Dev/03_Darknet.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### https://github.com/emilmelnikov/darknet53-pytorch/blob/master/darknet53.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import functools\n",
    "\n",
    "import torch\n",
    "from torch import nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "\n",
    "def applyseq(x, *fs):\n",
    "    \"\"\"Consecutively apply functions in a list to a value (see the examples).\n",
    "    >>> applyseq(1)\n",
    "    1\n",
    "    >>> applyseq(1, lambda x: x+1)\n",
    "    2\n",
    "    >>> applyseq(1, lambda x: x+1, lambda x: bin(x))\n",
    "    '0b10'\n",
    "    \"\"\"\n",
    "    return functools.reduce(lambda a, f: f(a), fs, x)\n",
    "\n",
    "\n",
    "def global_avg_pool2d(x):\n",
    "    \"\"\"Global average pooling.\"\"\"\n",
    "    return F.adaptive_avg_pool2d(x.view(x.size(0), -1), 1)\n",
    "\n",
    "\n",
    "class SameConvBlock(nn.Module):\n",
    "    \"\"\"Size-preserving convolution, followed by BatchNorm and activation.\"\"\"\n",
    "\n",
    "    def __init__(self, cin, cout, *, size, stride):\n",
    "        super().__init__()\n",
    "        self.conv = nn.Conv2d(cin, cout, size, stride=stride, padding=size//2, bias=False)\n",
    "        self.bn = nn.BatchNorm1d(cout, momentum=0.01)\n",
    "\n",
    "    def forward(self, x):\n",
    "        return F.leaky_relu(self.bn(self.conv(x)), negative_slope=0.1)\n",
    "\n",
    "\n",
    "class ConvBlock(SameConvBlock):\n",
    "    \"\"\"Ordinary 3x3 convolution block.\"\"\"\n",
    "    def __init__(self, cin, cout):\n",
    "        super().__init__(cin, cout, size=3, stride=1)\n",
    "\n",
    "\n",
    "class ReduceConvBlock(SameConvBlock):\n",
    "    \"\"\"Dimensionality-reducing 1x1 convolution block.\"\"\"\n",
    "    def __init__(self, cin, cout):\n",
    "        super().__init__(cin, cout, size=1, stride=1)\n",
    "\n",
    "\n",
    "class DownConvBlock(SameConvBlock):\n",
    "    \"\"\"Downsampling 3x3 convolution, divides the size by 2.\"\"\"\n",
    "    def __init__(self, cin, cout):\n",
    "        super().__init__(cin, cout, size=3, stride=2)\n",
    "\n",
    "\n",
    "class ResBlock(nn.Module):\n",
    "    \"\"\"Residual block. The residual path has an 1x1 and 3x3 convolution blocks.\"\"\"\n",
    "\n",
    "    def __init__(self, cin):\n",
    "        super().__init__()\n",
    "        self.convreduce = ReduceConvBlock(cin, cin//2)\n",
    "        self.conv = ConvBlock(cin//2, cin)\n",
    "\n",
    "    def forward(self, x):\n",
    "        return x + self.conv(self.convreduce(x))\n",
    "\n",
    "\n",
    "class DarkNetBlock(nn.Module):\n",
    "    \"\"\"Basic block for the DarkNet: downsampling followed by multiple residual blocks.\"\"\"\n",
    "\n",
    "    def __init__(self, cin, n):\n",
    "        super().__init__()\n",
    "        self.convdown = DownConvBlock(cin, 2*cin)\n",
    "        self.reslist = nn.ModuleList([ResBlock(2*cin) for i in range(n)])\n",
    "\n",
    "    def forward(self, x):\n",
    "        return applyseq(x, self.convdown, *self.reslist)\n",
    "\n",
    "\n",
    "class DarkNet(nn.Module):\n",
    "    \"\"\"DarkNet network.\n",
    "    - First, expand channels from `cin` to `c1`.\n",
    "    - Then, apply a sequence of DarkNetBlock modules; each element in the `rbs` list adds\n",
    "      a single such block with the number of residual blocks equal to the element's value.\n",
    "    - Finally, apply adaptive average pooling to the `nf` features.\n",
    "    \"\"\"\n",
    "\n",
    "    def __init__(self, cin, c1, rbs, nf):\n",
    "        super().__init__()\n",
    "        self.conv1 = ConvBlock(cin, c1)\n",
    "        self.dns = nn.ModuleList([DarkNetBlock(2**i * c1, n) for i, n in enumerate(rbs)])\n",
    "        self.fc = nn.Linear(2**len(rbs) * c1, nf)\n",
    "\n",
    "    def forward(self, x):\n",
    "        return applyseq(x, self.conv1, *self.dns, global_avg_pool2d, self.fc)\n",
    "\n",
    "\n",
    "def darknet53():\n",
    "    \"\"\"Darknet-53 with parameters defined in the paper [YOLOv3: An Incremental Improvement](https://pjreddie.com/media/files/papers/YOLOv3.pdf).\"\"\"\n",
    "    return DarkNet(cin=3, c1=32, rbs=(1, 2, 8, 8, 4), nf=1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "net = darknet53()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![a](darknet53.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DarkNet(\n",
       "  (conv1): ConvBlock(\n",
       "    (conv): Conv2d(3, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "    (bn): BatchNorm1d(32, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "  )\n",
       "  (dns): ModuleList(\n",
       "    (0): DarkNetBlock(\n",
       "      (convdown): DownConvBlock(\n",
       "        (conv): Conv2d(32, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm1d(64, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (reslist): ModuleList(\n",
       "        (0): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(64, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(32, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(64, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (1): DarkNetBlock(\n",
       "      (convdown): DownConvBlock(\n",
       "        (conv): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (reslist): ModuleList(\n",
       "        (0): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(128, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(64, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(128, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (1): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(128, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(64, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(128, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (2): DarkNetBlock(\n",
       "      (convdown): DownConvBlock(\n",
       "        (conv): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (reslist): ModuleList(\n",
       "        (0): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(128, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (1): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(128, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (2): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(128, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (3): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(128, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (4): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(128, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (5): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(128, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (6): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(128, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (7): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(128, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (3): DarkNetBlock(\n",
       "      (convdown): DownConvBlock(\n",
       "        (conv): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (reslist): ModuleList(\n",
       "        (0): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (1): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (2): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (3): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (4): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (5): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (6): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (7): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(256, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "    (4): DarkNetBlock(\n",
       "      (convdown): DownConvBlock(\n",
       "        (conv): Conv2d(512, 1024, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
       "        (bn): BatchNorm1d(1024, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "      )\n",
       "      (reslist): ModuleList(\n",
       "        (0): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(512, 1024, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(1024, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (1): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(512, 1024, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(1024, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (2): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(512, 1024, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(1024, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "        (3): ResBlock(\n",
       "          (convreduce): ReduceConvBlock(\n",
       "            (conv): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "          (conv): ConvBlock(\n",
       "            (conv): Conv2d(512, 1024, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
       "            (bn): BatchNorm1d(1024, eps=1e-05, momentum=0.01, affine=True, track_running_stats=True)\n",
       "          )\n",
       "        )\n",
       "      )\n",
       "    )\n",
       "  )\n",
       "  (fc): Linear(in_features=1024, out_features=1000, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "net"
   ]
  },
  {
   "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": "3.6.7"
  },
  "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
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}