Util_Data_Visualization.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Loss and Accuary Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>epoch</th>\n",
       "      <th>loss</th>\n",
       "      <th>prec</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>3.8293</td>\n",
       "      <td>0.0909</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>3.5094</td>\n",
       "      <td>0.1405</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>3.1706</td>\n",
       "      <td>0.1943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>3.2106</td>\n",
       "      <td>0.2052</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>2.7184</td>\n",
       "      <td>0.2853</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>155</th>\n",
       "      <td>155</td>\n",
       "      <td>1.5595</td>\n",
       "      <td>0.7187</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>156</th>\n",
       "      <td>156</td>\n",
       "      <td>1.5717</td>\n",
       "      <td>0.7197</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>157</th>\n",
       "      <td>157</td>\n",
       "      <td>1.5647</td>\n",
       "      <td>0.7150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>158</th>\n",
       "      <td>158</td>\n",
       "      <td>1.5658</td>\n",
       "      <td>0.7184</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>159</th>\n",
       "      <td>159</td>\n",
       "      <td>1.5623</td>\n",
       "      <td>0.7197</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>160 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     epoch    loss    prec\n",
       "0        0  3.8293  0.0909\n",
       "1        1  3.5094  0.1405\n",
       "2        2  3.1706  0.1943\n",
       "3        3  3.2106  0.2052\n",
       "4        4  2.7184  0.2853\n",
       "..     ...     ...     ...\n",
       "155    155  1.5595  0.7187\n",
       "156    156  1.5717  0.7197\n",
       "157    157  1.5647  0.7150\n",
       "158    158  1.5658  0.7184\n",
       "159    159  1.5623  0.7197\n",
       "\n",
       "[160 rows x 3 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import os\n",
    "import pandas as pd\n",
    "\n",
    "filepath = r'G:\\lizulin\\network-slimming\\logs\\sparsity_vgg19_cifar10_s_1e_4'\n",
    "filepath2 = r'G:\\lizulin\\network-slimming\\logs\\fine_tuning_vgg19_cifar100_percent_0.5'\n",
    "filepath3 = r'G:\\lizulin\\network-slimming\\logs\\fine_tuning_vgg19_cifar100_percent_0.5_init'\n",
    "filepath4 = r'G:\\lizulin\\network-slimming\\logs\\sparsity_vgg19_cifar100_s_1e_4'\n",
    "\n",
    "data = pd.read_csv(os.path.join(filepath4, 'record.csv'))\n",
    "\n",
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "line_loss, = plt.plot(data['loss'], 'r-')\n",
    "line_prec, = plt.plot(data['prec'], 'b-')\n",
    "plt.legend([line_loss, line_prec], ['loss', 'accuracy'], loc='upper right')\n",
    "plt.ylabel('value', fontsize=12)\n",
    "plt.xlabel('epoch', fontsize=12)\n",
    "plt.title('Train loss and accuracy (best_prec: {:.4f})'.format(max(data['prec'])), fontsize=14)\n",
    "plt.savefig(os.path.join(filepath3, \"record train loss.png\"))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. VGG Weight Visualization "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.1 pruned model file "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['cfg', 'state_dict'])"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch\n",
    "\n",
    "filepath = r'G:\\lizulin\\network-slimming\\logs\\prune_vgg19_cifar100_percent_0.5\\pruned.pth.tar'\n",
    "\n",
    "pruned_model = torch.load(filepath)\n",
    "pruned_model.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"[40, 64, 'M', 128, 128, 'M', 254, 255, 256, 251, 'M', 436, 288, 86, 36, 'M', 41, 45, 59, 384]\""
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pruned_dict = pruned_model['state_dict']\n",
    "pruned_cfg = pruned_model['cfg']\n",
    "str(pruned_model['cfg'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.2 vgg model "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "import math\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "from torch.autograd import Variable\n",
    "\n",
    "__all__ = ['vgg']\n",
    "\n",
    "defaultcfg = {\n",
    "    11: [64, 'M', 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512],\n",
    "    13: [64, 64, 'M', 128, 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512],\n",
    "    16: [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 'M', 512, 512, 512, 'M', 512, 512, 512],\n",
    "    19: [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 256, 'M', 512, 512, 512, 512, 'M', 512, 512, 512, 512],\n",
    "}\n",
    "\n",
    "\n",
    "class vgg(nn.Module):\n",
    "    def __init__(self, dataset='cifar10', depth=19, init_weights=True, cfg=None, batch_norm=True):\n",
    "        super(vgg, self).__init__()\n",
    "        if cfg is None:\n",
    "            cfg = defaultcfg[depth]\n",
    "\n",
    "        self.feature = self.make_layers(cfg, batch_norm)\n",
    "\n",
    "        if dataset == 'cifar10':\n",
    "            num_classes = 10\n",
    "        elif dataset == 'cifar100':\n",
    "            num_classes = 100\n",
    "        self.classifier = nn.Linear(cfg[-1], num_classes)\n",
    "        if init_weights:\n",
    "            self._initialize_weights()\n",
    "\n",
    "    def make_layers(self, cfg, batch_norm=False):\n",
    "        layers = []\n",
    "        in_channels = 3\n",
    "        for v in cfg:\n",
    "            if v == 'M':\n",
    "                layers += [nn.MaxPool2d(kernel_size=2, stride=2)]\n",
    "            else:\n",
    "                conv2d = nn.Conv2d(in_channels, v, kernel_size=3, padding=1, bias=False)\n",
    "                if batch_norm:\n",
    "                    layers += [conv2d, nn.BatchNorm2d(v), nn.ReLU(inplace=True)]\n",
    "                else:\n",
    "                    layers += [conv2d, nn.ReLU(inplace=True)]\n",
    "                in_channels = v\n",
    "        return nn.Sequential(*layers)\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.feature(x)\n",
    "        x = nn.AvgPool2d(2)(x)\n",
    "        x = x.view(x.size(0), -1)\n",
    "        y = self.classifier(x)\n",
    "        return y4\n",
    "\n",
    "    def _initialize_weights(self):\n",
    "        for m in self.modules():\n",
    "            if isinstance(m, nn.Conv2d):\n",
    "                n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels\n",
    "                m.weight.data.normal_(0, math.sqrt(2. / n))  # mean, std\n",
    "                if m.bias is not None:\n",
    "                    m.bias.data.zero_()\n",
    "            elif isinstance(m, nn.BatchNorm2d):\n",
    "                m.weight.data.fill_(0.5)\n",
    "                m.bias.data.zero_()\n",
    "            elif isinstance(m, nn.Linear):\n",
    "                m.weight.data.normal_(0, 0.01)\n",
    "                m.bias.data.zero_()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.3 pruned model load parameters "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<All keys matched successfully>"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = vgg(dataset='cifar100', depth=19, cfg=pruned_cfg)\n",
    "model.load_state_dict(pruned_dict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.4 add cfg to model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['epoch', 'state_dict', 'best_prec1', 'optimizer'])"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch\n",
    "import os\n",
    "\n",
    "filepath = r'G:\\lizulin\\network-slimming\\logs\\sparsity_vgg19_cifar10_s_1e_4'\n",
    "filepath2 = r'G:\\lizulin\\network-slimming\\logs\\fine_tuning_vgg19_cifar100_percent_0.5'\n",
    "filepath3 = r'G:\\lizulin\\network-slimming\\logs\\fine_tuning_vgg19_cifar100_percent_0.5_init'\n",
    "filepath4 = r'G:\\lizulin\\network-slimming\\logs\\fine_tuning_vgg19_cifar100_feature_percent_0.5'\n",
    "\n",
    "path = filepath4\n",
    "file = os.path.join(path, 'model_best.pth.tar')\n",
    "\n",
    "model_best = torch.load(file)\n",
    "model_best.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['epoch', 'state_dict', 'best_prec1', 'optimizer', 'cfg'])"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cfg = [47, 64, 'M', 128, 128, 'M', 254, 254, 249, 251, 'M', 233, 244, 169, 173, 'M', 123, 124, 125, 185]\n",
    "model_best['cfg'] = cfg\n",
    "model_best.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save(model_best, file)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:python36]",
   "language": "python",
   "name": "conda-env-python36-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}