code/46-语义分割.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Untitled0.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "PWaz7EsiS44f",
        "outputId": "61e5bcb2-bdf5-43ab-c40c-7c962f1244df"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: torch in /usr/local/lib/python3.7/dist-packages (1.10.0+cu111)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.7/dist-packages (from torch) (3.10.0.2)\n",
            "Collecting d2l\n",
            "  Downloading d2l-0.17.3-py3-none-any.whl (82 kB)\n",
            "\u001b[K     |████████████████████████████████| 82 kB 561 kB/s \n",
            "\u001b[?25hCollecting matplotlib==3.3.3\n",
            "  Downloading matplotlib-3.3.3-cp37-cp37m-manylinux1_x86_64.whl (11.6 MB)\n",
            "\u001b[K     |████████████████████████████████| 11.6 MB 17.1 MB/s \n",
            "\u001b[?25hCollecting requests==2.25.1\n",
            "  Downloading requests-2.25.1-py2.py3-none-any.whl (61 kB)\n",
            "\u001b[K     |████████████████████████████████| 61 kB 8.7 MB/s \n",
            "\u001b[?25hCollecting pandas==1.2.2\n",
            "  Downloading pandas-1.2.2-cp37-cp37m-manylinux1_x86_64.whl (9.9 MB)\n",
            "\u001b[K     |████████████████████████████████| 9.9 MB 20.8 MB/s \n",
            "\u001b[?25hCollecting numpy==1.18.5\n",
            "  Downloading numpy-1.18.5-cp37-cp37m-manylinux1_x86_64.whl (20.1 MB)\n",
            "\u001b[K     |████████████████████████████████| 20.1 MB 1.3 MB/s \n",
            "\u001b[?25hRequirement already satisfied: jupyter==1.0.0 in /usr/local/lib/python3.7/dist-packages (from d2l) (1.0.0)\n",
            "Requirement already satisfied: nbconvert in /usr/local/lib/python3.7/dist-packages (from jupyter==1.0.0->d2l) (5.6.1)\n",
            "Requirement already satisfied: ipykernel in /usr/local/lib/python3.7/dist-packages (from jupyter==1.0.0->d2l) (4.10.1)\n",
            "Requirement already satisfied: jupyter-console in /usr/local/lib/python3.7/dist-packages (from jupyter==1.0.0->d2l) (5.2.0)\n",
            "Requirement already satisfied: ipywidgets in /usr/local/lib/python3.7/dist-packages (from jupyter==1.0.0->d2l) (7.6.5)\n",
            "Requirement already satisfied: notebook in /usr/local/lib/python3.7/dist-packages (from jupyter==1.0.0->d2l) (5.3.1)\n",
            "Requirement already satisfied: qtconsole in /usr/local/lib/python3.7/dist-packages (from jupyter==1.0.0->d2l) (5.2.2)\n",
            "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib==3.3.3->d2l) (2.8.2)\n",
            "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.3 in /usr/local/lib/python3.7/dist-packages (from matplotlib==3.3.3->d2l) (3.0.7)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib==3.3.3->d2l) (0.11.0)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib==3.3.3->d2l) (1.3.2)\n",
            "Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.7/dist-packages (from matplotlib==3.3.3->d2l) (7.1.2)\n",
            "Requirement already satisfied: pytz>=2017.3 in /usr/local/lib/python3.7/dist-packages (from pandas==1.2.2->d2l) (2018.9)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.7/dist-packages (from requests==2.25.1->d2l) (2.10)\n",
            "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /usr/local/lib/python3.7/dist-packages (from requests==2.25.1->d2l) (1.24.3)\n",
            "Requirement already satisfied: chardet<5,>=3.0.2 in /usr/local/lib/python3.7/dist-packages (from requests==2.25.1->d2l) (3.0.4)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.7/dist-packages (from requests==2.25.1->d2l) (2021.10.8)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.1->matplotlib==3.3.3->d2l) (1.15.0)\n",
            "Requirement already satisfied: traitlets>=4.1.0 in /usr/local/lib/python3.7/dist-packages (from ipykernel->jupyter==1.0.0->d2l) (5.1.1)\n",
            "Requirement already satisfied: ipython>=4.0.0 in /usr/local/lib/python3.7/dist-packages (from ipykernel->jupyter==1.0.0->d2l) (5.5.0)\n",
            "Requirement already satisfied: tornado>=4.0 in /usr/local/lib/python3.7/dist-packages (from ipykernel->jupyter==1.0.0->d2l) (5.1.1)\n",
            "Requirement already satisfied: jupyter-client in /usr/local/lib/python3.7/dist-packages (from ipykernel->jupyter==1.0.0->d2l) (5.3.5)\n",
            "Requirement already satisfied: pygments in /usr/local/lib/python3.7/dist-packages (from ipython>=4.0.0->ipykernel->jupyter==1.0.0->d2l) (2.6.1)\n",
            "Requirement already satisfied: simplegeneric>0.8 in /usr/local/lib/python3.7/dist-packages (from ipython>=4.0.0->ipykernel->jupyter==1.0.0->d2l) (0.8.1)\n",
            "Requirement already satisfied: decorator in /usr/local/lib/python3.7/dist-packages (from ipython>=4.0.0->ipykernel->jupyter==1.0.0->d2l) (4.4.2)\n",
            "Requirement already satisfied: setuptools>=18.5 in /usr/local/lib/python3.7/dist-packages (from ipython>=4.0.0->ipykernel->jupyter==1.0.0->d2l) (57.4.0)\n",
            "Requirement already satisfied: pickleshare in /usr/local/lib/python3.7/dist-packages (from ipython>=4.0.0->ipykernel->jupyter==1.0.0->d2l) (0.7.5)\n",
            "Requirement already satisfied: prompt-toolkit<2.0.0,>=1.0.4 in /usr/local/lib/python3.7/dist-packages (from ipython>=4.0.0->ipykernel->jupyter==1.0.0->d2l) (1.0.18)\n",
            "Requirement already satisfied: pexpect in /usr/local/lib/python3.7/dist-packages (from ipython>=4.0.0->ipykernel->jupyter==1.0.0->d2l) (4.8.0)\n",
            "Requirement already satisfied: wcwidth in /usr/local/lib/python3.7/dist-packages (from prompt-toolkit<2.0.0,>=1.0.4->ipython>=4.0.0->ipykernel->jupyter==1.0.0->d2l) (0.2.5)\n",
            "Requirement already satisfied: ipython-genutils~=0.2.0 in /usr/local/lib/python3.7/dist-packages (from ipywidgets->jupyter==1.0.0->d2l) (0.2.0)\n",
            "Requirement already satisfied: nbformat>=4.2.0 in /usr/local/lib/python3.7/dist-packages (from ipywidgets->jupyter==1.0.0->d2l) (5.1.3)\n",
            "Requirement already satisfied: jupyterlab-widgets>=1.0.0 in /usr/local/lib/python3.7/dist-packages (from ipywidgets->jupyter==1.0.0->d2l) (1.0.2)\n",
            "Requirement already satisfied: widgetsnbextension~=3.5.0 in /usr/local/lib/python3.7/dist-packages (from ipywidgets->jupyter==1.0.0->d2l) (3.5.2)\n",
            "Requirement already satisfied: jsonschema!=2.5.0,>=2.4 in /usr/local/lib/python3.7/dist-packages (from nbformat>=4.2.0->ipywidgets->jupyter==1.0.0->d2l) (4.3.3)\n",
            "Requirement already satisfied: jupyter-core in /usr/local/lib/python3.7/dist-packages (from nbformat>=4.2.0->ipywidgets->jupyter==1.0.0->d2l) (4.9.1)\n",
            "Requirement already satisfied: importlib-metadata in /usr/local/lib/python3.7/dist-packages (from jsonschema!=2.5.0,>=2.4->nbformat>=4.2.0->ipywidgets->jupyter==1.0.0->d2l) (4.10.1)\n",
            "Requirement already satisfied: importlib-resources>=1.4.0 in /usr/local/lib/python3.7/dist-packages (from jsonschema!=2.5.0,>=2.4->nbformat>=4.2.0->ipywidgets->jupyter==1.0.0->d2l) (5.4.0)\n",
            "Requirement already satisfied: attrs>=17.4.0 in /usr/local/lib/python3.7/dist-packages (from jsonschema!=2.5.0,>=2.4->nbformat>=4.2.0->ipywidgets->jupyter==1.0.0->d2l) (21.4.0)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.7/dist-packages (from jsonschema!=2.5.0,>=2.4->nbformat>=4.2.0->ipywidgets->jupyter==1.0.0->d2l) (3.10.0.2)\n",
            "Requirement already satisfied: pyrsistent!=0.17.0,!=0.17.1,!=0.17.2,>=0.14.0 in /usr/local/lib/python3.7/dist-packages (from jsonschema!=2.5.0,>=2.4->nbformat>=4.2.0->ipywidgets->jupyter==1.0.0->d2l) (0.18.1)\n",
            "Requirement already satisfied: zipp>=3.1.0 in /usr/local/lib/python3.7/dist-packages (from importlib-resources>=1.4.0->jsonschema!=2.5.0,>=2.4->nbformat>=4.2.0->ipywidgets->jupyter==1.0.0->d2l) (3.7.0)\n",
            "Requirement already satisfied: Send2Trash in /usr/local/lib/python3.7/dist-packages (from notebook->jupyter==1.0.0->d2l) (1.8.0)\n",
            "Requirement already satisfied: terminado>=0.8.1 in /usr/local/lib/python3.7/dist-packages (from notebook->jupyter==1.0.0->d2l) (0.13.1)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.7/dist-packages (from notebook->jupyter==1.0.0->d2l) (2.11.3)\n",
            "Requirement already satisfied: pyzmq>=13 in /usr/local/lib/python3.7/dist-packages (from jupyter-client->ipykernel->jupyter==1.0.0->d2l) (22.3.0)\n",
            "Requirement already satisfied: ptyprocess in /usr/local/lib/python3.7/dist-packages (from terminado>=0.8.1->notebook->jupyter==1.0.0->d2l) (0.7.0)\n",
            "Requirement already satisfied: MarkupSafe>=0.23 in /usr/local/lib/python3.7/dist-packages (from jinja2->notebook->jupyter==1.0.0->d2l) (2.0.1)\n",
            "Requirement already satisfied: defusedxml in /usr/local/lib/python3.7/dist-packages (from nbconvert->jupyter==1.0.0->d2l) (0.7.1)\n",
            "Requirement already satisfied: mistune<2,>=0.8.1 in /usr/local/lib/python3.7/dist-packages (from nbconvert->jupyter==1.0.0->d2l) (0.8.4)\n",
            "Requirement already satisfied: entrypoints>=0.2.2 in /usr/local/lib/python3.7/dist-packages (from nbconvert->jupyter==1.0.0->d2l) (0.4)\n",
            "Requirement already satisfied: pandocfilters>=1.4.1 in /usr/local/lib/python3.7/dist-packages (from nbconvert->jupyter==1.0.0->d2l) (1.5.0)\n",
            "Requirement already satisfied: testpath in /usr/local/lib/python3.7/dist-packages (from nbconvert->jupyter==1.0.0->d2l) (0.5.0)\n",
            "Requirement already satisfied: bleach in /usr/local/lib/python3.7/dist-packages (from nbconvert->jupyter==1.0.0->d2l) (4.1.0)\n",
            "Requirement already satisfied: packaging in /usr/local/lib/python3.7/dist-packages (from bleach->nbconvert->jupyter==1.0.0->d2l) (21.3)\n",
            "Requirement already satisfied: webencodings in /usr/local/lib/python3.7/dist-packages (from bleach->nbconvert->jupyter==1.0.0->d2l) (0.5.1)\n",
            "Requirement already satisfied: qtpy in /usr/local/lib/python3.7/dist-packages (from qtconsole->jupyter==1.0.0->d2l) (2.0.1)\n",
            "Installing collected packages: numpy, requests, pandas, matplotlib, d2l\n",
            "  Attempting uninstall: numpy\n",
            "    Found existing installation: numpy 1.19.5\n",
            "    Uninstalling numpy-1.19.5:\n",
            "      Successfully uninstalled numpy-1.19.5\n",
            "  Attempting uninstall: requests\n",
            "    Found existing installation: requests 2.23.0\n",
            "    Uninstalling requests-2.23.0:\n",
            "      Successfully uninstalled requests-2.23.0\n",
            "  Attempting uninstall: pandas\n",
            "    Found existing installation: pandas 1.3.5\n",
            "    Uninstalling pandas-1.3.5:\n",
            "      Successfully uninstalled pandas-1.3.5\n",
            "  Attempting uninstall: matplotlib\n",
            "    Found existing installation: matplotlib 3.2.2\n",
            "    Uninstalling matplotlib-3.2.2:\n",
            "      Successfully uninstalled matplotlib-3.2.2\n",
            "\u001b[31mERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.\n",
            "google-colab 1.0.0 requires requests~=2.23.0, but you have requests 2.25.1 which is incompatible.\n",
            "datascience 0.10.6 requires folium==0.2.1, but you have folium 0.8.3 which is incompatible.\n",
            "albumentations 0.1.12 requires imgaug<0.2.7,>=0.2.5, but you have imgaug 0.2.9 which is incompatible.\u001b[0m\n",
            "Successfully installed d2l-0.17.3 matplotlib-3.3.3 numpy-1.18.5 pandas-1.2.2 requests-2.25.1\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.colab-display-data+json": {
              "pip_warning": {
                "packages": [
                  "matplotlib",
                  "mpl_toolkits",
                  "numpy",
                  "pandas"
                ]
              }
            }
          },
          "metadata": {}
        }
      ],
      "source": [
        "!pip3 install torch\n",
        "!pip3 install d2l"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%matplotlib inline\n",
        "import os\n",
        "import torch\n",
        "import torchvision\n",
        "from d2l import torch as d2l\n",
        "\n",
        "#最重要的语义分割数据集之一是Pascal VOC2012,之后的VOC都是2012基础上的改动\n",
        "d2l.DATA_HUB['voc2012'] = (d2l.DATA_URL + 'VOCtrainval_11-May-2012.tar',\n",
        "                           '4e443f8a2eca6b1dac8a6c57641b67dd40621a49')\n",
        "\n",
        "voc_dir = d2l.download_extract('voc2012', 'VOCdevkit/VOC2012')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0151QjwoTelN",
        "outputId": "5d3b8a24-3b50-48ec-c869-bc7f4aa0c993"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading ../data/VOCtrainval_11-May-2012.tar from http://d2l-data.s3-accelerate.amazonaws.com/VOCtrainval_11-May-2012.tar...\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#把所有图片读入内存\n",
        "def read_voc_images(voc_dir, is_train=True):\n",
        "    \"\"\"读取所有VOC图像并标注\"\"\"\n",
        "    txt_fname = os.path.join(voc_dir, 'ImageSets', 'Segmentation',\n",
        "                             'train.txt' if is_train else 'val.txt')\n",
        "    mode = torchvision.io.image.ImageReadMode.RGB\n",
        "    with open(txt_fname, 'r') as f:\n",
        "        images = f.read().split()\n",
        "    features, labels = [], []\n",
        "    for i, fname in enumerate(images):\n",
        "        features.append(torchvision.io.read_image(os.path.join(\n",
        "            voc_dir, 'JPEGImages', f'{fname}.jpg')))#根文件下JPEGImages为原始图片，用于训练\n",
        "        labels.append(torchvision.io.read_image(os.path.join(\n",
        "            voc_dir, 'SegmentationClass' ,f'{fname}.png'), mode))#语义分割需要对每个像素有label,也存成图片\n",
        "    return features, labels\n",
        "\n",
        "train_features, train_labels = read_voc_images(voc_dir, True)"
      ],
      "metadata": {
        "id": "lO1Pjwb6Vsss"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#展示一些图片和对应label\n",
        "n = 5\n",
        "imgs = train_features[0:n] + train_labels[0:n]\n",
        "imgs = [img.permute(1,2,0) for img in imgs]\n",
        "d2l.show_images(imgs, 2, n);"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 885
        },
        "id": "XFaYPwdPTa7-",
        "outputId": "c0dbc2e1-6677-41de-d0e1-007948116fd9"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "error",
          "ename": "ImportError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mImportError\u001b[0m                               Traceback (most recent call last)",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/IPython/core/formatters.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, obj)\u001b[0m\n\u001b[1;32m    332\u001b[0m                 \u001b[0;32mpass\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    333\u001b[0m             \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 334\u001b[0;31m                 \u001b[0;32mreturn\u001b[0m \u001b[0mprinter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    335\u001b[0m             \u001b[0;31m# Finally look for special method names\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    336\u001b[0m             \u001b[0mmethod\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_real_method\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprint_method\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/IPython/core/pylabtools.py\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m(fig)\u001b[0m\n\u001b[1;32m    239\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    240\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0;34m'png'\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mformats\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 241\u001b[0;31m         \u001b[0mpng_formatter\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfor_type\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mFigure\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mlambda\u001b[0m \u001b[0mfig\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mprint_figure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfig\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'png'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    242\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0;34m'retina'\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mformats\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0;34m'png2x'\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mformats\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    243\u001b[0m         \u001b[0mpng_formatter\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfor_type\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mFigure\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mlambda\u001b[0m \u001b[0mfig\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mretina_figure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfig\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/IPython/core/pylabtools.py\u001b[0m in \u001b[0;36mprint_figure\u001b[0;34m(fig, fmt, bbox_inches, **kwargs)\u001b[0m\n\u001b[1;32m    123\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    124\u001b[0m     \u001b[0mbytes_io\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mBytesIO\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 125\u001b[0;31m     \u001b[0mfig\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcanvas\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprint_figure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbytes_io\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkw\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    126\u001b[0m     \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbytes_io\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgetvalue\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    127\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mfmt\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'svg'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/matplotlib/backend_bases.py\u001b[0m in \u001b[0;36mprint_figure\u001b[0;34m(self, filename, dpi, facecolor, edgecolor, orientation, format, bbox_inches, **kwargs)\u001b[0m\n\u001b[1;32m   2092\u001b[0m         \u001b[0mhardcopy\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2093\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2094\u001b[0;31m         \u001b[0mParameters\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2095\u001b[0m         \u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2096\u001b[0m         \u001b[0mfilename\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mstr\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mpath\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mlike\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mfile\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mlike\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/matplotlib/backend_bases.py\u001b[0m in \u001b[0;36m_get_renderer\u001b[0;34m(figure, print_method)\u001b[0m\n\u001b[1;32m   1558\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mrenderer\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1559\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1560\u001b[0;31m             raise RuntimeError(f\"{print_method} did not call Figure.draw, so \"\n\u001b[0m\u001b[1;32m   1561\u001b[0m                                f\"no renderer is available\")\n\u001b[1;32m   1562\u001b[0m         \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/matplotlib/backends/backend_agg.py\u001b[0m in \u001b[0;36mprint_png\u001b[0;34m(self, filename_or_obj, metadata, pil_kwargs, *args, **kwargs)\u001b[0m\n\u001b[1;32m    503\u001b[0m         \u001b[0mpil_kwargs\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mdict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moptional\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    504\u001b[0m             \u001b[0mKeyword\u001b[0m \u001b[0marguments\u001b[0m \u001b[0mpassed\u001b[0m \u001b[0mto\u001b[0m\u001b[0;31m \u001b[0m\u001b[0;31m`\u001b[0m\u001b[0mPIL\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mImage\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mImage\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msave\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 505\u001b[0;31m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    506\u001b[0m             \u001b[0mIf\u001b[0m \u001b[0mthe\u001b[0m \u001b[0;34m'pnginfo'\u001b[0m \u001b[0mkey\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0mpresent\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mit\u001b[0m \u001b[0mcompletely\u001b[0m \u001b[0moverrides\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    507\u001b[0m             \u001b[0;34m*\u001b[0m\u001b[0mmetadata\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mincluding\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mdefault\u001b[0m \u001b[0;34m'Software'\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mImportError\u001b[0m: cannot import name '_png' from 'matplotlib' (/usr/local/lib/python3.7/dist-packages/matplotlib/__init__.py)"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 540x216 with 10 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#用不同像素的颜色表示不同的label\n",
        "VOC_COLORMAP = [[0, 0, 0], [128, 0, 0], [0, 128, 0], [128, 128, 0],\n",
        "                [0, 0, 128], [128, 0, 128], [0, 128, 128], [128, 128, 128],\n",
        "                [64, 0, 0], [192, 0, 0], [64, 128, 0], [192, 128, 0],\n",
        "                [64, 0, 128], [192, 0, 128], [64, 128, 128], [192, 128, 128],\n",
        "                [0, 64, 0], [128, 64, 0], [0, 192, 0], [128, 192, 0],\n",
        "                [0, 64, 128]]\n",
        "\n",
        "\n",
        "VOC_CLASSES = ['background', 'aeroplane', 'bicycle', 'bird', 'boat',\n",
        "               'bottle', 'bus', 'car', 'cat', 'chair', 'cow',\n",
        "               'diningtable', 'dog', 'horse', 'motorbike', 'person',\n",
        "               'potted plant', 'sheep', 'sofa', 'train', 'tv/monitor']"
      ],
      "metadata": {
        "id": "9C0e4K-ZVsbK"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def voc_colormap2label():\n",
        "    \"\"\"构建从RGB到VOC类别索引的映射\"\"\"#把每个RGBlabel对应一个整数\n",
        "    colormap2label = torch.zeros(256 ** 3, dtype=torch.long)\n",
        "    for i, colormap in enumerate(VOC_COLORMAP):\n",
        "        colormap2label[\n",
        "            (colormap[0] * 256 + colormap[1]) * 256 + colormap[2]] = i\n",
        "    return colormap2label\n",
        "\n",
        "\n",
        "def voc_label_indices(colormap, colormap2label):\n",
        "    \"\"\"将VOC标签中的RGB值映射到它们的类别索引\"\"\"\n",
        "    colormap = colormap.permute(1, 2, 0).numpy().astype('int32')\n",
        "    idx = ((colormap[:, :, 0] * 256 + colormap[:, :, 1]) * 256\n",
        "           + colormap[:, :, 2])\n",
        "    return colormap2label[idx]"
      ],
      "metadata": {
        "id": "FXMKAP5YVxKX"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y = voc_label_indices(train_labels[0], voc_colormap2label())\n",
        "y[105:115, 130:140], VOC_CLASSES[1]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ii8EvpdsV7xC",
        "outputId": "4f6a2d73-f022-4f59-b2ab-ce13fd5ba044"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(tensor([[0, 0, 0, 0, 0, 0, 0, 0, 0, 1],\n",
              "         [0, 0, 0, 0, 0, 0, 0, 1, 1, 1],\n",
              "         [0, 0, 0, 0, 0, 0, 1, 1, 1, 1],\n",
              "         [0, 0, 0, 0, 0, 1, 1, 1, 1, 1],\n",
              "         [0, 0, 0, 0, 0, 1, 1, 1, 1, 1],\n",
              "         [0, 0, 0, 0, 1, 1, 1, 1, 1, 1],\n",
              "         [0, 0, 0, 0, 0, 1, 1, 1, 1, 1],\n",
              "         [0, 0, 0, 0, 0, 1, 1, 1, 1, 1],\n",
              "         [0, 0, 0, 0, 0, 0, 1, 1, 1, 1],\n",
              "         [0, 0, 0, 0, 0, 0, 0, 0, 1, 1]]), 'aeroplane')"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#图片增广\n",
        "def voc_rand_crop(feature, label, height, width):\n",
        "    \"\"\"随机裁剪特征和标签图像\"\"\"\n",
        "    rect = torchvision.transforms.RandomCrop.get_params(\n",
        "        feature, (height, width))#get_params()返回RandomCrop随机得到的框\n",
        "    feature = torchvision.transforms.functional.crop(feature, *rect)\n",
        "    label = torchvision.transforms.functional.crop(label, *rect)#用这个框分别切feature和label\n",
        "    return feature, label\n",
        "\n",
        "imgs = []\n",
        "for _ in range(n):\n",
        "    imgs += voc_rand_crop(train_features[0], train_labels[0], 200, 300)\n",
        "\n",
        "imgs = [img.permute(1, 2, 0) for img in imgs]\n",
        "d2l.show_images(imgs[::2] + imgs[1::2], 2, n);#展示效果"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 392
        },
        "id": "702H4XnBWApq",
        "outputId": "7e782bd9-85d6-415c-e782-afd76324f37a"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "error",
          "ename": "ImportError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mImportError\u001b[0m                               Traceback (most recent call last)",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/IPython/core/formatters.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, obj)\u001b[0m\n\u001b[1;32m    332\u001b[0m                 \u001b[0;32mpass\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    333\u001b[0m             \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 334\u001b[0;31m                 \u001b[0;32mreturn\u001b[0m \u001b[0mprinter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    335\u001b[0m             \u001b[0;31m# Finally look for special method names\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    336\u001b[0m             \u001b[0mmethod\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_real_method\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprint_method\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/IPython/core/pylabtools.py\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m(fig)\u001b[0m\n\u001b[1;32m    239\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    240\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0;34m'png'\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mformats\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 241\u001b[0;31m         \u001b[0mpng_formatter\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfor_type\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mFigure\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mlambda\u001b[0m \u001b[0mfig\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mprint_figure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfig\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'png'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    242\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0;34m'retina'\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mformats\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0;34m'png2x'\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mformats\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    243\u001b[0m         \u001b[0mpng_formatter\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfor_type\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mFigure\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mlambda\u001b[0m \u001b[0mfig\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mretina_figure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfig\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/IPython/core/pylabtools.py\u001b[0m in \u001b[0;36mprint_figure\u001b[0;34m(fig, fmt, bbox_inches, **kwargs)\u001b[0m\n\u001b[1;32m    123\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    124\u001b[0m     \u001b[0mbytes_io\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mBytesIO\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 125\u001b[0;31m     \u001b[0mfig\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcanvas\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprint_figure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbytes_io\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkw\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    126\u001b[0m     \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbytes_io\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgetvalue\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    127\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mfmt\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'svg'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/matplotlib/backend_bases.py\u001b[0m in \u001b[0;36mprint_figure\u001b[0;34m(self, filename, dpi, facecolor, edgecolor, orientation, format, bbox_inches, **kwargs)\u001b[0m\n\u001b[1;32m   2092\u001b[0m         \u001b[0mhardcopy\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2093\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2094\u001b[0;31m         \u001b[0mParameters\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2095\u001b[0m         \u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2096\u001b[0m         \u001b[0mfilename\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mstr\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mpath\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mlike\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mfile\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mlike\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/matplotlib/backend_bases.py\u001b[0m in \u001b[0;36m_get_renderer\u001b[0;34m(figure, print_method)\u001b[0m\n\u001b[1;32m   1558\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mrenderer\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1559\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1560\u001b[0;31m             raise RuntimeError(f\"{print_method} did not call Figure.draw, so \"\n\u001b[0m\u001b[1;32m   1561\u001b[0m                                f\"no renderer is available\")\n\u001b[1;32m   1562\u001b[0m         \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/matplotlib/backends/backend_agg.py\u001b[0m in \u001b[0;36mprint_png\u001b[0;34m(self, filename_or_obj, metadata, pil_kwargs, *args, **kwargs)\u001b[0m\n\u001b[1;32m    503\u001b[0m         \u001b[0mpil_kwargs\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mdict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moptional\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    504\u001b[0m             \u001b[0mKeyword\u001b[0m \u001b[0marguments\u001b[0m \u001b[0mpassed\u001b[0m \u001b[0mto\u001b[0m\u001b[0;31m \u001b[0m\u001b[0;31m`\u001b[0m\u001b[0mPIL\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mImage\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mImage\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msave\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 505\u001b[0;31m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    506\u001b[0m             \u001b[0mIf\u001b[0m \u001b[0mthe\u001b[0m \u001b[0;34m'pnginfo'\u001b[0m \u001b[0mkey\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0mpresent\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mit\u001b[0m \u001b[0mcompletely\u001b[0m \u001b[0moverrides\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    507\u001b[0m             \u001b[0;34m*\u001b[0m\u001b[0mmetadata\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mincluding\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mdefault\u001b[0m \u001b[0;34m'Software'\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mImportError\u001b[0m: cannot import name '_png' from 'matplotlib' (/usr/local/lib/python3.7/dist-packages/matplotlib/__init__.py)"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 540x216 with 10 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class VOCSegDataset(torch.utils.data.Dataset):\n",
        "    \"\"\"一个用于加载VOC数据集的自定义数据集\"\"\"\n",
        "\n",
        "    def __init__(self, is_train, crop_size, voc_dir):\n",
        "        self.transform = torchvision.transforms.Normalize(\n",
        "            mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])\n",
        "        self.crop_size = crop_size#用crop不用拉伸是因为拉伸需要插值而对label插值操作很麻烦\n",
        "        features, labels = read_voc_images(voc_dir, is_train=is_train)\n",
        "        self.features = [self.normalize_image(feature)\n",
        "                         for feature in self.filter(features)]\n",
        "        self.labels = self.filter(labels)\n",
        "        self.colormap2label = voc_colormap2label()\n",
        "        print('read ' + str(len(self.features)) + ' examples')\n",
        "\n",
        "    def normalize_image(self, img):\n",
        "        return self.transform(img.float() / 255)\n",
        "\n",
        "    def filter(self, imgs):#如果图片尺寸比裁剪大小还小就舍弃掉（也可以做padding）\n",
        "        return [img for img in imgs if (\n",
        "            img.shape[1] >= self.crop_size[0] and\n",
        "            img.shape[2] >= self.crop_size[1])]\n",
        "\n",
        "    def __getitem__(self, idx):\n",
        "        feature, label = voc_rand_crop(self.features[idx], self.labels[idx],\n",
        "                                       *self.crop_size)\n",
        "        return (feature, voc_label_indices(label, self.colormap2label))#RGBlabellabel转index label\n",
        "\n",
        "    def __len__(self):\n",
        "        return len(self.features)"
      ],
      "metadata": {
        "id": "FCL-lddxq9_A"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#读取数据集\n",
        "crop_size = (320, 480)\n",
        "voc_train = VOCSegDataset(True, crop_size, voc_dir)\n",
        "voc_test = VOCSegDataset(False, crop_size, voc_dir)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "M3KOdUMSr4Yz",
        "outputId": "ff0077a1-4228-4f10-d56f-a403081c94ad"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "read 1114 examples\n",
            "read 1078 examples\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "batch_size = 64\n",
        "train_iter = torch.utils.data.DataLoader(voc_train, batch_size, shuffle=True,\n",
        "                                    drop_last=True,\n",
        "                                    num_workers=d2l.get_dataloader_workers())\n",
        "for X, Y in train_iter:\n",
        "    print(X.shape)\n",
        "    print(Y.shape)\n",
        "    break"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cQOAOa6avo9A",
        "outputId": "dc0e6ddf-8101-43f3-f865-c1e64e7a89e6"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py:481: UserWarning: This DataLoader will create 4 worker processes in total. Our suggested max number of worker in current system is 2, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.\n",
            "  cpuset_checked))\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "torch.Size([64, 3, 320, 480])\n",
            "torch.Size([64, 320, 480])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#整合所有组件\n",
        "def load_data_voc(batch_size, crop_size):\n",
        "    \"\"\"加载VOC语义分割数据集\"\"\"\n",
        "    voc_dir = d2l.download_extract('voc2012', os.path.join(\n",
        "        'VOCdevkit', 'VOC2012'))\n",
        "    num_workers = d2l.get_dataloader_workers()\n",
        "    train_iter = torch.utils.data.DataLoader(\n",
        "        VOCSegDataset(True, crop_size, voc_dir), batch_size,\n",
        "        shuffle=True, drop_last=True, num_workers=num_workers)\n",
        "    test_iter = torch.utils.data.DataLoader(\n",
        "        VOCSegDataset(False, crop_size, voc_dir), batch_size,\n",
        "        drop_last=True, num_workers=num_workers)\n",
        "    return train_iter, test_iter"
      ],
      "metadata": {
        "id": "Tgd9VvqSwMlt"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}