Resnet_pytorch.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "goeLmKFXDEM5"
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "i9hBUOU5kwzf",
    "outputId": "78fc62e1-b005-4ae6-bc4f-3daaccf0b318"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[-2.8201e-23,  3.0715e-41,  7.0065e-44,  6.8664e-44],\n",
      "        [ 6.3058e-44,  6.7262e-44,  7.4269e-44,  6.3058e-44],\n",
      "        [ 6.7262e-44,  7.9874e-44,  1.1771e-43,  6.7262e-44],\n",
      "        [ 7.4269e-44,  8.1275e-44,  7.0065e-44,  7.0065e-44],\n",
      "        [ 8.1275e-44,  7.2868e-44,  7.8473e-44,  6.4460e-44]])\n"
     ]
    }
   ],
   "source": [
    "x = torch.empty(5,4)\n",
    "print(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "eJLerQSlk_pG",
    "outputId": "17df1637-ae82-42cd-a508-40deb3318256"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[0.3976, 0.2362, 0.3250, 0.7433, 0.3457, 0.1334, 0.6145, 0.2543],\n",
      "        [0.3360, 0.4803, 0.8231, 0.5072, 0.8074, 0.8740, 0.1817, 0.4878],\n",
      "        [0.2481, 0.4676, 0.5134, 0.4437, 0.7316, 0.7485, 0.7084, 0.6157],\n",
      "        [0.9833, 0.1059, 0.5228, 0.3079, 0.9208, 0.8198, 0.8263, 0.9913],\n",
      "        [0.4339, 0.9590, 0.0707, 0.6564, 0.6033, 0.1882, 0.8471, 0.4325],\n",
      "        [0.8924, 0.3524, 0.5912, 0.1681, 0.7513, 0.5760, 0.3734, 0.6122],\n",
      "        [0.2243, 0.3824, 0.7740, 0.1810, 0.9710, 0.5640, 0.4942, 0.5299],\n",
      "        [0.8625, 0.4748, 0.1725, 0.9206, 0.9410, 0.7507, 0.4352, 0.3791]])\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "torch.Size([8, 8])"
      ]
     },
     "execution_count": 5,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = torch.rand(8,8)\n",
    "print(x)\n",
    "x.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "hE7Cr51smehn",
    "outputId": "db541c01-8f71-4fe6-fc55-c18af8f92670"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([64])"
      ]
     },
     "execution_count": 8,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y=x.view(64)\n",
    "y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "CJTvfwfCmlY4",
    "outputId": "38a70640-ccaf-459e-f0e7-0e93537d109d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[0.3976, 0.2362, 0.3250, 0.7433, 0.3457, 0.1334, 0.6145, 0.2543, 0.3360,\n",
       "         0.4803, 0.8231, 0.5072, 0.8074, 0.8740, 0.1817, 0.4878],\n",
       "        [0.2481, 0.4676, 0.5134, 0.4437, 0.7316, 0.7485, 0.7084, 0.6157, 0.9833,\n",
       "         0.1059, 0.5228, 0.3079, 0.9208, 0.8198, 0.8263, 0.9913],\n",
       "        [0.4339, 0.9590, 0.0707, 0.6564, 0.6033, 0.1882, 0.8471, 0.4325, 0.8924,\n",
       "         0.3524, 0.5912, 0.1681, 0.7513, 0.5760, 0.3734, 0.6122],\n",
       "        [0.2243, 0.3824, 0.7740, 0.1810, 0.9710, 0.5640, 0.4942, 0.5299, 0.8625,\n",
       "         0.4748, 0.1725, 0.9206, 0.9410, 0.7507, 0.4352, 0.3791]])"
      ]
     },
     "execution_count": 9,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x.view(4,16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "siuamgEfragG",
    "outputId": "4838eaa3-fb10-4f1d-86f1-9983288c37b8"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[0.3976, 0.2362, 0.3250, 0.7433, 0.3457, 0.1334, 0.6145, 0.2543, 0.3360,\n",
       "         0.4803, 0.8231, 0.5072, 0.8074, 0.8740, 0.1817, 0.4878],\n",
       "        [0.2481, 0.4676, 0.5134, 0.4437, 0.7316, 0.7485, 0.7084, 0.6157, 0.9833,\n",
       "         0.1059, 0.5228, 0.3079, 0.9208, 0.8198, 0.8263, 0.9913],\n",
       "        [0.4339, 0.9590, 0.0707, 0.6564, 0.6033, 0.1882, 0.8471, 0.4325, 0.8924,\n",
       "         0.3524, 0.5912, 0.1681, 0.7513, 0.5760, 0.3734, 0.6122],\n",
       "        [0.2243, 0.3824, 0.7740, 0.1810, 0.9710, 0.5640, 0.4942, 0.5299, 0.8625,\n",
       "         0.4748, 0.1725, 0.9206, 0.9410, 0.7507, 0.4352, 0.3791]])"
      ]
     },
     "execution_count": 10,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x.view(-1,16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "vqff3vAQrddL",
    "outputId": "2259b07e-6951-4b6d-d634-e25bdea777a3"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.39762217 0.23619479 0.32497042 0.74326366 0.34572303 0.1334269\n",
      "  0.61449426 0.2543121 ]\n",
      " [0.3359866  0.48026216 0.82313406 0.5071559  0.8074372  0.8739712\n",
      "  0.18168235 0.4877885 ]\n",
      " [0.24812645 0.4676389  0.51340616 0.44367492 0.7316265  0.74847585\n",
      "  0.7084268  0.6156904 ]\n",
      " [0.9832553  0.10591137 0.52277523 0.3079158  0.920777   0.8198265\n",
      "  0.82629675 0.9913481 ]\n",
      " [0.43390596 0.9589873  0.07069868 0.65644807 0.6032713  0.18816608\n",
      "  0.8471476  0.43247557]\n",
      " [0.8924207  0.35235912 0.5912032  0.16812909 0.75130737 0.5760197\n",
      "  0.37343603 0.6122213 ]\n",
      " [0.22430557 0.3824473  0.7740357  0.18098116 0.9710045  0.5639605\n",
      "  0.49423772 0.52994597]\n",
      " [0.86252207 0.47476012 0.17245269 0.9205597  0.9410002  0.750687\n",
      "  0.43517268 0.37913066]]\n"
     ]
    }
   ],
   "source": [
    "y=x.numpy()\n",
    "print(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "tqZlqxXErsoX"
   },
   "outputs": [],
   "source": [
    "x = torch.ones(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "crPPdPOyrxW0",
    "outputId": "53a23b87-cf0e-4fda-ca04-8d181a257479"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([1.])"
      ]
     },
     "execution_count": 13,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "2s81fdSdrx2B",
    "outputId": "a784ab4c-1ad2-4426-d7ae-78418d2d3930"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 14,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x.item()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "6YPTfknpnUtJ"
   },
   "source": [
    "item : loss 출력 시 사용"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "ziOuXcfNr0ph"
   },
   "outputs": [],
   "source": [
    "import torch #기복적인 토치\n",
    "import torchvision #기본적인 토치 모듈, dataset 다운로드\n",
    "import torchvision.transforms as tr #데이터 불러오면서 바로 전처리 할수있게 해주는 라이브러리\n",
    "from torch.utils.data import DataLoader, Dataset  # 배치사이즈 형태로 만들어줘서 살제학습 시 사용, 데이터 튜닝\n",
    "import numpy as np\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "iqC-D5VHI_ph"
   },
   "outputs": [],
   "source": [
    "import torch #기복적인 토치\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "from torchsummary import summary #summary\n",
    "from torch import optim #optimizer\n",
    "from torch.optim.lr_scheduler import StepLR\n",
    "\n",
    "# dataset and transformation\n",
    "import torchvision\n",
    "from torchvision import datasets #데이터 튜닝\n",
    "import torchvision.transforms as transforms #데이터 불러오면서 바로 전처리 할수있게 해주는 라이브러리\n",
    "from torch.utils.data import DataLoader # 배치사이즈 형태로 만들어줘서 살제학습 시 사용\n",
    "import os\n",
    "\n",
    "# display images\n",
    "from torchvision import utils\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "# utils\n",
    "import numpy as np\n",
    "from torchsummary import summary\n",
    "import time\n",
    "import copy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 527
    },
    "id": "_ICLoeKapmB8",
    "outputId": "f89f0db0-baa2-408c-c97c-125ee3a016b2"
   },
   "outputs": [
    {
     "ename": "KeyboardInterrupt",
     "evalue": "ignored",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/ipykernel/kernelbase.py\u001b[0m in \u001b[0;36m_input_request\u001b[0;34m(self, prompt, ident, parent, password)\u001b[0m\n\u001b[1;32m    728\u001b[0m             \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 729\u001b[0;31m                 \u001b[0mident\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mreply\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msession\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrecv\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstdin_socket\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0\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    730\u001b[0m             \u001b[0;32mexcept\u001b[0m \u001b[0mException\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/jupyter_client/session.py\u001b[0m in \u001b[0;36mrecv\u001b[0;34m(self, socket, mode, content, copy)\u001b[0m\n\u001b[1;32m    802\u001b[0m         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 803\u001b[0;31m             \u001b[0mmsg_list\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msocket\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrecv_multipart\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmode\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcopy\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcopy\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    804\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0mzmq\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mZMQError\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\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/zmq/sugar/socket.py\u001b[0m in \u001b[0;36mrecv_multipart\u001b[0;34m(self, flags, copy, track)\u001b[0m\n\u001b[1;32m    582\u001b[0m         \"\"\"\n\u001b[0;32m--> 583\u001b[0;31m         \u001b[0mparts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrecv\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mflags\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcopy\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcopy\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtrack\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mtrack\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    584\u001b[0m         \u001b[0;31m# have first part already, only loop while more to receive\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32mzmq/backend/cython/socket.pyx\u001b[0m in \u001b[0;36mzmq.backend.cython.socket.Socket.recv\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32mzmq/backend/cython/socket.pyx\u001b[0m in \u001b[0;36mzmq.backend.cython.socket.Socket.recv\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32mzmq/backend/cython/socket.pyx\u001b[0m in \u001b[0;36mzmq.backend.cython.socket._recv_copy\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/zmq/backend/cython/checkrc.pxd\u001b[0m in \u001b[0;36mzmq.backend.cython.checkrc._check_rc\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: ",
      "\nDuring handling of the above exception, another exception occurred:\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-2-4996ee3d8d09>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mgoogle\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcolab\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mdrive\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mdrive\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmount\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'/content/gdrive'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/google/colab/drive.py\u001b[0m in \u001b[0;36mmount\u001b[0;34m(mountpoint, force_remount, timeout_ms, use_metadata_server)\u001b[0m\n\u001b[1;32m    258\u001b[0m       \u001b[0;32mwith\u001b[0m \u001b[0m_output\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0muse_tags\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'dfs-auth-dance'\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    259\u001b[0m         \u001b[0;32mwith\u001b[0m \u001b[0mopen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfifo\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'w'\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mfifo_file\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 260\u001b[0;31m           \u001b[0mfifo_file\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwrite\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mget_code\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mauth_prompt\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m'\\n'\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    261\u001b[0m       \u001b[0mwrote_to_fifo\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    262\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0mcase\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m5\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/ipykernel/kernelbase.py\u001b[0m in \u001b[0;36mraw_input\u001b[0;34m(self, prompt)\u001b[0m\n\u001b[1;32m    702\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_parent_ident\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    703\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_parent_header\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 704\u001b[0;31m             \u001b[0mpassword\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\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    705\u001b[0m         )\n\u001b[1;32m    706\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/ipykernel/kernelbase.py\u001b[0m in \u001b[0;36m_input_request\u001b[0;34m(self, prompt, ident, parent, password)\u001b[0m\n\u001b[1;32m    732\u001b[0m             \u001b[0;32mexcept\u001b[0m \u001b[0mKeyboardInterrupt\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    733\u001b[0m                 \u001b[0;31m# re-raise KeyboardInterrupt, to truncate traceback\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 734\u001b[0;31m                 \u001b[0;32mraise\u001b[0m \u001b[0mKeyboardInterrupt\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    735\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[1;32m    736\u001b[0m                 \u001b[0;32mbreak\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "from google.colab import drive\n",
    "drive.mount('/content/gdrive')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Mq0n0wGyr066"
   },
   "source": [
    "데이터셋 다운받을 경로를 지정하고, 데이터셋을 불러옵니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 181,
     "referenced_widgets": [
      "9cf8ad1f2ea6405a82d3e357ac09cc88",
      "95baf3c673e24eea8d17611287242c7b",
      "6d79ea7663d747d2b0eda6504334f2e9",
      "1c7d33e05afb4d61baef5e9c3693019a",
      "9cf4f0cd473e42a79373412dbd606074",
      "52daecb58f844a7da50028c62001495e",
      "a8e5ca3c10b049e68ae3726593e25e06",
      "b06b0515a88a41999dd70ffffe9f9516"
     ]
    },
    "id": "-04KymR3pBYM",
    "outputId": "ad5f12a1-cc20-4afc-db47-24426cd3af31"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading http://ai.stanford.edu/~acoates/stl10/stl10_binary.tar.gz to /content/gdrive/My Drive/stl10_binary.tar.gz\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9cf8ad1f2ea6405a82d3e357ac09cc88",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=2640397119.0), HTML(value='')))"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Extracting /content/gdrive/My Drive/stl10_binary.tar.gz to /content/gdrive/My Drive/\n",
      "Files already downloaded and verified\n",
      "5000\n",
      "8000\n"
     ]
    }
   ],
   "source": [
    "# specify the data path\n",
    "path2data = '/content/gdrive/My Drive/'\n",
    "\n",
    "# if not exists the path, make the directory\n",
    "if not os.path.exists(path2data):\n",
    "    os.mkdir(path2data)\n",
    "\n",
    "# load dataset\n",
    "train_ds = torchvision.datasets.STL10(path2data, split='train', download=True, transform=transforms.ToTensor()) #transform : 미리 전처리를 해주겠다., tr.Compose로 설정\n",
    "val_ds = torchvision.datasets.STL10(path2data, split='test', download=True, transform=transforms.ToTensor())\n",
    "\n",
    "print(len(train_ds))\n",
    "print(len(val_ds))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "8tKamGRZNr1J"
   },
   "outputs": [],
   "source": [
    "train_ds = torchvision.datasets.cityscapes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "u1dSZEZZrqAp"
   },
   "source": [
    "이미지에 Normalization을 적용하기 위해, 이미지 픽셀값의 평균, 표준편차를 계산합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "CnAvr0asrxMm",
    "outputId": "b0d2c1e2-6b0d-467d-e847-038f17ad78b3"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.4467106 0.43980986 0.40664646\n",
      "0.44723064 0.4396425 0.40495726\n"
     ]
    }
   ],
   "source": [
    "# To normalize the dataset, calculate the mean and std\n",
    "train_meanRGB = [np.mean(x.numpy(), axis=(1,2)) for x, _ in train_ds]\n",
    "train_stdRGB = [np.std(x.numpy(), axis=(1,2)) for x, _ in train_ds]\n",
    "\n",
    "train_meanR = np.mean([m[0] for m in train_meanRGB])\n",
    "train_meanG = np.mean([m[1] for m in train_meanRGB])\n",
    "train_meanB = np.mean([m[2] for m in train_meanRGB])\n",
    "train_stdR = np.mean([s[0] for s in train_stdRGB])\n",
    "train_stdG = np.mean([s[1] for s in train_stdRGB])\n",
    "train_stdB = np.mean([s[2] for s in train_stdRGB])\n",
    "\n",
    "\n",
    "val_meanRGB = [np.mean(x.numpy(), axis=(1,2)) for x, _ in val_ds]\n",
    "val_stdRGB = [np.std(x.numpy(), axis=(1,2)) for x, _ in val_ds]\n",
    "\n",
    "val_meanR = np.mean([m[0] for m in val_meanRGB])\n",
    "val_meanG = np.mean([m[1] for m in val_meanRGB])\n",
    "val_meanB = np.mean([m[2] for m in val_meanRGB])\n",
    "\n",
    "val_stdR = np.mean([s[0] for s in val_stdRGB])\n",
    "val_stdG = np.mean([s[1] for s in val_stdRGB])\n",
    "val_stdB = np.mean([s[2] for s in val_stdRGB])\n",
    "\n",
    "print(train_meanR, train_meanG, train_meanB)\n",
    "print(val_meanR, val_meanG, val_meanB)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "PQhQ94sRsHvl"
   },
   "source": [
    "dataset에 적용할 transformation을 정의합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "Dx_jPUMpsJHq"
   },
   "outputs": [],
   "source": [
    "train_transformation = transforms.Compose([\n",
    "                        transforms.ToTensor(),\n",
    "                        transforms.Resize(224),\n",
    "                        transforms.Normalize([train_meanR, train_meanG, train_meanB],[train_stdR, train_stdG, train_stdB]),\n",
    "                        transforms.RandomHorizontalFlip(),\n",
    "])\n",
    "\n",
    "val_transformation = transforms.Compose([\n",
    "                        transforms.ToTensor(),\n",
    "                        transforms.Resize(224),\n",
    "                        transforms.Normalize([train_meanR, train_meanG, train_meanB],[train_stdR, train_stdG, train_stdB]),\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "p345PesGsMU9"
   },
   "source": [
    "transformation을 dataset에 적용하고, dataloader를 생성합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "-MQhFHwtsRdj"
   },
   "outputs": [],
   "source": [
    "train_ds.transform = train_transformation\n",
    "val_ds.transform = val_transformation\n",
    "\n",
    "# create DataLoader\n",
    "train_dl = DataLoader(train_ds, batch_size=32, shuffle=True)\n",
    "val_dl = DataLoader(val_ds, batch_size=32, shuffle=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "5P0mLw8wsejr"
   },
   "source": [
    "transformation이 적용된 샘플 이미지를 확인하겠습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 205
    },
    "id": "YDKnzY2IsXSe",
    "outputId": "110dfa35-f655-43d9-d2e4-d947a2e71888"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "image indices: [ 235 3980  905 2763]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# display sample images \n",
    "def show(img, y=None, color=True):\n",
    "    npimg = img.numpy()\n",
    "    npimg_tr = np.transpose(npimg, (1,2,0))\n",
    "    plt.imshow(npimg_tr)\n",
    "\n",
    "    if y is not None:\n",
    "        plt.title('labels :' + str(y))\n",
    "\n",
    "np.random.seed(1)\n",
    "torch.manual_seed(1)\n",
    "\n",
    "grid_size = 4\n",
    "rnd_inds = np.random.randint(0, len(train_ds), grid_size)\n",
    "print('image indices:',rnd_inds)\n",
    "\n",
    "x_grid = [train_ds[i][0] for i in rnd_inds]\n",
    "y_grid = [train_ds[i][1] for i in rnd_inds]\n",
    "\n",
    "x_grid = utils.make_grid(x_grid, nrow=grid_size, padding=2)\n",
    "\n",
    "show(x_grid, y_grid)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "oXWSfNkUuUwG"
   },
   "source": [
    "모델 구축하기 : 각각의 residual block을 정의합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "qt0qScxyuT3c"
   },
   "outputs": [],
   "source": [
    "class BasicBlock(nn.Module):\n",
    "    expansion = 1\n",
    "    def __init__(self, in_channels, out_channels, stride=1):\n",
    "        super().__init__()\n",
    "\n",
    "        # BatchNorm에 bias가 포함되어 있으므로, conv2d는 bias=False로 설정합니다.\n",
    "        self.residual_function = nn.Sequential(\n",
    "            nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=stride, padding=1, bias=False),\n",
    "            nn.BatchNorm2d(out_channels),\n",
    "            nn.ReLU(),\n",
    "            nn.Conv2d(out_channels, out_channels * BasicBlock.expansion, kernel_size=3, stride=1, padding=1, bias=False),\n",
    "            nn.BatchNorm2d(out_channels * BasicBlock.expansion),\n",
    "        )\n",
    "\n",
    "        # identity mapping, input과 output의 feature map size, filter 수가 동일한 경우 사용.\n",
    "        self.shortcut = nn.Sequential()\n",
    "\n",
    "        self.relu = nn.ReLU()\n",
    "\n",
    "        # projection mapping using 1x1conv\n",
    "        if stride != 1 or in_channels != BasicBlock.expansion * out_channels:\n",
    "            self.shortcut = nn.Sequential(\n",
    "                nn.Conv2d(in_channels, out_channels * BasicBlock.expansion, kernel_size=1, stride=stride, bias=False),\n",
    "                nn.BatchNorm2d(out_channels * BasicBlock.expansion)\n",
    "            )\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.residual_function(x) + self.shortcut(x)\n",
    "        x = self.relu(x)\n",
    "        return x\n",
    "\n",
    "\n",
    "class BottleNeck(nn.Module):\n",
    "    expansion = 4\n",
    "    def __init__(self, in_channels, out_channels, stride=1):\n",
    "        super().__init__()\n",
    "\n",
    "        self.residual_function = nn.Sequential(\n",
    "            nn.Conv2d(in_channels, out_channels, kernel_size=1, stride=1, bias=False),\n",
    "            nn.BatchNorm2d(out_channels),\n",
    "            nn.ReLU(),\n",
    "            nn.Conv2d(out_channels, out_channels, kernel_size=3, stride=stride, padding=1, bias=False),\n",
    "            nn.BatchNorm2d(out_channels),\n",
    "            nn.ReLU(),\n",
    "            nn.Conv2d(out_channels, out_channels * BottleNeck.expansion, kernel_size=1, stride=1, bias=False),\n",
    "            nn.BatchNorm2d(out_channels * BottleNeck.expansion),\n",
    "        )\n",
    "\n",
    "        self.shortcut = nn.Sequential()\n",
    "\n",
    "        self.relu = nn.ReLU()\n",
    "\n",
    "        if stride != 1 or in_channels != out_channels * BottleNeck.expansion:\n",
    "            self.shortcut = nn.Sequential(\n",
    "                nn.Conv2d(in_channels, out_channels*BottleNeck.expansion, kernel_size=1, stride=stride, bias=False),\n",
    "                nn.BatchNorm2d(out_channels*BottleNeck.expansion)\n",
    "            )\n",
    "            \n",
    "    def forward(self, x):\n",
    "        x = self.residual_function(x) + self.shortcut(x)\n",
    "        x = self.relu(x)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "5rhz_r2xudBM"
   },
   "outputs": [],
   "source": [
    "class ResNet(nn.Module):\n",
    "    def __init__(self, block, num_block, num_classes=10, init_weights=True):\n",
    "        super().__init__()\n",
    "\n",
    "        self.in_channels=64\n",
    "\n",
    "        self.conv1 = nn.Sequential(\n",
    "            nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3, bias=False),\n",
    "            nn.BatchNorm2d(64),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(kernel_size=3, stride=2, padding=1)\n",
    "        )\n",
    "\n",
    "        self.conv2_x = self._make_layer(block, 64, num_block[0], 1)\n",
    "        self.conv3_x = self._make_layer(block, 128, num_block[1], 2)\n",
    "        self.conv4_x = self._make_layer(block, 256, num_block[2], 2)\n",
    "        self.conv5_x = self._make_layer(block, 512, num_block[3], 2)\n",
    "\n",
    "        self.avg_pool = nn.AdaptiveAvgPool2d((1,1))\n",
    "        self.fc = nn.Linear(512 * block.expansion, num_classes)\n",
    "\n",
    "        # weights inittialization\n",
    "        if init_weights:\n",
    "            self._initialize_weights()\n",
    "\n",
    "    def _make_layer(self, block, out_channels, num_blocks, stride):\n",
    "        strides = [stride] + [1] * (num_blocks - 1)\n",
    "        layers = []\n",
    "        for stride in strides:\n",
    "            layers.append(block(self.in_channels, out_channels, stride))\n",
    "            self.in_channels = out_channels * block.expansion\n",
    "\n",
    "        return nn.Sequential(*layers)\n",
    "\n",
    "    def forward(self,x):\n",
    "        output = self.conv1(x)\n",
    "        output = self.conv2_x(output)\n",
    "        x = self.conv3_x(output)\n",
    "        x = self.conv4_x(x)\n",
    "        x = self.conv5_x(x)\n",
    "        x = self.avg_pool(x)\n",
    "        x = x.view(x.size(0), -1)\n",
    "        x = self.fc(x)\n",
    "        return x\n",
    "\n",
    "    # define weight initialization function\n",
    "    def _initialize_weights(self):\n",
    "        for m in self.modules():\n",
    "            if isinstance(m, nn.Conv2d):\n",
    "                nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')\n",
    "                if m.bias is not None:\n",
    "                    nn.init.constant_(m.bias, 0)\n",
    "            elif isinstance(m, nn.BatchNorm2d):\n",
    "                nn.init.constant_(m.weight, 1)\n",
    "                nn.init.constant_(m.bias, 0)\n",
    "            elif isinstance(m, nn.Linear):\n",
    "                nn.init.normal_(m.weight, 0, 0.01)\n",
    "                nn.init.constant_(m.bias, 0)\n",
    "\n",
    "def resnet18():\n",
    "    return ResNet(BasicBlock, [2,2,2,2])\n",
    "\n",
    "def resnet34():\n",
    "    return ResNet(BasicBlock, [3, 4, 6, 3])\n",
    "\n",
    "def resnet50():\n",
    "    return ResNet(BottleNeck, [3,4,6,3])\n",
    "\n",
    "def resnet101():\n",
    "    return ResNet(BottleNeck, [3, 4, 23, 3])\n",
    "\n",
    "def resnet152():\n",
    "    return ResNet(BottleNeck, [3, 8, 36, 3])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Smn5p_lQug30"
   },
   "source": [
    "모델이 잘 구축됬는지 확인합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "W1Wkq7mCufds"
   },
   "outputs": [],
   "source": [
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "model = resnet50().to(device)\n",
    "x = torch.randn(3, 3, 224, 224).to(device)\n",
    "output = model(x)\n",
    "print(output.size())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "0IA4nTwSunXF",
    "outputId": "5391f2e7-017e-4f56-b5d5-77894f495c99"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1         [-1, 64, 112, 112]           9,408\n",
      "       BatchNorm2d-2         [-1, 64, 112, 112]             128\n",
      "              ReLU-3         [-1, 64, 112, 112]               0\n",
      "         MaxPool2d-4           [-1, 64, 56, 56]               0\n",
      "            Conv2d-5           [-1, 64, 56, 56]           4,096\n",
      "       BatchNorm2d-6           [-1, 64, 56, 56]             128\n",
      "              ReLU-7           [-1, 64, 56, 56]               0\n",
      "            Conv2d-8           [-1, 64, 56, 56]          36,864\n",
      "       BatchNorm2d-9           [-1, 64, 56, 56]             128\n",
      "             ReLU-10           [-1, 64, 56, 56]               0\n",
      "           Conv2d-11          [-1, 256, 56, 56]          16,384\n",
      "      BatchNorm2d-12          [-1, 256, 56, 56]             512\n",
      "           Conv2d-13          [-1, 256, 56, 56]          16,384\n",
      "      BatchNorm2d-14          [-1, 256, 56, 56]             512\n",
      "             ReLU-15          [-1, 256, 56, 56]               0\n",
      "       BottleNeck-16          [-1, 256, 56, 56]               0\n",
      "           Conv2d-17           [-1, 64, 56, 56]          16,384\n",
      "      BatchNorm2d-18           [-1, 64, 56, 56]             128\n",
      "             ReLU-19           [-1, 64, 56, 56]               0\n",
      "           Conv2d-20           [-1, 64, 56, 56]          36,864\n",
      "      BatchNorm2d-21           [-1, 64, 56, 56]             128\n",
      "             ReLU-22           [-1, 64, 56, 56]               0\n",
      "           Conv2d-23          [-1, 256, 56, 56]          16,384\n",
      "      BatchNorm2d-24          [-1, 256, 56, 56]             512\n",
      "             ReLU-25          [-1, 256, 56, 56]               0\n",
      "       BottleNeck-26          [-1, 256, 56, 56]               0\n",
      "           Conv2d-27           [-1, 64, 56, 56]          16,384\n",
      "      BatchNorm2d-28           [-1, 64, 56, 56]             128\n",
      "             ReLU-29           [-1, 64, 56, 56]               0\n",
      "           Conv2d-30           [-1, 64, 56, 56]          36,864\n",
      "      BatchNorm2d-31           [-1, 64, 56, 56]             128\n",
      "             ReLU-32           [-1, 64, 56, 56]               0\n",
      "           Conv2d-33          [-1, 256, 56, 56]          16,384\n",
      "      BatchNorm2d-34          [-1, 256, 56, 56]             512\n",
      "             ReLU-35          [-1, 256, 56, 56]               0\n",
      "       BottleNeck-36          [-1, 256, 56, 56]               0\n",
      "           Conv2d-37          [-1, 128, 56, 56]          32,768\n",
      "      BatchNorm2d-38          [-1, 128, 56, 56]             256\n",
      "             ReLU-39          [-1, 128, 56, 56]               0\n",
      "           Conv2d-40          [-1, 128, 28, 28]         147,456\n",
      "      BatchNorm2d-41          [-1, 128, 28, 28]             256\n",
      "             ReLU-42          [-1, 128, 28, 28]               0\n",
      "           Conv2d-43          [-1, 512, 28, 28]          65,536\n",
      "      BatchNorm2d-44          [-1, 512, 28, 28]           1,024\n",
      "           Conv2d-45          [-1, 512, 28, 28]         131,072\n",
      "      BatchNorm2d-46          [-1, 512, 28, 28]           1,024\n",
      "             ReLU-47          [-1, 512, 28, 28]               0\n",
      "       BottleNeck-48          [-1, 512, 28, 28]               0\n",
      "           Conv2d-49          [-1, 128, 28, 28]          65,536\n",
      "      BatchNorm2d-50          [-1, 128, 28, 28]             256\n",
      "             ReLU-51          [-1, 128, 28, 28]               0\n",
      "           Conv2d-52          [-1, 128, 28, 28]         147,456\n",
      "      BatchNorm2d-53          [-1, 128, 28, 28]             256\n",
      "             ReLU-54          [-1, 128, 28, 28]               0\n",
      "           Conv2d-55          [-1, 512, 28, 28]          65,536\n",
      "      BatchNorm2d-56          [-1, 512, 28, 28]           1,024\n",
      "             ReLU-57          [-1, 512, 28, 28]               0\n",
      "       BottleNeck-58          [-1, 512, 28, 28]               0\n",
      "           Conv2d-59          [-1, 128, 28, 28]          65,536\n",
      "      BatchNorm2d-60          [-1, 128, 28, 28]             256\n",
      "             ReLU-61          [-1, 128, 28, 28]               0\n",
      "           Conv2d-62          [-1, 128, 28, 28]         147,456\n",
      "      BatchNorm2d-63          [-1, 128, 28, 28]             256\n",
      "             ReLU-64          [-1, 128, 28, 28]               0\n",
      "           Conv2d-65          [-1, 512, 28, 28]          65,536\n",
      "      BatchNorm2d-66          [-1, 512, 28, 28]           1,024\n",
      "             ReLU-67          [-1, 512, 28, 28]               0\n",
      "       BottleNeck-68          [-1, 512, 28, 28]               0\n",
      "           Conv2d-69          [-1, 128, 28, 28]          65,536\n",
      "      BatchNorm2d-70          [-1, 128, 28, 28]             256\n",
      "             ReLU-71          [-1, 128, 28, 28]               0\n",
      "           Conv2d-72          [-1, 128, 28, 28]         147,456\n",
      "      BatchNorm2d-73          [-1, 128, 28, 28]             256\n",
      "             ReLU-74          [-1, 128, 28, 28]               0\n",
      "           Conv2d-75          [-1, 512, 28, 28]          65,536\n",
      "      BatchNorm2d-76          [-1, 512, 28, 28]           1,024\n",
      "             ReLU-77          [-1, 512, 28, 28]               0\n",
      "       BottleNeck-78          [-1, 512, 28, 28]               0\n",
      "           Conv2d-79          [-1, 256, 28, 28]         131,072\n",
      "      BatchNorm2d-80          [-1, 256, 28, 28]             512\n",
      "             ReLU-81          [-1, 256, 28, 28]               0\n",
      "           Conv2d-82          [-1, 256, 14, 14]         589,824\n",
      "      BatchNorm2d-83          [-1, 256, 14, 14]             512\n",
      "             ReLU-84          [-1, 256, 14, 14]               0\n",
      "           Conv2d-85         [-1, 1024, 14, 14]         262,144\n",
      "      BatchNorm2d-86         [-1, 1024, 14, 14]           2,048\n",
      "           Conv2d-87         [-1, 1024, 14, 14]         524,288\n",
      "      BatchNorm2d-88         [-1, 1024, 14, 14]           2,048\n",
      "             ReLU-89         [-1, 1024, 14, 14]               0\n",
      "       BottleNeck-90         [-1, 1024, 14, 14]               0\n",
      "           Conv2d-91          [-1, 256, 14, 14]         262,144\n",
      "      BatchNorm2d-92          [-1, 256, 14, 14]             512\n",
      "             ReLU-93          [-1, 256, 14, 14]               0\n",
      "           Conv2d-94          [-1, 256, 14, 14]         589,824\n",
      "      BatchNorm2d-95          [-1, 256, 14, 14]             512\n",
      "             ReLU-96          [-1, 256, 14, 14]               0\n",
      "           Conv2d-97         [-1, 1024, 14, 14]         262,144\n",
      "      BatchNorm2d-98         [-1, 1024, 14, 14]           2,048\n",
      "             ReLU-99         [-1, 1024, 14, 14]               0\n",
      "      BottleNeck-100         [-1, 1024, 14, 14]               0\n",
      "          Conv2d-101          [-1, 256, 14, 14]         262,144\n",
      "     BatchNorm2d-102          [-1, 256, 14, 14]             512\n",
      "            ReLU-103          [-1, 256, 14, 14]               0\n",
      "          Conv2d-104          [-1, 256, 14, 14]         589,824\n",
      "     BatchNorm2d-105          [-1, 256, 14, 14]             512\n",
      "            ReLU-106          [-1, 256, 14, 14]               0\n",
      "          Conv2d-107         [-1, 1024, 14, 14]         262,144\n",
      "     BatchNorm2d-108         [-1, 1024, 14, 14]           2,048\n",
      "            ReLU-109         [-1, 1024, 14, 14]               0\n",
      "      BottleNeck-110         [-1, 1024, 14, 14]               0\n",
      "          Conv2d-111          [-1, 256, 14, 14]         262,144\n",
      "     BatchNorm2d-112          [-1, 256, 14, 14]             512\n",
      "            ReLU-113          [-1, 256, 14, 14]               0\n",
      "          Conv2d-114          [-1, 256, 14, 14]         589,824\n",
      "     BatchNorm2d-115          [-1, 256, 14, 14]             512\n",
      "            ReLU-116          [-1, 256, 14, 14]               0\n",
      "          Conv2d-117         [-1, 1024, 14, 14]         262,144\n",
      "     BatchNorm2d-118         [-1, 1024, 14, 14]           2,048\n",
      "            ReLU-119         [-1, 1024, 14, 14]               0\n",
      "      BottleNeck-120         [-1, 1024, 14, 14]               0\n",
      "          Conv2d-121          [-1, 256, 14, 14]         262,144\n",
      "     BatchNorm2d-122          [-1, 256, 14, 14]             512\n",
      "            ReLU-123          [-1, 256, 14, 14]               0\n",
      "          Conv2d-124          [-1, 256, 14, 14]         589,824\n",
      "     BatchNorm2d-125          [-1, 256, 14, 14]             512\n",
      "            ReLU-126          [-1, 256, 14, 14]               0\n",
      "          Conv2d-127         [-1, 1024, 14, 14]         262,144\n",
      "     BatchNorm2d-128         [-1, 1024, 14, 14]           2,048\n",
      "            ReLU-129         [-1, 1024, 14, 14]               0\n",
      "      BottleNeck-130         [-1, 1024, 14, 14]               0\n",
      "          Conv2d-131          [-1, 256, 14, 14]         262,144\n",
      "     BatchNorm2d-132          [-1, 256, 14, 14]             512\n",
      "            ReLU-133          [-1, 256, 14, 14]               0\n",
      "          Conv2d-134          [-1, 256, 14, 14]         589,824\n",
      "     BatchNorm2d-135          [-1, 256, 14, 14]             512\n",
      "            ReLU-136          [-1, 256, 14, 14]               0\n",
      "          Conv2d-137         [-1, 1024, 14, 14]         262,144\n",
      "     BatchNorm2d-138         [-1, 1024, 14, 14]           2,048\n",
      "            ReLU-139         [-1, 1024, 14, 14]               0\n",
      "      BottleNeck-140         [-1, 1024, 14, 14]               0\n",
      "          Conv2d-141          [-1, 512, 14, 14]         524,288\n",
      "     BatchNorm2d-142          [-1, 512, 14, 14]           1,024\n",
      "            ReLU-143          [-1, 512, 14, 14]               0\n",
      "          Conv2d-144            [-1, 512, 7, 7]       2,359,296\n",
      "     BatchNorm2d-145            [-1, 512, 7, 7]           1,024\n",
      "            ReLU-146            [-1, 512, 7, 7]               0\n",
      "          Conv2d-147           [-1, 2048, 7, 7]       1,048,576\n",
      "     BatchNorm2d-148           [-1, 2048, 7, 7]           4,096\n",
      "          Conv2d-149           [-1, 2048, 7, 7]       2,097,152\n",
      "     BatchNorm2d-150           [-1, 2048, 7, 7]           4,096\n",
      "            ReLU-151           [-1, 2048, 7, 7]               0\n",
      "      BottleNeck-152           [-1, 2048, 7, 7]               0\n",
      "          Conv2d-153            [-1, 512, 7, 7]       1,048,576\n",
      "     BatchNorm2d-154            [-1, 512, 7, 7]           1,024\n",
      "            ReLU-155            [-1, 512, 7, 7]               0\n",
      "          Conv2d-156            [-1, 512, 7, 7]       2,359,296\n",
      "     BatchNorm2d-157            [-1, 512, 7, 7]           1,024\n",
      "            ReLU-158            [-1, 512, 7, 7]               0\n",
      "          Conv2d-159           [-1, 2048, 7, 7]       1,048,576\n",
      "     BatchNorm2d-160           [-1, 2048, 7, 7]           4,096\n",
      "            ReLU-161           [-1, 2048, 7, 7]               0\n",
      "      BottleNeck-162           [-1, 2048, 7, 7]               0\n",
      "          Conv2d-163            [-1, 512, 7, 7]       1,048,576\n",
      "     BatchNorm2d-164            [-1, 512, 7, 7]           1,024\n",
      "            ReLU-165            [-1, 512, 7, 7]               0\n",
      "          Conv2d-166            [-1, 512, 7, 7]       2,359,296\n",
      "     BatchNorm2d-167            [-1, 512, 7, 7]           1,024\n",
      "            ReLU-168            [-1, 512, 7, 7]               0\n",
      "          Conv2d-169           [-1, 2048, 7, 7]       1,048,576\n",
      "     BatchNorm2d-170           [-1, 2048, 7, 7]           4,096\n",
      "            ReLU-171           [-1, 2048, 7, 7]               0\n",
      "      BottleNeck-172           [-1, 2048, 7, 7]               0\n",
      "AdaptiveAvgPool2d-173           [-1, 2048, 1, 1]               0\n",
      "          Linear-174                   [-1, 10]          20,490\n",
      "================================================================\n",
      "Total params: 23,528,522\n",
      "Trainable params: 23,528,522\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.57\n",
      "Forward/backward pass size (MB): 286.55\n",
      "Params size (MB): 89.75\n",
      "Estimated Total Size (MB): 376.88\n",
      "----------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "summary(model, (3, 224, 224), device=device.type)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Z2btGDrwutgQ"
   },
   "source": [
    "3. 모델 학습하기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "3_fcfUTXusel"
   },
   "outputs": [],
   "source": [
    "loss_func = nn.CrossEntropyLoss(reduction='sum')\n",
    "opt = optim.Adam(model.parameters(), lr=0.001)\n",
    "\n",
    "from torch.optim.lr_scheduler import ReduceLROnPlateau\n",
    "lr_scheduler = ReduceLROnPlateau(opt, mode='min', factor=0.1, patience=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "sm-IcZx7uyRn"
   },
   "source": [
    " 현재 lr을 계산하는 함수를 정의합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "_oIPvxzWuz67"
   },
   "outputs": [],
   "source": [
    "# function to get current lr\n",
    "def get_lr(opt):\n",
    "    for param_group in opt.param_groups:\n",
    "        return param_group['lr']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "lr2j12tQu6rl"
   },
   "source": [
    " 배치당 loss와 metric을 계산하는 함수를 정의합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "gGWL_yNEu8Fe"
   },
   "outputs": [],
   "source": [
    "# function to calculate metric per mini-batch\n",
    "def metric_batch(output, target):\n",
    "    pred = output.argmax(1, keepdim=True)\n",
    "    corrects = pred.eq(target.view_as(pred)).sum().item()\n",
    "    return corrects\n",
    "\n",
    "\n",
    "# function to calculate loss per mini-batch\n",
    "def loss_batch(loss_func, output, target, opt=None):\n",
    "    loss = loss_func(output, target)\n",
    "    metric_b = metric_batch(output, target)\n",
    "\n",
    "    if opt is not None:\n",
    "        opt.zero_grad()\n",
    "        loss.backward()\n",
    "        opt.step()\n",
    "\n",
    "    return loss.item(), metric_b"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "keqhkMyIu_TS"
   },
   "source": [
    "epoch당 loss를 정의하는 함수입니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "MTaqxBbevCVo"
   },
   "outputs": [],
   "source": [
    "# function to calculate loss and metric per epoch\n",
    "def loss_epoch(model, loss_func, dataset_dl, sanity_check=False, opt=None):\n",
    "    running_loss = 0.0\n",
    "    running_metric = 0.0\n",
    "    len_data = len(dataset_dl.dataset)\n",
    "\n",
    "    for xb, yb in dataset_dl:\n",
    "        xb = xb.to(device)\n",
    "        yb = yb.to(device)\n",
    "        output = model(xb)\n",
    "\n",
    "        loss_b, metric_b = loss_batch(loss_func, output, yb, opt)\n",
    "\n",
    "        running_loss += loss_b\n",
    "        \n",
    "        if metric_b is not None:\n",
    "            running_metric += metric_b\n",
    "        \n",
    "        if sanity_check is True:\n",
    "            break\n",
    "\n",
    "    loss = running_loss / len_data\n",
    "    metric = running_metric / len_data\n",
    "\n",
    "    return loss, metric"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "FuzOP_JkvEy-"
   },
   "source": [
    "이제, 위에서 정의한 함수를 활용해서 학습을 시작하는 함수를 정의합니다.\n",
    "\n",
    " val_loss가 가장 낮을 때 모델의 가중치를 저장하는 코드를 구현했는데 GPU out of memory가 발생해서 주석처리 했습니다"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "TpK_wRQqvKIC"
   },
   "outputs": [],
   "source": [
    "# function to start training\n",
    "def train_val(model, params):\n",
    "    num_epochs=params['num_epochs']\n",
    "    loss_func=params[\"loss_func\"]\n",
    "    opt=params[\"optimizer\"]\n",
    "    train_dl=params[\"train_dl\"]\n",
    "    val_dl=params[\"val_dl\"]\n",
    "    sanity_check=params[\"sanity_check\"]\n",
    "    lr_scheduler=params[\"lr_scheduler\"]\n",
    "    path2weights=params[\"path2weights\"]\n",
    "\n",
    "    loss_history = {'train': [], 'val': []}\n",
    "    metric_history = {'train': [], 'val': []}\n",
    "\n",
    "    # # GPU out of memoty error\n",
    "    # best_model_wts = copy.deepcopy(model.state_dict())\n",
    "\n",
    "    best_loss = float('inf')\n",
    "\n",
    "    start_time = time.time()\n",
    "\n",
    "    for epoch in range(num_epochs):\n",
    "        current_lr = get_lr(opt)\n",
    "        print('Epoch {}/{}, current lr={}'.format(epoch, num_epochs-1, current_lr))\n",
    "\n",
    "        model.train()\n",
    "        train_loss, train_metric = loss_epoch(model, loss_func, train_dl, sanity_check, opt)\n",
    "        loss_history['train'].append(train_loss)\n",
    "        metric_history['train'].append(train_metric)\n",
    "\n",
    "        model.eval()\n",
    "        with torch.no_grad():\n",
    "            val_loss, val_metric = loss_epoch(model, loss_func, val_dl, sanity_check)\n",
    "        loss_history['val'].append(val_loss)\n",
    "        metric_history['val'].append(val_metric)\n",
    "\n",
    "        if val_loss < best_loss:\n",
    "            best_loss = val_loss\n",
    "            # best_model_wts = copy.deepcopy(model.state_dict())\n",
    "\n",
    "            # torch.save(model.state_dict(), path2weights)\n",
    "            # print('Copied best model weights!')\n",
    "            print('Get best val_loss')\n",
    "\n",
    "        lr_scheduler.step(val_loss)\n",
    "\n",
    "        print('train loss: %.6f, val loss: %.6f, accuracy: %.2f, time: %.4f min' %(train_loss, val_loss, 100*val_metric, (time.time()-start_time)/60))\n",
    "        print('-'*10)\n",
    "\n",
    "    # model.load_state_dict(best_model_wts)\n",
    "\n",
    "    return model, loss_history, metric_history"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ACaAHTQrvLp9"
   },
   "source": [
    "하이퍼 파라미터를 정의합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "Lqsf73VnvNa9"
   },
   "outputs": [],
   "source": [
    "# definc the training parameters\n",
    "params_train = {\n",
    "    'num_epochs':20,\n",
    "    'optimizer':opt,\n",
    "    'loss_func':loss_func,\n",
    "    'train_dl':train_dl,\n",
    "    'val_dl':val_dl,\n",
    "    'sanity_check':False,\n",
    "    'lr_scheduler':lr_scheduler,\n",
    "    'path2weights':'./models/weights.pt',\n",
    "}\n",
    "\n",
    "# create the directory that stores weights.pt\n",
    "def createFolder(directory):\n",
    "    try:\n",
    "        if not os.path.exists(directory):\n",
    "            os.makedirs(directory)\n",
    "    except OSerror:\n",
    "        print('Error')\n",
    "createFolder('./models')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "IbKMx-edvS9L"
   },
   "source": [
    "학습 시작"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 176
    },
    "id": "vYklZ239vYSi",
    "outputId": "9518840f-a916-491d-8ce1-afde8a2d8d91"
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "ignored",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-1-40a68c4f9fe2>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mloss_hist\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmetric_hist\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtrain_val\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mparams_train\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'train_val' is not defined"
     ]
    }
   ],
   "source": [
    "model, loss_hist, metric_hist = train_val(model, params_train)"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "Resnet_pytorch.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.9"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "1c7d33e05afb4d61baef5e9c3693019a": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_b06b0515a88a41999dd70ffffe9f9516",
      "placeholder": "​",
      "style": "IPY_MODEL_a8e5ca3c10b049e68ae3726593e25e06",
      "value": " 2640397312/? [11:59&lt;00:00, 3671891.12it/s]"
     }
    },
    "52daecb58f844a7da50028c62001495e": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "6d79ea7663d747d2b0eda6504334f2e9": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_52daecb58f844a7da50028c62001495e",
      "max": 2640397119,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_9cf4f0cd473e42a79373412dbd606074",
      "value": 2640397119
     }
    },
    "95baf3c673e24eea8d17611287242c7b": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "9cf4f0cd473e42a79373412dbd606074": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": "initial"
     }
    },
    "9cf8ad1f2ea6405a82d3e357ac09cc88": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_6d79ea7663d747d2b0eda6504334f2e9",
       "IPY_MODEL_1c7d33e05afb4d61baef5e9c3693019a"
      ],
      "layout": "IPY_MODEL_95baf3c673e24eea8d17611287242c7b"
     }
    },
    "a8e5ca3c10b049e68ae3726593e25e06": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "b06b0515a88a41999dd70ffffe9f9516": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}