notebooks/CIFAR10_VGG11_volume_estimation.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "sys.path.insert(0,'..')\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "import torch.backends.cudnn as cudnn\n",
    "import torchvision.datasets as datasets\n",
    "import torchvision.transforms as transforms\n",
    "import pandas as pd\n",
    "from tqdm.notebook import tqdm\n",
    "\n",
    "from resnet import resnet18\n",
    "from vgg import VGG\n",
    "from relu import ReLUAlpha\n",
    "\n",
    "\n",
    "def init(batchnorm, relu=0):\n",
    "    if batchnorm:\n",
    "        net = VGG('VGG11', batch_norm=True, relu_fn=lambda: ReLUAlpha(relu)).to(device)\n",
    "    else:\n",
    "        net = VGG('VGG11', relu_fn=lambda: ReLUAlpha(relu)).to(device)\n",
    "    return net\n",
    "\n",
    "def weight_reset(m):\n",
    "    if isinstance(m, nn.Conv2d) or isinstance(m, nn.Linear):\n",
    "        m.reset_parameters()\n",
    "\n",
    "def copy_params(net1, net2):\n",
    "    params1 = net1.named_parameters()\n",
    "    params2 = net2.named_parameters()\n",
    "\n",
    "    dict_params2 = dict(params2)\n",
    "    for name, param in params1:\n",
    "        if name in dict_params2:\n",
    "            dict_params2[name].data.copy_(param.data)\n",
    "\n",
    "def rebuild_model(net1, net2):\n",
    "    net1.apply(weight_reset)\n",
    "    copy_params(net1, net2)\n",
    "            \n",
    "def build_models(batchnorm, precision=32):\n",
    "    net1 = init(batchnorm, 0)\n",
    "    net2 = init(batchnorm, 1)\n",
    "    copy_params(net1, net2)\n",
    "\n",
    "    for net in [net1, net2]:\n",
    "        net.to(device)\n",
    "        if precision == 16:\n",
    "            net.half()\n",
    "        if precision == 64:\n",
    "            net = net.double()\n",
    "        \n",
    "    return net1, net2\n",
    "\n",
    "def get_cifar10_train_loader():\n",
    "    transform = transforms.Compose([\n",
    "        transforms.RandomCrop(32, padding=4),\n",
    "        transforms.RandomHorizontalFlip(),\n",
    "        transforms.ToTensor(),\n",
    "        transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010)),\n",
    "    ])\n",
    "    trainset = datasets.CIFAR10(root='./data', train=True, download=True, transform=transform)\n",
    "    return torch.utils.data.DataLoader(trainset, batch_size=128, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Files already downloaded and verified\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "74ee764d23514a559c6d30499de236b3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "precision:   0%|          | 0/2 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "batchnorm:   0%|          | 0/2 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "run:   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "run:   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "batchnorm:   0%|          | 0/2 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "run:   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "run:   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "nb_experiments = 100 #full paper experiment: nb_experiments = 1000\n",
    "\n",
    "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
    "if device == 'cuda':\n",
    "    cudnn.benchmark = False\n",
    "    # Deterministic convolutions\n",
    "    torch.backends.cudnn.deterministic = True\n",
    "\n",
    "trainloader = get_cifar10_train_loader()\n",
    "df = pd.DataFrame(columns=[\"network\", \"batchnorm\", \"same\", \"precision\", \"batch_idx\", \"run_id\",\n",
    "                           \"net1_L1\", \"net1_L2\", \"net2_L1\", \"net2_L2\", \"diff_L1\", \"diff_L2\"])\n",
    "\n",
    "for precision in tqdm([16, 32], desc='precision'):\n",
    "    for batchnorm in tqdm([False, True], desc='batchnorm', leave=False):\n",
    "        count = 0\n",
    "        net1, net2 = build_models(batchnorm, precision)\n",
    "        opt = optim.SGD([{'params': net1.parameters()}, {'params': net2.parameters()}], lr=0)\n",
    "        for run_id in tqdm(range(nb_experiments), desc='run', leave=False):\n",
    "            rebuild_model(net1, net2)\n",
    "            # Optimizers are just used to free the gradients accumulated during the backward pass\n",
    "            # No optimization step is performed\n",
    "            opt.zero_grad()\n",
    "            for i, (x, y) in enumerate(trainloader):\n",
    "                x = x.to(device)\n",
    "                if precision == 16:\n",
    "                    x = x.half()\n",
    "                if precision == 64:\n",
    "                    x = x.double()\n",
    "                y = y.to(device)\n",
    "                same = True\n",
    "                y1, y2 = net1(x), net2(x)\n",
    "                loss = nn.CrossEntropyLoss()(y1, y) + nn.CrossEntropyLoss()(y2, y)\n",
    "                loss.backward()\n",
    "                if not torch.equal(net1.features[0].weight.grad, net2.features[0].weight.grad):\n",
    "                    same = False\n",
    "                diff = net1.features[0].weight.grad - net2.features[0].weight.grad\n",
    "                df = df.append({\"run_id\": run_id,\n",
    "                                \"network\": 'resnet',\n",
    "                                \"batchnorm\": batchnorm,\n",
    "                                \"same\": same,\n",
    "                                \"precision\": precision,\n",
    "                                \"net1_L1\": torch.linalg.norm(net1.features[0].weight.grad.view(-1), 1).item(),\n",
    "                                \"net1_L2\": torch.linalg.norm(net1.features[0].weight.grad.view(-1), 2).item(),\n",
    "                                \"net2_L1\": torch.linalg.norm(net2.features[0].weight.grad.view(-1), 1).item(),\n",
    "                                \"net2_L2\": torch.linalg.norm(net2.features[0].weight.grad.view(-1), 2).item(),\n",
    "                                \"diff_L1\": torch.linalg.norm(diff.view(-1), 1),\n",
    "                                \"diff_L2\": torch.linalg.norm(diff.view(-1), 2),\n",
    "                                \"batch_idx\": i}, ignore_index=True)\n",
    "                if same == False:\n",
    "                    break\n",
    "\n",
    "df.to_csv('volume_estimation_sample_vgg11.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('../paper_results/section_4/cifar10/vgg11/volume_estimation_sample_vgg11.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def prepare_volume_plot(df, max_size, batch_size):\n",
    "    dfs = []\n",
    "    #for precision in [16, 32, 64]:\n",
    "    for precision in [16, 32]:\n",
    "        tmp = df[df['precision'] == precision]\n",
    "        max_len = len(tmp.run_id.unique())\n",
    "        tmp = tmp[['batch_idx', 'same']].groupby(['batch_idx']).sum().reset_index()\n",
    "        tmp['div'] = (max_len-tmp['same'])/max_len\n",
    "        for i in range(tmp.batch_idx.count(), max_size):\n",
    "            tmp = tmp.append({\"div\": 1.,\n",
    "                                \"same\": max_len,\n",
    "                                \"batch_idx\": i}, ignore_index=True)\n",
    "        tmp['batch_idx'] = tmp['batch_idx'] * batch_size\n",
    "        tmp = tmp.astype({\"div\":float})\n",
    "        dfs.append(tmp)\n",
    "    #return dfs[0], dfs[1], dfs[2]\n",
    "    return dfs[0], dfs[1]\n",
    "\n",
    "def plot_volume_estimation(df, max_size, batch_size=128):\n",
    "    #df16, df32, df64 = prepare_volume_plot(df, max_size, batch_size)\n",
    "    df16, df32 = prepare_volume_plot(df, max_size, batch_size)\n",
    "    plt.figure(figsize=(15,5))\n",
    "    #plt.plot(df64['batch_idx'], df64['div'].values, label='64 bits')\n",
    "    plt.plot(df32['batch_idx'], df32['div'].values, label='32 bits')\n",
    "    plt.plot(df16['batch_idx'], df16['div'].values, label='16 bits')\n",
    "    #plt.fill_between(df64['batch_idx'].values, df64['div'].values, alpha=.3)\n",
    "    plt.fill_between(df32['batch_idx'].values, df32['div'].values, alpha=0.3)\n",
    "    plt.fill_between(df16['batch_idx'].values, df16['div'].values, alpha=.3)\n",
    "    leg = plt.legend(title=\"Precision\", fontsize='25', title_fontsize='25', loc='lower right')\n",
    "    plt.xlabel('sample size', fontsize='25');\n",
    "    plt.ylabel('Ratio of ' + r'$\\theta \\in S$', fontsize='30');plt.yticks(fontsize='25');\n",
    "    plt.xticks(fontsize='25');\n",
    "    for line in leg.get_lines():\n",
    "        line.set_linewidth(10.0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_volume_estimation(df[df['batchnorm']==True], 100, batch_size=128)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_volume_estimation(df[df['batchnorm']==False], 100, batch_size=128)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": "3.7.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}