' * space_num_before_prob,\n",
            "                                          probs[id]))\n",
            "        print(\"\\n\")\n",
            "    log.info(\"This sample is an API example, for any performance measurements please use the dedicated benchmark_app tool\\n\")\n",
            "\n",
            "if __name__ == '__main__':\n",
            "    sys.exit(main() or 0)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mgHMBC5-0Zk2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 537
        },
        "outputId": "c42a7c39-098c-4b9d-d655-5afbea3b4e76"
      },
      "source": [
        "# 執行環境設定批次檔並執行影像分類\n",
        "# 輸出結果 car, sport car 置信度（機率） 0.4242041\n",
        "!source /opt/intel/openvino/bin/setupvars.sh && \\\n",
        "python \\\n",
        "/opt/intel/openvino/inference_engine/samples/python/classification_sample/classification_sample.py \\\n",
        "-m /content/public/mobilenet-v1-1.0-224/FP16/mobilenet-v1-1.0-224.xml \\\n",
        "-i /opt/intel/openvino/deployment_tools/demo/car.png \\\n",
        "--labels /opt/intel/openvino/deployment_tools/demo/squeezenet1.1.labels"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[setupvars.sh] OpenVINO environment initialized\n",
            "[ INFO ] Creating Inference Engine\n",
            "[ INFO ] Loading network files:\n",
            "\t/content/public/mobilenet-v1-1.0-224/FP16/mobilenet-v1-1.0-224.xml\n",
            "\t/content/public/mobilenet-v1-1.0-224/FP16/mobilenet-v1-1.0-224.bin\n",
            "[ INFO ] Preparing input blobs\n",
            "[ WARNING ] Image /opt/intel/openvino/deployment_tools/demo/car.png is resized from (259, 787) to (224, 224)\n",
            "[ INFO ] Batch size is 1\n",
            "[ INFO ] Loading model to the plugin\n",
            "[ INFO ] Starting inference in synchronous mode\n",
            "[ INFO ] Processing output blob\n",
            "[ INFO ] Top 10 results: \n",
            "Image /opt/intel/openvino/deployment_tools/demo/car.png\n",
            "\n",
            "classid probability\n",
            "------- -----------\n",
            "car, sport car0.4242041\n",
            "convertible0.2728059\n",
            " wheel   0.1330034\n",
            "wagon, station wagon, wagon, estate car, beach waggon, station waggon, waggon 0.0699498\n",
            "race car, racing car0.0339604\n",
            "radiator grille0.0263465\n",
            "pickup truck0.0155464\n",
            "hack, taxi, taxicab0.0127090\n",
            "minivan  0.0051457\n",
            "   T     0.0018786\n",
            "\n",
            "\n",
            "[ INFO ] This sample is an API example, for any performance measurements please use the dedicated benchmark_app tool\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}