workshop.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pathlib import Path\n",
    "import numpy as np\n",
    "from PIL import Image\n",
    "import seaborn as sns\n",
    "from collections import defaultdict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1702, 2)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "rootdir = Path(\"crops_square/bird_dataset\")\n",
    "\n",
    "sizes = np.empty((0, 2))\n",
    "for imfn in rootdir.rglob('*.jpg'):\n",
    "    if imfn.name.split('.')[-2] == '00' or True:\n",
    "        im = Image.open(imfn)\n",
    "        w, h = im.size\n",
    "        sizes = np.concatenate((sizes, [[w, h]]), axis=0)\n",
    "        if w < 50 or h < 50:\n",
    "            print(imfn)\n",
    "print(sizes.shape)\n",
    "sns.scatterplot(x=sizes[:, 0], y=sizes[:, 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1082, 2)\n",
      "('014.Indigo_Bunting', 57)\n",
      "('023.Brandt_Cormorant', 53)\n",
      "('013.Bobolink', 58)\n",
      "('020.Yellow_breasted_Chat', 51)\n",
      "('016.Painted_Bunting', 52)\n",
      "('015.Lazuli_Bunting', 54)\n",
      "('034.Gray_crowned_Rosy_Finch', 51)\n",
      "('012.Yellow_headed_Blackbird', 54)\n",
      "('030.Fish_Crow', 54)\n",
      "('021.Eastern_Towhee', 54)\n",
      "('009.Brewer_Blackbird', 55)\n",
      "('004.Groove_billed_Ani', 58)\n",
      "('019.Gray_Catbird', 54)\n",
      "('033.Yellow_billed_Cuckoo', 55)\n",
      "('011.Rusty_Blackbird', 57)\n",
      "('026.Bronzed_Cowbird', 52)\n",
      "('031.Black_billed_Cuckoo', 55)\n",
      "('010.Red_winged_Blackbird', 53)\n",
      "('028.Brown_Creeper', 52)\n",
      "('029.American_Crow', 53)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "rootdir = Path(\"crops_square/bird_dataset/train_images\")\n",
    "\n",
    "classes = defaultdict(int)\n",
    "\n",
    "sizes = np.empty((0, 2))\n",
    "for imfn in rootdir.rglob('*.jpg'):\n",
    "    if imfn.name.split('.')[-2] == '00' or True:\n",
    "        classes[imfn.parent.name] += 1\n",
    "        im = Image.open(imfn)\n",
    "        w, h = im.size\n",
    "        sizes = np.concatenate((sizes, [[w, h]]), axis=0)\n",
    "        if w < 50 or h < 50:\n",
    "            print(imfn)\n",
    "print(sizes.shape)\n",
    "print(*classes.items(), sep='\\n')\n",
    "sns.scatterplot(x=sizes[:, 0], y=sizes[:, 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(103, 2)\n",
      "('014.Indigo_Bunting', 3)\n",
      "('023.Brandt_Cormorant', 6)\n",
      "('013.Bobolink', 2)\n",
      "('020.Yellow_breasted_Chat', 8)\n",
      "('016.Painted_Bunting', 6)\n",
      "('015.Lazuli_Bunting', 4)\n",
      "('034.Gray_crowned_Rosy_Finch', 8)\n",
      "('012.Yellow_headed_Blackbird', 2)\n",
      "('030.Fish_Crow', 6)\n",
      "('021.Eastern_Towhee', 6)\n",
      "('009.Brewer_Blackbird', 4)\n",
      "('004.Groove_billed_Ani', 2)\n",
      "('019.Gray_Catbird', 5)\n",
      "('033.Yellow_billed_Cuckoo', 4)\n",
      "('011.Rusty_Blackbird', 3)\n",
      "('026.Bronzed_Cowbird', 8)\n",
      "('031.Black_billed_Cuckoo', 5)\n",
      "('010.Red_winged_Blackbird', 7)\n",
      "('028.Brown_Creeper', 7)\n",
      "('029.American_Crow', 7)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "rootdir = Path(\"crops_square/bird_dataset/val_images\")\n",
    "\n",
    "classes = defaultdict(int)\n",
    "\n",
    "sizes = np.empty((0, 2))\n",
    "for imfn in rootdir.rglob('*.jpg'):\n",
    "    if imfn.name.split('.')[-2] == '00' or True:\n",
    "        classes[imfn.parent.name] += 1\n",
    "        im = Image.open(imfn)\n",
    "        w, h = im.size\n",
    "        sizes = np.concatenate((sizes, [[w, h]]), axis=0)\n",
    "        if w < 50 or h < 50:\n",
    "            print(imfn)\n",
    "print(sizes.shape)\n",
    "print(*classes.items(), sep='\\n')\n",
    "sns.scatterplot(x=sizes[:, 0], y=sizes[:, 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(517, 2)\n",
      "('mistery_category', 517)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "rootdir = Path(\"crops_square/bird_dataset/test_images\")\n",
    "\n",
    "classes = defaultdict(int)\n",
    "\n",
    "sizes = np.empty((0, 2))\n",
    "for imfn in rootdir.rglob('*.jpg'):\n",
    "    if imfn.name.split('.')[-2] == '00' or True:\n",
    "        classes[imfn.parent.name] += 1\n",
    "        im = Image.open(imfn)\n",
    "        w, h = im.size\n",
    "        sizes = np.concatenate((sizes, [[w, h]]), axis=0)\n",
    "        if w < 50 or h < 50:\n",
    "            print(imfn)\n",
    "print(sizes.shape)\n",
    "print(*classes.items(), sep='\\n')\n",
    "sns.scatterplot(x=sizes[:, 0], y=sizes[:, 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torchvision import models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def n_params(model):\n",
    "    return sum(p.numel() for p in model.parameters())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "m_list = [\n",
    "    \"vgg11\",\n",
    "    \"vgg16\",\n",
    "    \"vgg19\",\n",
    "    \"resnet18\",\n",
    "    \"resnet34\",\n",
    "    \"resnet50\",\n",
    "    \"resnet101\",\n",
    "    \"resnet152\",\n",
    "    \"squeezenet1_0\",\n",
    "    \"squeezenet1_1\",\n",
    "    \"densenet121\",\n",
    "    \"densenet169\",\n",
    "    \"densenet161\",\n",
    "    \"densenet201\",\n",
    "    \"inception_v3\",\n",
    "    \"googlenet\",\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "vgg11 1.328633e+08\n",
      "vgg16 1.383575e+08\n",
      "vgg19 1.436672e+08\n",
      "resnet18 1.168951e+07\n",
      "resnet34 2.179767e+07\n",
      "resnet50 2.555703e+07\n",
      "resnet101 4.454916e+07\n",
      "resnet152 6.019281e+07\n",
      "squeezenet1_0 1.248424e+06\n",
      "squeezenet1_1 1.235496e+06\n",
      "densenet121 7.978856e+06\n",
      "densenet169 1.414948e+07\n",
      "densenet161 2.868100e+07\n",
      "densenet201 2.001393e+07\n",
      "inception_v3 2.716126e+07\n",
      "googlenet 1.300489e+07\n"
     ]
    }
   ],
   "source": [
    "model_sizes = {}\n",
    "for m_name in m_list:\n",
    "    m = models.__dict__[m_name]()\n",
    "    model_sizes[m_name] = n_params(m)\n",
    "    print(m_name, f'{model_sizes[m_name]:e}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "vgg11 Sequential(\n",
      "  (0): Linear(in_features=25088, out_features=4096, bias=True)\n",
      "  (1): ReLU(inplace=True)\n",
      "  (2): Dropout(p=0.5, inplace=False)\n",
      "  (3): Linear(in_features=4096, out_features=4096, bias=True)\n",
      "  (4): ReLU(inplace=True)\n",
      "  (5): Dropout(p=0.5, inplace=False)\n",
      "  (6): Linear(in_features=4096, out_features=1000, bias=True)\n",
      ")\n",
      "vgg16 Sequential(\n",
      "  (0): Linear(in_features=25088, out_features=4096, bias=True)\n",
      "  (1): ReLU(inplace=True)\n",
      "  (2): Dropout(p=0.5, inplace=False)\n",
      "  (3): Linear(in_features=4096, out_features=4096, bias=True)\n",
      "  (4): ReLU(inplace=True)\n",
      "  (5): Dropout(p=0.5, inplace=False)\n",
      "  (6): Linear(in_features=4096, out_features=1000, bias=True)\n",
      ")\n",
      "vgg19 Sequential(\n",
      "  (0): Linear(in_features=25088, out_features=4096, bias=True)\n",
      "  (1): ReLU(inplace=True)\n",
      "  (2): Dropout(p=0.5, inplace=False)\n",
      "  (3): Linear(in_features=4096, out_features=4096, bias=True)\n",
      "  (4): ReLU(inplace=True)\n",
      "  (5): Dropout(p=0.5, inplace=False)\n",
      "  (6): Linear(in_features=4096, out_features=1000, bias=True)\n",
      ")\n",
      "resnet18 Linear(in_features=512, out_features=1000, bias=True)\n",
      "resnet34 Linear(in_features=512, out_features=1000, bias=True)\n",
      "resnet50 Linear(in_features=2048, out_features=1000, bias=True)\n",
      "resnet101 Linear(in_features=2048, out_features=1000, bias=True)\n",
      "resnet152 Linear(in_features=2048, out_features=1000, bias=True)\n",
      "squeezenet1_0 Sequential(\n",
      "  (0): Dropout(p=0.5, inplace=False)\n",
      "  (1): Conv2d(512, 1000, kernel_size=(1, 1), stride=(1, 1))\n",
      "  (2): ReLU(inplace=True)\n",
      "  (3): AdaptiveAvgPool2d(output_size=(1, 1))\n",
      ")\n",
      "squeezenet1_1 Sequential(\n",
      "  (0): Dropout(p=0.5, inplace=False)\n",
      "  (1): Conv2d(512, 1000, kernel_size=(1, 1), stride=(1, 1))\n",
      "  (2): ReLU(inplace=True)\n",
      "  (3): AdaptiveAvgPool2d(output_size=(1, 1))\n",
      ")\n",
      "densenet121 Linear(in_features=1024, out_features=1000, bias=True)\n",
      "densenet169 Linear(in_features=1664, out_features=1000, bias=True)\n",
      "densenet161 Linear(in_features=2208, out_features=1000, bias=True)\n",
      "densenet201 Linear(in_features=1920, out_features=1000, bias=True)\n",
      "inception_v3 Linear(in_features=2048, out_features=1000, bias=True)\n",
      "googlenet Linear(in_features=1024, out_features=1000, bias=True)\n"
     ]
    }
   ],
   "source": [
    "model_sizes = {}\n",
    "for m_name in m_list:\n",
    "    m = models.__dict__[m_name]()\n",
    "    model_sizes[m_name] = n_params(m)\n",
    "    print(m_name, list(m.children())[-1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GoogLeNet(\n",
       "  (conv1): BasicConv2d(\n",
       "    (conv): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
       "    (bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  )\n",
       "  (maxpool1): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)\n",
       "  (conv2): BasicConv2d(\n",
       "    (conv): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "    (bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  )\n",
       "  (conv3): BasicConv2d(\n",
       "    (conv): Conv2d(64, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "    (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  )\n",
       "  (maxpool2): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)\n",
       "  (inception3a): Inception(\n",
       "    (branch1): BasicConv2d(\n",
       "      (conv): Conv2d(192, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (branch2): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(192, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(96, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch3): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(192, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(16, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(16, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch4): Sequential(\n",
       "      (0): MaxPool2d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=True)\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(192, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (inception3b): Inception(\n",
       "    (branch1): BasicConv2d(\n",
       "      (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (branch2): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(128, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch3): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(32, 96, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch4): Sequential(\n",
       "      (0): MaxPool2d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=True)\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (maxpool3): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)\n",
       "  (inception4a): Inception(\n",
       "    (branch1): BasicConv2d(\n",
       "      (conv): Conv2d(480, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (branch2): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(480, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(96, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(96, 208, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(208, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch3): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(480, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(16, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(16, 48, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(48, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch4): Sequential(\n",
       "      (0): MaxPool2d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=True)\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(480, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (inception4b): Inception(\n",
       "    (branch1): BasicConv2d(\n",
       "      (conv): Conv2d(512, 160, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (branch2): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(512, 112, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(112, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(112, 224, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(224, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch3): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(512, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(24, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(24, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch4): Sequential(\n",
       "      (0): MaxPool2d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=True)\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(512, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (inception4c): Inception(\n",
       "    (branch1): BasicConv2d(\n",
       "      (conv): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (branch2): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(256, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch3): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(512, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(24, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(24, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch4): Sequential(\n",
       "      (0): MaxPool2d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=True)\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(512, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (inception4d): Inception(\n",
       "    (branch1): BasicConv2d(\n",
       "      (conv): Conv2d(512, 112, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn): BatchNorm2d(112, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (branch2): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(512, 144, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(144, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(144, 288, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(288, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch3): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(512, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch4): Sequential(\n",
       "      (0): MaxPool2d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=True)\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(512, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(64, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (inception4e): Inception(\n",
       "    (branch1): BasicConv2d(\n",
       "      (conv): Conv2d(528, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn): BatchNorm2d(256, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (branch2): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(528, 160, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(160, 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(320, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch3): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(528, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(32, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch4): Sequential(\n",
       "      (0): MaxPool2d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=True)\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(528, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (maxpool4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=True)\n",
       "  (inception5a): Inception(\n",
       "    (branch1): BasicConv2d(\n",
       "      (conv): Conv2d(832, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn): BatchNorm2d(256, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (branch2): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(832, 160, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(160, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(160, 320, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(320, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch3): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(832, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(32, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(32, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch4): Sequential(\n",
       "      (0): MaxPool2d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=True)\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(832, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (inception5b): Inception(\n",
       "    (branch1): BasicConv2d(\n",
       "      (conv): Conv2d(832, 384, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (branch2): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(832, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(192, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(192, 384, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(384, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch3): Sequential(\n",
       "      (0): BasicConv2d(\n",
       "        (conv): Conv2d(832, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(48, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(48, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (branch4): Sequential(\n",
       "      (0): MaxPool2d(kernel_size=3, stride=1, padding=1, dilation=1, ceil_mode=True)\n",
       "      (1): BasicConv2d(\n",
       "        (conv): Conv2d(832, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "        (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (aux1): InceptionAux(\n",
       "    (conv): BasicConv2d(\n",
       "      (conv): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (fc1): Linear(in_features=2048, out_features=1024, bias=True)\n",
       "    (fc2): Linear(in_features=1024, out_features=1000, bias=True)\n",
       "  )\n",
       "  (aux2): InceptionAux(\n",
       "    (conv): BasicConv2d(\n",
       "      (conv): Conv2d(528, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "      (bn): BatchNorm2d(128, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (fc1): Linear(in_features=2048, out_features=1024, bias=True)\n",
       "    (fc2): Linear(in_features=1024, out_features=1000, bias=True)\n",
       "  )\n",
       "  (avgpool): AdaptiveAvgPool2d(output_size=(1, 1))\n",
       "  (dropout): Dropout(p=0.2, inplace=False)\n",
       "  (fc): Linear(in_features=1024, out_features=1000, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "models.googlenet()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ResNet(\n",
       "  (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
       "  (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "  (relu): ReLU(inplace=True)\n",
       "  (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
       "  (layer1): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (layer2): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (layer3): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (layer4): Sequential(\n",
       "    (0): BasicBlock(\n",
       "      (conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (downsample): Sequential(\n",
       "        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
       "        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      )\n",
       "    )\n",
       "    (1): BasicBlock(\n",
       "      (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (relu): ReLU(inplace=True)\n",
       "      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "    )\n",
       "  )\n",
       "  (avgpool): AdaptiveAvgPool2d(output_size=(1, 1))\n",
       "  (fc): Linear(in_features=512, out_features=1000, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "models.resnet18()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "m = models.resnet18()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "512"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "m.fc.in_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "VGG(\n",
       "  (features): Sequential(\n",
       "    (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (1): ReLU(inplace=True)\n",
       "    (2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (3): ReLU(inplace=True)\n",
       "    (4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "    (5): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (6): ReLU(inplace=True)\n",
       "    (7): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (8): ReLU(inplace=True)\n",
       "    (9): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "    (10): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (11): ReLU(inplace=True)\n",
       "    (12): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (13): ReLU(inplace=True)\n",
       "    (14): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (15): ReLU(inplace=True)\n",
       "    (16): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "    (17): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (18): ReLU(inplace=True)\n",
       "    (19): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (20): ReLU(inplace=True)\n",
       "    (21): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (22): ReLU(inplace=True)\n",
       "    (23): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "    (24): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (25): ReLU(inplace=True)\n",
       "    (26): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (27): ReLU(inplace=True)\n",
       "    (28): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (29): ReLU(inplace=True)\n",
       "    (30): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "  )\n",
       "  (avgpool): AdaptiveAvgPool2d(output_size=(7, 7))\n",
       "  (classifier): Sequential(\n",
       "    (0): Linear(in_features=25088, out_features=4096, bias=True)\n",
       "    (1): ReLU(inplace=True)\n",
       "    (2): Dropout(p=0.5, inplace=False)\n",
       "    (3): Linear(in_features=4096, out_features=4096, bias=True)\n",
       "    (4): ReLU(inplace=True)\n",
       "    (5): Dropout(p=0.5, inplace=False)\n",
       "    (6): Linear(in_features=4096, out_features=1000, bias=True)\n",
       "  )\n",
       ")"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "models.vgg16()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SqueezeNet(\n",
       "  (features): Sequential(\n",
       "    (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(2, 2))\n",
       "    (1): ReLU(inplace=True)\n",
       "    (2): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)\n",
       "    (3): Fire(\n",
       "      (squeeze): Conv2d(64, 16, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(16, 64, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(16, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (4): Fire(\n",
       "      (squeeze): Conv2d(128, 16, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(16, 64, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(16, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (5): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)\n",
       "    (6): Fire(\n",
       "      (squeeze): Conv2d(128, 32, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(32, 128, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(32, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (7): Fire(\n",
       "      (squeeze): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(32, 128, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(32, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (8): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)\n",
       "    (9): Fire(\n",
       "      (squeeze): Conv2d(256, 48, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(48, 192, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(48, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (10): Fire(\n",
       "      (squeeze): Conv2d(384, 48, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(48, 192, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(48, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (11): Fire(\n",
       "      (squeeze): Conv2d(384, 64, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(64, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (12): Fire(\n",
       "      (squeeze): Conv2d(512, 64, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(64, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "  )\n",
       "  (classifier): Sequential(\n",
       "    (0): Dropout(p=0.5, inplace=False)\n",
       "    (1): Conv2d(512, 1000, kernel_size=(1, 1), stride=(1, 1))\n",
       "    (2): ReLU(inplace=True)\n",
       "    (3): AdaptiveAvgPool2d(output_size=(1, 1))\n",
       "  )\n",
       ")"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "models.squeezenet1_1()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'2020-11-24T19:53:06.124246'"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import datetime\n",
    "datetime.datetime.now().isoformat()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.8.5 64-bit ('MASH': conda)",
   "language": "python",
   "name": "python38564bitmashconda7a43979bf891458b9945659045d4d708"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}