Implementing Darknet19 from scratch using fast.ai - MNIST

fastai-darknet19-mnist.ipynb

{
  "cells": [
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Darknet19 - MNIST"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "_Adapted from Lesson 7 of fast.ai course-v3_"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "This is the initial setup. Scroll down for the Darknet implementation. "
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "%reload_ext autoreload\n%autoreload 2\n%matplotlib inline",
      "execution_count": 1,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "from fastai.vision import *",
      "execution_count": 2,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "### Data"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "path = untar_data(URLs.MNIST)",
      "execution_count": 3,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "path.ls()",
      "execution_count": 4,
      "outputs": [
        {
          "execution_count": 4,
          "output_type": "execute_result",
          "data": {
            "text/plain": "[PosixPath('/home/jupyter/.fastai/data/mnist_png/testing'),\n PosixPath('/home/jupyter/.fastai/data/mnist_png/training'),\n PosixPath('/home/jupyter/.fastai/data/mnist_png/models')]"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "il = ImageList.from_folder(path, convert_mode='L')",
      "execution_count": 5,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "il.items[0]",
      "execution_count": 6,
      "outputs": [
        {
          "execution_count": 6,
          "output_type": "execute_result",
          "data": {
            "text/plain": "PosixPath('/home/jupyter/.fastai/data/mnist_png/testing/1/8914.png')"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "defaults.cmap='binary'",
      "execution_count": 7,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "il",
      "execution_count": 8,
      "outputs": [
        {
          "execution_count": 8,
          "output_type": "execute_result",
          "data": {
            "text/plain": "ImageList (70000 items)\nImage (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28)\nPath: /home/jupyter/.fastai/data/mnist_png"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "il[0].show()",
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 216x216 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "sd = il.split_by_folder(train='training', valid='testing')",
      "execution_count": 10,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "sd",
      "execution_count": 11,
      "outputs": [
        {
          "execution_count": 11,
          "output_type": "execute_result",
          "data": {
            "text/plain": "ItemLists;\n\nTrain: ImageList (60000 items)\nImage (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28)\nPath: /home/jupyter/.fastai/data/mnist_png;\n\nValid: ImageList (10000 items)\nImage (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28)\nPath: /home/jupyter/.fastai/data/mnist_png;\n\nTest: None"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "(path/'training').ls()",
      "execution_count": 12,
      "outputs": [
        {
          "execution_count": 12,
          "output_type": "execute_result",
          "data": {
            "text/plain": "[PosixPath('/home/jupyter/.fastai/data/mnist_png/training/1'),\n PosixPath('/home/jupyter/.fastai/data/mnist_png/training/6'),\n PosixPath('/home/jupyter/.fastai/data/mnist_png/training/8'),\n PosixPath('/home/jupyter/.fastai/data/mnist_png/training/2'),\n PosixPath('/home/jupyter/.fastai/data/mnist_png/training/5'),\n PosixPath('/home/jupyter/.fastai/data/mnist_png/training/9'),\n PosixPath('/home/jupyter/.fastai/data/mnist_png/training/0'),\n PosixPath('/home/jupyter/.fastai/data/mnist_png/training/3'),\n PosixPath('/home/jupyter/.fastai/data/mnist_png/training/7'),\n PosixPath('/home/jupyter/.fastai/data/mnist_png/training/4')]"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "ll = sd.label_from_folder()",
      "execution_count": 13,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "ll",
      "execution_count": 14,
      "outputs": [
        {
          "execution_count": 14,
          "output_type": "execute_result",
          "data": {
            "text/plain": "LabelLists;\n\nTrain: LabelList (60000 items)\nx: ImageList\nImage (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28)\ny: CategoryList\n1,1,1,1,1\nPath: /home/jupyter/.fastai/data/mnist_png;\n\nValid: LabelList (10000 items)\nx: ImageList\nImage (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28),Image (1, 28, 28)\ny: CategoryList\n1,1,1,1,1\nPath: /home/jupyter/.fastai/data/mnist_png;\n\nTest: None"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "x,y = ll.train[0]  # input and output (label)",
      "execution_count": 15,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "x.show()\nprint(y,x.shape)",
      "execution_count": 16,
      "outputs": [
        {
          "text": "1 torch.Size([1, 28, 28])\n",
          "output_type": "stream",
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 216x216 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "tfms = ([*rand_pad(padding=3, size=28, mode='zeros')], [])  # only minimal tfms",
      "execution_count": 17,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "The `*` above unpacks the cropping and padding transforms that return from `rand_pad()` into the surrounding list. The empty array `[]` indicates __no transforms for the validation set__."
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "ll = ll.transform(tfms)  # append the tfms",
      "execution_count": 18,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "bs = 128",
      "execution_count": 19,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "# not using imagenet_stats because not using pretrained model\ndata = ll.databunch(bs=bs).normalize()  # normalize according to the values in the dataset",
      "execution_count": 20,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "x,y = data.train_ds[0]",
      "execution_count": 21,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "x.show()\nprint(y)",
      "execution_count": 22,
      "outputs": [
        {
          "text": "1\n",
          "output_type": "stream",
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 216x216 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "Data augmentation occurs on-the-fly when the object itself is called; thus, every time the `data.train_ds[0][0]` object is called, it will produce a new version with tfms applied. `plot_multi` will then call that object multiple times, showing our tfms. "
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "def _plot(i,j,ax): data.train_ds[0][0].show(ax, cmap='gray')  # i is row, j is col, ax is the figsize\nplot_multi(_plot, 3, 3, figsize=(8,8))",
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 576x576 with 9 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "xb,yb = data.one_batch()\nxb.shape,yb.shape",
      "execution_count": 24,
      "outputs": [
        {
          "execution_count": 24,
          "output_type": "execute_result",
          "data": {
            "text/plain": "(torch.Size([128, 1, 28, 28]), torch.Size([128]))"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "data.show_batch(rows=3, figsize=(5,5))",
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 360x360 with 9 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "---"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Implementing Darknet19 from scratch"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "https://github.com/pjreddie/darknet/blob/master/cfg/darknet19.cfg"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "Best results are found at the bottom. "
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "def dn_maxp():\n    return nn.MaxPool2d(2, stride=2)\n\ndef dn_conv(ni, nf, size=3, stride=1):\n    for_pad = lambda s: s if s > 2 else 3\n    return nn.Sequential(\n        nn.Conv2d(ni, nf, kernel_size=size, stride=stride,\n                  padding=(for_pad(size) - 1)//2, bias=False),  # bias default is True in pytorch\n        nn.BatchNorm2d(nf),\n        nn.LeakyReLU(negative_slope=0.1, inplace=True)  # https://github.com/pjreddie/darknet/blob/master/src/activations.h\n    )\n\ndef dn_trip(ni, nf):\n    return nn.Sequential(\n        dn_conv(ni, nf),\n        dn_conv(nf, ni, size=1),  # note the flipping of (ni, nf)\n        dn_conv(ni, nf)\n    )",
      "execution_count": 21,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "The inventor of Darknet uses MaxPool, so we'll use that here, though fastai generally uses AdaptiveAvgPool instead. We will, however, select a different number of filters per layer given that our images are so small (28x28)."
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "model = nn.Sequential(\n    dn_conv(1, 8),\n    dn_maxp(),\n    dn_conv(8, 16),\n    dn_maxp(),\n    dn_trip(16, 32),\n    dn_maxp(),\n    dn_trip(32, 64),\n    dn_maxp(),\n    dn_trip(64, 128),\n    dn_maxp(),\n    dn_trip(128, 256),\n    dn_conv(256, 128, size=1),\n    dn_conv(128, 256),\n    nn.Conv2d(256, 10, 1),  # (ni, nf, size)\n    nn.AvgPool2d(10),  # (size) Keep the kernel size the same as the input to retain its dimensionality\n    Flatten()  # We don't include a softmax layer because CrossEntropyLoss includes LogSoftMax\n)",
      "execution_count": 31,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "learn = Learner(data, model, loss_func = nn.CrossEntropyLoss(), metrics=accuracy)",
      "execution_count": 32,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "print(learn.summary())",
      "execution_count": 33,
      "outputs": [
        {
          "text": "======================================================================\nLayer (type)         Output Shape         Param #    Trainable \n======================================================================\nConv2d               [1, 8, 28, 28]       72         True      \n______________________________________________________________________\nBatchNorm2d          [1, 8, 28, 28]       16         True      \n______________________________________________________________________\nLeakyReLU            [1, 8, 28, 28]       0          False     \n______________________________________________________________________\nMaxPool2d            [1, 8, 14, 14]       0          False     \n______________________________________________________________________\nConv2d               [1, 16, 14, 14]      1,152      True      \n______________________________________________________________________\nBatchNorm2d          [1, 16, 14, 14]      32         True      \n______________________________________________________________________\nLeakyReLU            [1, 16, 14, 14]      0          False     \n______________________________________________________________________\nMaxPool2d            [1, 16, 7, 7]        0          False     \n______________________________________________________________________\nConv2d               [1, 32, 7, 7]        4,608      True      \n______________________________________________________________________\nBatchNorm2d          [1, 32, 7, 7]        64         True      \n______________________________________________________________________\nLeakyReLU            [1, 32, 7, 7]        0          False     \n______________________________________________________________________\nConv2d               [1, 16, 9, 9]        512        True      \n______________________________________________________________________\nBatchNorm2d          [1, 16, 9, 9]        32         True      \n______________________________________________________________________\nLeakyReLU            [1, 16, 9, 9]        0          False     \n______________________________________________________________________\nConv2d               [1, 32, 9, 9]        4,608      True      \n______________________________________________________________________\nBatchNorm2d          [1, 32, 9, 9]        64         True      \n______________________________________________________________________\nLeakyReLU            [1, 32, 9, 9]        0          False     \n______________________________________________________________________\nMaxPool2d            [1, 32, 4, 4]        0          False     \n______________________________________________________________________\nConv2d               [1, 64, 4, 4]        18,432     True      \n______________________________________________________________________\nBatchNorm2d          [1, 64, 4, 4]        128        True      \n______________________________________________________________________\nLeakyReLU            [1, 64, 4, 4]        0          False     \n______________________________________________________________________\nConv2d               [1, 32, 6, 6]        2,048      True      \n______________________________________________________________________\nBatchNorm2d          [1, 32, 6, 6]        64         True      \n______________________________________________________________________\nLeakyReLU            [1, 32, 6, 6]        0          False     \n______________________________________________________________________\nConv2d               [1, 64, 6, 6]        18,432     True      \n______________________________________________________________________\nBatchNorm2d          [1, 64, 6, 6]        128        True      \n______________________________________________________________________\nLeakyReLU            [1, 64, 6, 6]        0          False     \n______________________________________________________________________\nMaxPool2d            [1, 64, 3, 3]        0          False     \n______________________________________________________________________\nConv2d               [1, 128, 3, 3]       73,728     True      \n______________________________________________________________________\nBatchNorm2d          [1, 128, 3, 3]       256        True      \n______________________________________________________________________\nLeakyReLU            [1, 128, 3, 3]       0          False     \n______________________________________________________________________\nConv2d               [1, 64, 5, 5]        8,192      True      \n______________________________________________________________________\nBatchNorm2d          [1, 64, 5, 5]        128        True      \n______________________________________________________________________\nLeakyReLU            [1, 64, 5, 5]        0          False     \n______________________________________________________________________\nConv2d               [1, 128, 5, 5]       73,728     True      \n______________________________________________________________________\nBatchNorm2d          [1, 128, 5, 5]       256        True      \n______________________________________________________________________\nLeakyReLU            [1, 128, 5, 5]       0          False     \n______________________________________________________________________\nMaxPool2d            [1, 128, 2, 2]       0          False     \n______________________________________________________________________\nConv2d               [1, 256, 2, 2]       294,912    True      \n______________________________________________________________________\nBatchNorm2d          [1, 256, 2, 2]       512        True      \n______________________________________________________________________\nLeakyReLU            [1, 256, 2, 2]       0          False     \n______________________________________________________________________\nConv2d               [1, 128, 4, 4]       32,768     True      \n______________________________________________________________________\nBatchNorm2d          [1, 128, 4, 4]       256        True      \n______________________________________________________________________\nLeakyReLU            [1, 128, 4, 4]       0          False     \n______________________________________________________________________\nConv2d               [1, 256, 4, 4]       294,912    True      \n______________________________________________________________________\nBatchNorm2d          [1, 256, 4, 4]       512        True      \n______________________________________________________________________\nLeakyReLU            [1, 256, 4, 4]       0          False     \n______________________________________________________________________\nConv2d               [1, 128, 6, 6]       32,768     True      \n______________________________________________________________________\nBatchNorm2d          [1, 128, 6, 6]       256        True      \n______________________________________________________________________\nLeakyReLU            [1, 128, 6, 6]       0          False     \n______________________________________________________________________\nConv2d               [1, 256, 6, 6]       294,912    True      \n______________________________________________________________________\nBatchNorm2d          [1, 256, 6, 6]       512        True      \n______________________________________________________________________\nLeakyReLU            [1, 256, 6, 6]       0          False     \n______________________________________________________________________\nConv2d               [1, 10, 6, 6]        2,570      True      \n______________________________________________________________________\nAvgPool2d            [1, 10, 1, 1]        0          False     \n______________________________________________________________________\nFlatten              [1, 10]              0          False     \n______________________________________________________________________\n\nTotal params: 1,161,570\nTotal trainable params: 1,161,570\nTotal non-trainable params: 0\n\n",
          "output_type": "stream",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "learn.lr_find(end_lr=300)\nlearn.recorder.plot()",
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<IPython.core.display.HTML object>",
            "text/html": ""
          },
          "metadata": {}
        },
        {
          "text": "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n",
          "output_type": "stream",
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 432x288 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "learn.fit_one_cycle(12, max_lr=5e-3)",
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<IPython.core.display.HTML object>",
            "text/html": "Total time: 01:54 <p><table style='width:375px; margin-bottom:10px'>\n  <tr>\n    <th>epoch</th>\n    <th>train_loss</th>\n    <th>valid_loss</th>\n    <th>accuracy</th>\n    <th>time</th>\n  </tr>\n  <tr>\n    <th>1</th>\n    <th>0.185423</th>\n    <th>0.181024</th>\n    <th>0.947200</th>\n    <th>00:10</th>\n  </tr>\n  <tr>\n    <th>2</th>\n    <th>0.143844</th>\n    <th>0.255105</th>\n    <th>0.921600</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>3</th>\n    <th>0.111953</th>\n    <th>0.088630</th>\n    <th>0.975800</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>4</th>\n    <th>0.082952</th>\n    <th>0.087798</th>\n    <th>0.976500</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>5</th>\n    <th>0.066971</th>\n    <th>0.074697</th>\n    <th>0.979400</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>6</th>\n    <th>0.056713</th>\n    <th>0.049992</th>\n    <th>0.985900</th>\n    <th>00:10</th>\n  </tr>\n  <tr>\n    <th>7</th>\n    <th>0.044795</th>\n    <th>0.037763</th>\n    <th>0.989400</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>8</th>\n    <th>0.034625</th>\n    <th>0.034965</th>\n    <th>0.990600</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>9</th>\n    <th>0.025673</th>\n    <th>0.024432</th>\n    <th>0.991800</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>10</th>\n    <th>0.019318</th>\n    <th>0.022267</th>\n    <th>0.993300</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>11</th>\n    <th>0.014915</th>\n    <th>0.016843</th>\n    <th>0.994100</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>12</th>\n    <th>0.019533</th>\n    <th>0.016252</th>\n    <th>0.994700</th>\n    <th>00:09</th>\n  </tr>\n</table>\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "terp = learn.interpret()\nterp.plot_top_losses(9, figsize=(7, 7))",
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 504x504 with 9 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "### Trying it with AdaptiveAvgPool"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "Here, we try replacing the MaxPool layers with AdaptiveAvgPool; we call it by passing the output shape we expect to see. Thus, if we're to end up with the same output shapes as we do with a stride 2 MaxPool of size 2, we will follow the same sequential reduction of the feature map: 28px -> 14px -> 7px -> 4px -> 3px -> 2px. This is a different sequence from the lesson's Resnet CNNs because of how Darknet's inner layers work, here defined using `dn_trip`."
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "model = nn.Sequential(\n    dn_conv(1, 8),\n    nn.AdaptiveAvgPool2d(14),\n    dn_conv(8, 16),\n    nn.AdaptiveAvgPool2d(7),\n    dn_trip(16, 32),\n    nn.AdaptiveAvgPool2d(4),\n    dn_trip(32, 64),\n    nn.AdaptiveAvgPool2d(3),\n    dn_trip(64, 128),\n    nn.AdaptiveAvgPool2d(2),\n    dn_trip(128, 256),\n    dn_conv(256, 128, size=1),\n    dn_conv(128, 256),\n    nn.Conv2d(256, 10, 1),  # (ni, nf, size)\n    nn.AvgPool2d(10),\n    Flatten()\n)",
      "execution_count": 49,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "learn = Learner(data, model, loss_func = nn.CrossEntropyLoss(), metrics=accuracy)",
      "execution_count": 50,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "print(learn.summary())",
      "execution_count": 51,
      "outputs": [
        {
          "text": "======================================================================\nLayer (type)         Output Shape         Param #    Trainable \n======================================================================\nConv2d               [1, 8, 28, 28]       72         True      \n______________________________________________________________________\nBatchNorm2d          [1, 8, 28, 28]       16         True      \n______________________________________________________________________\nLeakyReLU            [1, 8, 28, 28]       0          False     \n______________________________________________________________________\nAdaptiveAvgPool2d    [1, 8, 14, 14]       0          False     \n______________________________________________________________________\nConv2d               [1, 16, 14, 14]      1,152      True      \n______________________________________________________________________\nBatchNorm2d          [1, 16, 14, 14]      32         True      \n______________________________________________________________________\nLeakyReLU            [1, 16, 14, 14]      0          False     \n______________________________________________________________________\nAdaptiveAvgPool2d    [1, 16, 7, 7]        0          False     \n______________________________________________________________________\nConv2d               [1, 32, 7, 7]        4,608      True      \n______________________________________________________________________\nBatchNorm2d          [1, 32, 7, 7]        64         True      \n______________________________________________________________________\nLeakyReLU            [1, 32, 7, 7]        0          False     \n______________________________________________________________________\nConv2d               [1, 16, 9, 9]        512        True      \n______________________________________________________________________\nBatchNorm2d          [1, 16, 9, 9]        32         True      \n______________________________________________________________________\nLeakyReLU            [1, 16, 9, 9]        0          False     \n______________________________________________________________________\nConv2d               [1, 32, 9, 9]        4,608      True      \n______________________________________________________________________\nBatchNorm2d          [1, 32, 9, 9]        64         True      \n______________________________________________________________________\nLeakyReLU            [1, 32, 9, 9]        0          False     \n______________________________________________________________________\nAdaptiveAvgPool2d    [1, 32, 4, 4]        0          False     \n______________________________________________________________________\nConv2d               [1, 64, 4, 4]        18,432     True      \n______________________________________________________________________\nBatchNorm2d          [1, 64, 4, 4]        128        True      \n______________________________________________________________________\nLeakyReLU            [1, 64, 4, 4]        0          False     \n______________________________________________________________________\nConv2d               [1, 32, 6, 6]        2,048      True      \n______________________________________________________________________\nBatchNorm2d          [1, 32, 6, 6]        64         True      \n______________________________________________________________________\nLeakyReLU            [1, 32, 6, 6]        0          False     \n______________________________________________________________________\nConv2d               [1, 64, 6, 6]        18,432     True      \n______________________________________________________________________\nBatchNorm2d          [1, 64, 6, 6]        128        True      \n______________________________________________________________________\nLeakyReLU            [1, 64, 6, 6]        0          False     \n______________________________________________________________________\nAdaptiveAvgPool2d    [1, 64, 3, 3]        0          False     \n______________________________________________________________________\nConv2d               [1, 128, 3, 3]       73,728     True      \n______________________________________________________________________\nBatchNorm2d          [1, 128, 3, 3]       256        True      \n______________________________________________________________________\nLeakyReLU            [1, 128, 3, 3]       0          False     \n______________________________________________________________________\nConv2d               [1, 64, 5, 5]        8,192      True      \n______________________________________________________________________\nBatchNorm2d          [1, 64, 5, 5]        128        True      \n______________________________________________________________________\nLeakyReLU            [1, 64, 5, 5]        0          False     \n______________________________________________________________________\nConv2d               [1, 128, 5, 5]       73,728     True      \n______________________________________________________________________\nBatchNorm2d          [1, 128, 5, 5]       256        True      \n______________________________________________________________________\nLeakyReLU            [1, 128, 5, 5]       0          False     \n______________________________________________________________________\nAdaptiveAvgPool2d    [1, 128, 2, 2]       0          False     \n______________________________________________________________________\nConv2d               [1, 256, 2, 2]       294,912    True      \n______________________________________________________________________\nBatchNorm2d          [1, 256, 2, 2]       512        True      \n______________________________________________________________________\nLeakyReLU            [1, 256, 2, 2]       0          False     \n______________________________________________________________________\nConv2d               [1, 128, 4, 4]       32,768     True      \n______________________________________________________________________\nBatchNorm2d          [1, 128, 4, 4]       256        True      \n______________________________________________________________________\nLeakyReLU            [1, 128, 4, 4]       0          False     \n______________________________________________________________________\nConv2d               [1, 256, 4, 4]       294,912    True      \n______________________________________________________________________\nBatchNorm2d          [1, 256, 4, 4]       512        True      \n______________________________________________________________________\nLeakyReLU            [1, 256, 4, 4]       0          False     \n______________________________________________________________________\nConv2d               [1, 128, 6, 6]       32,768     True      \n______________________________________________________________________\nBatchNorm2d          [1, 128, 6, 6]       256        True      \n______________________________________________________________________\nLeakyReLU            [1, 128, 6, 6]       0          False     \n______________________________________________________________________\nConv2d               [1, 256, 6, 6]       294,912    True      \n______________________________________________________________________\nBatchNorm2d          [1, 256, 6, 6]       512        True      \n______________________________________________________________________\nLeakyReLU            [1, 256, 6, 6]       0          False     \n______________________________________________________________________\nConv2d               [1, 10, 6, 6]        2,570      True      \n______________________________________________________________________\nAvgPool2d            [1, 10, 1, 1]        0          False     \n______________________________________________________________________\nFlatten              [1, 10]              0          False     \n______________________________________________________________________\n\nTotal params: 1,161,570\nTotal trainable params: 1,161,570\nTotal non-trainable params: 0\n\n",
          "output_type": "stream",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "learn.lr_find(end_lr=300)\nlearn.recorder.plot()",
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<IPython.core.display.HTML object>",
            "text/html": ""
          },
          "metadata": {}
        },
        {
          "text": "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n",
          "output_type": "stream",
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 432x288 with 1 Axes>",
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAEKCAYAAAAB0GKPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3Xd8VfX9+PHXO3svEiBkEPZSWRGxCKJ1oHW2Lhxf29pSWq3aau22w9r216+1Vlv1Sx1djjqwxY1tXYgiYQaIQMIMBBIIZO+8f3/cg15Dxg3ce09y834+HueRc8/5nHven9wk73zO53M+R1QVY4wxpidhbgdgjDGmf7CEYYwxxieWMIwxxvjEEoYxxhifWMIwxhjjE0sYxhhjfGIJwxhjjE8sYRhjjPGJJQxjjDE+iXA7AH9KT0/XvLw8t8Mwxph+Y9WqVQdUNcOXsiGVMPLy8igoKHA7DGOM6TdEZKevZe2SlDHGGJ9YwjDGGOMTSxjGGGN8YgnDGGOMTyxhGGOM8YklDGOMMT6xhGGMMcYnIXUfxrF6e0sF0RFhJERHkBgTQWJMJDGRYVTWNbO/upH91U3sq2qktqmV1Pgo0uOjGJQQTVp8FJnJMcRH27fRGBP67C8d8LW/FdDY0n7Mxw9NimHU4HhGZSQwKiOB1PgooiPCiIkMJzoijLiocMYOSSQmMtyPURtjTHBZwgCe/dpnqGlqoaaxldrGVmqbWqlrbiUtLoohyTEMSYxhaHIMCdERHG5o5mCts9Q1UXqogZKKWkoq6li8eg+1Ta2dniM6Ioz8vFRmjU5n1qh0TshKJjxMglxTY4w5dpYwgBOzk30uOzgxhsGJMZ3uU1UqapqobmyhsaWdptZ2mlrbqG5o4cPth1hecoDfvLYZ2ExMZBjpzmWt1LgoBsVHER8dQZiAiCACYSJkp8YyJSeFicOSiI6wFooxxj2WMPxIRBicFMPgpKMTyrwTMgGoqGlieckBCkurqKxrprK+mcq6ZkoqaqlrakUBVWhXpb1dqWtuAyAyXJiYmcTknBRGZSSQmxZH7qA4slNjLZEYY4JCVNXtGPwmPz9fQ23ywX1VjazdfYg1uw+zdtdhCvdUUe8kEQARSE+IJjEmgoToCOKjIkhw1uOiwok/8jUqgpS4SDISoz9eBsVH22UxYwY4EVmlqvm+lLUWRh83NDmGecmZH7dQVJWK2iZ2V9azq7KeXQcbKKtqoKaplbomTx/M7sp6aptaaWhuo665tcsO/fAwYeyQRKYPT2FabirTclMZPiiO2qZWyqoaKatqZF9VAw3NbaTERZEcF0lqXBSpcZEMSYqxTnxjBhhLGP2MiHzcjzJ9eJpPx7S1K3XNrRyua6GitpGKmiYqapooq2pkfWkV/1yzl79/sAuAqPAwmtt6HjEmAsOSYxmRHs+I9HiGD4ojJS6K2Mhw4qLCiY0KJzEmgpHpCcRGWWIxJhRYwhgAwsOEpJhIkmIiyR0Ud9T+tnZla3kNq3ceZvuBWjISoxmaHMuwZM/osLioCA7XN3OovuXjr3sONbD9QC3bD9bzr7V7qG7sfHSYCOQNimfckETGDU1kQmYSJ2Unk5kcg4hdDjOmP7GEYQgPE8YPTWL80KQuy6TFR3W5T1U5XO8ZltzQ0kZ9s+dyWGV9M1v217J5XzUf7avm9U37ONJlNig+ihOzkzkxK5kpOSlMH55KSlzX5zDGuC9gCUNEcoC/AkOBdmCRqv6+Q5mLgbuc/a3Araq6zNl3PfAjp+gvVPUvgYrVHB8RITU+itRukgpAfXMrH+2rYcOeKgpLqyjcU8W7Ww/Q1u7JIqMHJ5A/PJX8vDTOnjiE5NjIYIRvjPFRwEZJiUgmkKmqq0UkEVgFXKKqm7zKJAB1qqoichLwjKqOF5E0oADIB9Q5drqqHurunKE4SirUNTS3sb70MAU7D1Gwo5JVOw9R3dhKXFQ4V+TncMNpI8hJO/oymjHGP/rEKClVLQPKnPUaESkCsoBNXmVqvQ6Jx5McAM4F3lDVSgAReQOYBzwVqHiNO2Kjwjll5CBOGTkIgPZ2pXBPFX9ZvoO/f7CTv76/g3MnDeWG00YwfXiq9XsY46Kg9GGISB4wFVjRyb5LgV8Bg4HPOZuzgN1exUqdbSbEhYUJk3NSuPfKKdwxbzx/eX8HT3ywk1c37CMnLZbzT8jk/BMzOSk72ZKHMUEW8Bv3nMtObwN3q+ribsrNAe5U1bNE5DtAtKr+wtn3Y6BeVX/byXELgAUAubm503fu3BmIahgX1TW18nJhGa8UlrFs6wFa25Xs1Fg+PzWLr84ZSWKM9XUYc6x6c0kqoAlDRCKBl4DXVfVeH8pvB04GzgbmqurXnO3/B7ylqt1ekrI+jNBXVd/C0k37eLmwjLc2V5CRGM335o3n0qlZhNld68b0Wp9IGOK5XvAXoFJVb+2izGigxOn0nga8CGQDqXg6uqc5RVfj6fSu7O6cljAGlrW7D/PTJRtZu/swU3NT+OmFk5ick+J2WMb0K71JGIF84t4s4DrgTBFZ6yzni8hCEVnolPkCsEFE1gJ/BK5Uj0o8w21XOsvPe0oWZuCZkpPC4q9/hnsun8zuygYuefA9fvTPQuqbO7+J0BhzfGzyQRMSahpbuO/fW3nsve2MGBTPfVdN4aRsa20Y05O+0sIwJmgSYyL58QUTeeIrp9DQ0sbnH1zOH98s/vimQGPM8bOEYULKZ0al89otc5h3wlD+9/XNXLXofXZX1rsdljEhwRKGCTnJcZE8MH8qv7tyMkVlNZx//7u8tH6v22EZ0+9ZwjAhSUS4dGo2r9w8m1EZCdz05BrueG6ddYgbcxwsYZiQljsojmcXnsqNZ4zi2VWlXHD/MjbsqXI7LGP6JUsYJuRFhofxnXPH88RXTqGuuZVLH3yPJ1fscjssY/odSxhmwDjSIX7qqHR+8EIh319cSFNrW88HGmMASxhmgEmNj+LxL57MN+aO4qkPdzF/0Qfsr250Oyxj+gVLGGbACQ8T7pg3ngevmcZH+2q44IFlrNppEwkY0xNLGGbAOv/ETF74xiziosK59pEP2bq/xu2QjOnTLGGYAW3c0ESe/dqpxEWFc9OTa2hotj4NY7piCcMMeIOTYvjdlVPYvL+Gn7240e1wjOmzLGEYA8wZm8E35o7i6ZW7+dfaPW6HY0yfZAnDGMe3zx5L/vBUfrC4kG0VtT0fYMwAYwnDGEdEeBj3z59KZEQYNz65hsYW688wxpslDGO8DEuJ5d4rJlNUVs3dLxe5HY4xfYolDGM6OHP8EL46ewR/+2Anr20oczscY/oMSxjGdOI7547npOxk7nhuPXsON7gdjjF9giUMYzoRFRHG/VdNpV3hlqfW0NrW7nZIxrjOEoYxXchLj+fuS0+gYOchfv+frW6HY4zrApYwRCRHRN4UkSIR2Sgit3RS5hoRWe8sy0Vkste+HSJSKCJrRaQgUHEa052Lp2Rx2fRs/vBmMctLDrgdjjGuCmQLoxW4TVUnADOBG0VkYocy24HTVfUk4C5gUYf9Z6jqFFXND2CcxnTrZxdNYkR6PLc+vZaDtU1uh2OMawKWMFS1TFVXO+s1QBGQ1aHMclU95Lz8AMgOVDzGHKv46AgemD+Vw/UtfH9xIarqdkjGuCIofRgikgdMBVZ0U+wG4FWv1wosFZFVIrKgm/deICIFIlJQUVHhj3CNOcqkYcncds5Ylm7azwtrbOoQMzAFPGGISALwPHCrqlZ3UeYMPAnju16bZ6nqNOA8PJez5nR2rKouUtV8Vc3PyMjwc/TGfOIrs0dycl4qP1mykb021NYMQAFNGCISiSdZPKGqi7socxLwCHCxqh48sl1V9zpfy4EXgBmBjNWYnoSHCfdcPpm2duW7z6+3S1NmwAnkKCkBHgWKVPXeLsrkAouB61R1i9f2eBFJPLIOnANsCFSsxvhq+KB4fnD+BN7deoC/f7DT7XCMCaqIAL73LOA6oFBE1jrbfgDkAqjqw8CdwCDgQU9+odUZETUEeMHZFgE8qaqvBTBWY3x2zSm5vL5xH7985SNmj8kgLz3e7ZCMCQoJpWZ1fn6+FhTYLRsm8MqqGjjnd+8wdkgiz3ztVMLDxO2QjDkmIrLK11sX7E5vY45BZnIsP794Eqt2HuLx97a7HY4xQWEJw5hjdMmULM4cP5jfLt3C7sp6t8MxJuAsYRhzjESEX1xyAuFhYjf0mQHBEoYxx2FYSizfPW88y4oP8NyqUrfDMSagLGEYc5yumZHLjLw0fvFyERU1NteUCV2WMIw5TmFhwq++cCINLW38dMlGt8MxJmAsYRjjB6MyErjls2N4ubCMpRv3uR2OMQFhCcMYP1kwZyQTMpO4818baWxpczscY/zOEoYxfhIZHsadF0xkX3UjzxTsdjscY/zOEoYxfjRzZBon56Xy8FslNLfac8BNaLGEYYwfiQg3nTmGvVWNLF5tw2xNaLGEYYyfzRmTzuTsZB58q4TWNmtlmNBhCcMYPzvSythVWc+L6/e6HY4xfmMJw5gA+Oz4wYwfmsgf/ltMe7tNGWJCgyUMYwIgLEy46czRlFTU8Zrdl2FChCUMYwLkvBMyGZkRzwP/LbaJCU1IsIRhTICEhwk3nTGaorJq/lNU7nY4xhw3SxjGBNBFk4eRmxbHA29aK8P0f5YwjAmgiPAwFp4+inW7D/Ne8UG3wzHmuAQsYYhIjoi8KSJFIrJRRG7ppMw1IrLeWZaLyGSvffNEZLOIFIvI9wIVpzGB9oXpWQxNiuGB/251OxRjjksgWxitwG2qOgGYCdwoIhM7lNkOnK6qJwF3AYsARCQc+CNwHjARmN/Jscb0C9ER4SyYM5IV2ysp2FHpdjjGHLOAJQxVLVPV1c56DVAEZHUos1xVDzkvPwCynfUZQLGqblPVZuBp4OJAxWpMoM2fkcug+Cj+8Gax26EYc8yC0ochInnAVGBFN8VuAF511rMA7+k+S+mQbIzpT2Kjwrlh9gje2lxBYWmV2+EYc0wCnjBEJAF4HrhVVau7KHMGnoTx3SObOinW6RATEVkgIgUiUlBRUeGPkI0JiOtmDicpJoI/WivD9FMBTRgiEoknWTyhqou7KHMS8AhwsaoeGUZSCuR4FcsGOp2UR1UXqWq+quZnZGT4L3hj/CwxJpIvfiaP1zbuY8v+GrfDMabXAjlKSoBHgSJVvbeLMrnAYuA6Vd3itWslMEZERohIFHAVsCRQsRoTLF+aNYK4qHAetFaG6YcC2cKYBVwHnCkia53lfBFZKCILnTJ3AoOAB539BQCq2grcBLyOp7P8GVXdGMBYjQmK1Pgorp05nCXr9rLzYJ3b4RjTKxJKd5/m5+drQUGB22EY063ymkZm/783ueCkYfz2isk9H2BMAInIKlXN96Ws3eltTJANTozh+s/k8cKaUrZaX4bpRyxhGOOCr58+ivioCO5ZutntUIzxmSUMY1yQGh/FV+eM5PWN+1mz61DPBxjTB1jCMMYlN5w2gkHxUfzmtc02k63pFyxhGOOS+OgIbjpzNO9vO8iy4gNuh2NMjyxhGOOiq0/JJSsl1loZpl+whGGMi6IjwvnW2WMp3FPFqxvs2d+mb7OEYYzLLp2axZjBCdyzdDOtbe1uh2NMlyxhGOOy8DDhtnPGsa2ijpcLy9wOx5guWcIwpg84Z+IQRmXE86d3t1lfhumzLGEY0weEhQlfmT2SDXuqeX+bPfvb9E2WMIzpIy6dmkV6QhSPvLvd7VCM6ZQlDGP6iJjIcK6bmcd/Pyq3OaZMn2QJw5g+5LpThxMdEWatDNMnWcIwpg9Ji4/isunZvLBmD+U1jW6HY8yn+JQwRGSUiEQ763NF5GYRSQlsaMYMTDecNoKW9nb+9v5Ot0Mx5lN8bWE8D7SJyGg8j10dATwZsKiMGcBGZiRw1oQh/O2DnTQ0t7kdjjEf8zVhtDuPTb0UuE9VvwVkBi4sYwa2BXNGcri+hedW7XY7FGM+5mvCaBGR+cD1wEvOtsjAhGSMyR+eyuScFB5dtp32druRz/QNviaMLwGnAner6nYRGQH8vbsDRCRHRN4UkSIR2Sgit3RSZryIvC8iTSJye4d9O0SkUETWiog9qNsMKCLCl2flseNgvU19bvqMCF8Kqeom4GYAEUkFElX11z0c1grcpqqrRSQRWCUibzjvdUSl876XdPEeZ6iq/baYAWneCUNJi4/iiRU7mTM2w+1wjPF5lNRbIpIkImnAOuBxEbm3u2NUtUxVVzvrNUARkNWhTLmqrgRajil6Y0JYdEQ4l0/P5t9F5eyvtiG2xn2+XpJKVtVq4PPA46o6HTjL15OISB4wFVjRi9gUWCoiq0RkQS+OMyZkzJ+RS1u78sxK6/w27vM1YUSISCZwBZ90evtERBLwDMu91Uk6vpqlqtOA84AbRWROF++/QEQKRKSgoqKiN6EZ0+flpccza/Qgnl65mzbr/DYu8zVh/Bx4HShR1ZUiMhLY2tNBIhKJJ1k8oaqLexOYqu51vpYDLwAzuii3SFXzVTU/I8Ou85rQc/WM4ew53MA7W+wfIuMunxKGqj6rqiep6ted19tU9QvdHSMigucmvyJV7ba/o5Nj452OckQkHjgH2NCb9zAmVJw9cQjpCZ7Ob2Pc5NMoKRHJBh4AZuHpW1gG3KKqpd0cNgu4DigUkbXOth8AuQCq+rCIDAUKgCSgXURuBSYC6cALnpxDBPCkqr7Wy7oZExKiIsK4Ij+Hh98uYe/hBoalxLodkhmgfEoYwON4pgK53Hl9rbPt7K4OUNVlgHT3pqq6D8juZFc1MNnH2IwJefNn5PLQ2yX8Y+VuvnX2WLfDMQOUr30YGar6uKq2OsufAeswMCZIctLimD0mg3+s3E1rW7vb4ZgByteEcUBErhWRcGe5FrDnSBoTRFfPyGVfdSNvbrbOb+MOXxPGl/EMqd0HlAGX4ZkuxBgTJJ+dMJjBidH87QPr/Dbu8HWU1C5VvUhVM1R1sKpegucmPmNMkESGh/E/pw7nnS0VFJX15pYmY/zjeJ64922/RWGM8cm1M4cTFxXOn97Z5nYoZgA6noTR7QgoY4z/pcRFcdXJuSxZt5c9hxvcDscMMMeTMGyeAmNccMPsEQA8+u52lyMxA023CUNEakSkupOlBhgWpBiNMV6yUmK5aPIwnl65i8P1zW6HYwaQbhOGqiaqalInS6Kq+nrTnzHGzxacPpL65jb+biOmTBAdzyUpY4xLxg9NYu64DP68fAeNLW1uh2MGCEsYxvRTX5szigO1zTy/ursp3YzxH0sYxvRTM0emMTknhT+9s82elWGCwhKGMf2UiLBwzkh2HKznpfV73Q7HDACWMIzpx86ZNJRJw5K4818b2V1Z73Y4JsRZwjCmHwsPEx68ZhrtqnzjidXWAT4A1Ta1Bm0GY0sYxvRzwwfF89vLJ1O4p4qfv7TJ7XBMkN3z+mZOvvvfqAa+H8sShjEh4JxJQ1l4+iieXLGL51bZqKmBpLi8lpy0OJwnlAaUJQxjQsTt54xl5sg0fvhCoc1mO4CUVNQyOiMhKOeyhGFMiIgID+OB+dNIjo3k639fRW1Tq9shmQCrbWqlrKqRUYMtYRhjeikjMZrfXzWVHQfrea5gt9vhmADbVlELwKiM+KCcL2AJQ0RyRORNESkSkY0icksnZcaLyPsi0iQit3fYN09ENotIsYh8L1BxGhNqTh01iBOyknimwPoyQl1xuSdhjA6BFkYrcJuqTgBmAjeKyMQOZSqBm4F7vDeKSDjwR+A8YCIwv5NjjTFduDI/h01l1WzYU+V2KCaASipqCQ8TctP6eQtDVctUdbWzXgMUAVkdypSr6kqgpcPhM4BiVd2mqs3A08DFgYrVmFBz0eQsoiLCeMYuS4W04vJahg+KIyoiOL0LQTmLiOQBU4EVPh6SBXj/pJfSIdkYY7qWHBfJvElD+eeaPXYzXwgrqahjVJBGSEEQEoaIJADPA7eqqq9j/TobUNzpXSkiskBECkSkoKKi4ljDNCbkXHlyDtWNrby+cZ/boZgAaGlrZ+fBuqD1X0CAE4aIROJJFk+o6uJeHFoK5Hi9zgY6nV1NVRepar6q5mdkZBx7sMaEmFNHDiI7NZZnrfM7JO2qrKelTUOjhSGe2w4fBYpU9d5eHr4SGCMiI0QkCrgKWOLvGI0JZWFhwuXTc1hWfMAmJgxBJUEeIQWBbWHMAq4DzhSRtc5yvogsFJGFACIyVERKgW8DPxKRUhFJUtVW4CbgdTyd5c+o6sYAxmpMSLosPxsRbLqQEFTs3IMxMkj3YAAE7LncqrqMzvsivMvsw3O5qbN9rwCvBCA0YwaMrJRYThudznOrSrn5s2MIDwv8fEMmOErK6xiSFE1STGTQzml3ehsT4q7Iz2HP4QaWlxxwOxTjR8UVtUHtvwBLGMaEvHMmDSElLpJ/rLR7MkKFqrKtvDao/RdgCcOYkBcdEc4lU7JYunE/h+ub3Q7H+EF5TRM1Ta3WwjDG+N9l07NpbmvnxfVlbodi/MCNEVJgCcOYAWHSsCTGD0200VIhouTjWWotYRhj/ExEuGx6Nut2H2br/hq3wzHHqbi8loToCIYkRQf1vJYwjBkgLpmaRUSY8Nxqa2X0d545pOKD8lhWb5YwjBkg0hOimTtuMC+s3kNrW7vb4ZjjUFxeG7Sn7HmzhGHMAHLZ9GzKa5p4t9juyeivapta2VfdGPT+C7CEYcyAcub4waTGRVrndz/m1ggpsIRhzIASFRHGxVOyeGPjfqrqOz63zPQHbo2QAksYxgw4R+7JWLK+0ycGmD6uuLyWiDBh+KC4oJ/bEoYxA4zdk9G/lVR4HssaGR78P9+WMIwZYOyejP6tpCK4T9nzZgnDmAHI7snon1ra2tlxILjP8fZmCcOYASg9IZo5YzN4eX0Zqup2OMZHuyrraW1Xa2EYY4LrzPGDKT3UQElFnduhGB9t3FsNuDNCCixhGDNgzR2XAcBbm8tdjsT4QlV55N1tZKfGMnFYkisxWMIwZoDKTo1jzOAE3tpc4XYoxgf/LipnfWkVN392jCsjpCCACUNEckTkTREpEpGNInJLJ2VERO4XkWIRWS8i07z2tYnIWmdZEqg4jRnI5o7L4MPtldQ1tbodiulGe7ty7xtbGD4ojs9PzXItjkCmqVbgNlWdAMwEbhSRiR3KnAeMcZYFwENe+xpUdYqzXBTAOI0ZsOaOG0xzWzvvlxx0OxTTjaWb9lFUVs0tnx1DhEutCwhgwlDVMlVd7azXAEVAx9R4MfBX9fgASBGRzEDFZIz5tPy8VOKiwnlri/Vj9FXt7crv3tjKyIx4Lp7iXusCgtSHISJ5wFRgRYddWYD3k+lL+SSpxIhIgYh8ICKXBDxIYwag6IhwZo1O582PKmx4bR/1cmEZm/fXcOtZYwkPC+7zLzoKeMIQkQTgeeBWVa3uuLuTQ4781Oaqaj5wNXCfiIzq4v0XOImloKLCOu+M6a254zLYc7jh40ntTN/R1q7c9+8tjB2SwAUnun/xJaAJQ0Qi8SSLJ1R1cSdFSoEcr9fZwF4AVT3ydRvwFp4WylFUdZGq5qtqfkZGhh+jN2ZgmDtuMICNluqDlqzbQ0lFHd86ayxhLrcuILCjpAR4FChS1Xu7KLYE+B9ntNRMoEpVy0QkVUSinfdJB2YBmwIVqzEDWVZKLGMGJ/Cm3Y/RpzS3tvP7f29lQmYS504a6nY4AEQE8L1nAdcBhSKy1tn2AyAXQFUfBl4BzgeKgXrgS065CcD/iUg7nqT2a1W1hGFMgJwxfjB/fm8HdU2txEcH8s+C8dU9Szez42A9j3/x5D7RuoAAJgxVXUbnfRTeZRS4sZPty4ETAxSaMaaDuWMzWPTONpaXHOTsiUP88p7bD9SRnhBFYkykX95vIHlrczmL3tnGtTNzOWP8YLfD+Zjd6W2MIT8vjfiocL9NE/Lkil2cde/bnHXv2zb1SC+VVzdy2zPrGD80kR99ruOta+6yhGGMISoijFmj03lr8/ENr21vV3796kf84IVCTh05iKSYSL74+Eq+9/x6ahrtkbA9aW9XvvXMWuqaW/nD1VOJiQx3O6RPsYRhjAE8o6X2HG6guPzYhtc2trTxzafX8PDbJVx9Si5//tLJvPjN01h4+iieKdjNvPve5b3iA36OOrQ89HYJ7xUf5GcXTWL04ES3wzmK9W4ZYwDv2WsrGDOkd3+sDtY2seBvq1i18xDfP288C+aMRESICIfvnTeecyYN4fZn1nHNIyvIGxTHyXlpzBiRxikjBpGTFotnUOXAtmpnJfe+sYULTsrkivycng9wgSUMYwwAw5zhte9sreCrc0b26tg7l2ykcE8Vf7x6Gp876egbzKblpvLKLbN56sNdLC85yBtF+3nWeaZ4SlwkybGRxEVFkBAdTnx0BCPS4/nCtGwmDUsaEMmksq6Zm59ay7CUGH75+RP7bJ0tYRhjPjZ7TAZPrNhJY0tbr66fb9hTxdkTh3SaLI6IiQznS7NG8KVZI2hvV7aW1/Lh9oMU7auhrqnVWdqorGtmeclBHn9vB+OHJnLZ9GwumZpFekK0P6rY57S1Kzc/tYaK2iaeW3gqSX14VJklDGPMx2aPTeex97azckcls8f4NnNCY0sbuyvrezUxXliYMG5oIuOGdn7pq6q+hSXr9/LcqlJ+8XIRv371I747b3yvWz79wT1LN7Os+AC/+cJJnJSd4nY43bJOb2PMx04ZkUZUeBjvbvW9c3rnwXraFUZlxPstjuS4SK6bOZx/3TiLN741h7njMvjlq0Uh12n+2oZ9PPRWCfNn5HLFyX2z38KbJQxjzMfioiKYPjyVd7b4Pq/UkUkLA/Wc6TFDErl//lRGZyRwy9Nr2F/dGJDzBFtxeS23P7uOyTkp/PSivnW/RVcsYRhjPmX22HQ+2ldDeY1vf5hLnGG4I/3YwugoLiqCB6+ZRl1TG998cg2tbe0BO1cw1Da1svDvq4iOCOOha6YRHdG37rfoiiUMY8ynzHH6Lny9/FNSUUtWSixxUYHtEh0zJJFffv4EPtxRyT1LtwT0XIGkqnz3+fVsq6jlgaunMiwl1u2QfGYJwxjzKRMzk0iLj+LdLb4mjLqAti68XTo1m/kMCGjfAAAQPUlEQVQzcnn47RL+U7Q/KOf0t78s38HL68u4Y954PjMq3e1wesUShjHmU8LChFmj03m3+ECP04SoKiUVtQHrv+jMTy6cyMTMJL79zDqW97NO8NW7DnH3K0WcNWEIC2b3vxFfljCMMUeZPSadipomNu+v6bbcvupG6pvbGDU4eAkjJjKcB6+ZRnxUOFc/soKrFr3Ph9srg3b+Y3WorpmbnljNkKQYfnv55D4zZXlvWMIwxhxl9hjPpZKeLkuVlNcB/h1S64u89Hj+e/tcfnLhRIrL67ji/97nukdXsGbXoaDG4av2duXWf6zlQG0zD10zneS4vntzXncsYRhjjpKZ/Mk0Id05MqR2dBAvSR1x5M7xd+84gx+eP4FNe6v5/EPLeWzZ9qDH0pM/vlnM21squPPCiZyYnex2OMfMEoYxplOnjUnnw+2VNLa0dVlmW0UtidERZCS6N21HbFQ4X50zknfuOIOzJwzh5y9t4u6XN9HefuzTtPvTh9srufffW7h4yjCuOSXX7XCOiyUMY0yn5ozJoKm1nZU7uu4fKKmoY+TghD4xWV58dAQPXTud/zl1OH96dzu3/GMtTa1dJ7tgaGlr54cvFJKVEssvL+27kwr6yhKGMaZTp4xMIzJcWNbNNCGeEVLB7b/oTniY8LOLJvHdeeN5cd1ern/sQ6oa3Htw02PLtrO1vJafXjgpJJ6VbgnDGNOpuKgI8oen8U4XCaO2qZWyqsagDqn1hYjw9bmjuO/KKazaeYjrHl1Bc2vw7wwvq2rg9//ZylkTBnOWn56T7raAJQwRyRGRN0WkSEQ2isgtnZQREblfRIpFZL2ITPPad72IbHWW6wMVpzGma7PHplNUVt3pNCHbK46MkOpbCeOIS6Zm8cD8qawvreK3b2wO+vnvemkTbe3KTy6cFPRzB0ogWxitwG2qOgGYCdwoIh1n2DoPGOMsC4CHAEQkDfgJcAowA/iJiKQGMFZjTCeOTBPy36Lyo/Z9Mulg37kk1dG8EzKZPyOXRe9s4/2Sg0E77ztbKnilcB/fPHM0OWlxQTtvoAUsYahqmaqudtZrgCKg44T5FwN/VY8PgBQRyQTOBd5Q1UpVPQS8AcwLVKzGmM5NGpZE3qA4Xly/96h9JRW1hIcJuYP69h/EH18wgbxB8dz2zFqq6gPfn9HU2sZPlmxkZHp8yD2/Iyh9GCKSB0wFVnTYlQXs9npd6mzrarsxJohEhAsnD+P9koNHXZYqqaglNy2uz8+0GhcVwX1XTqG8pokf/WtDj9OdHK9Fb29j+4E6fnbxpD7/vemtgCcMEUkAngduVdXqjrs7OUS72d7Z+y8QkQIRKaio8H0Of2OMby6aPIx2hVfWl31qe0l5XZ++HOVtck4Kt541hhfX7eVfa49uLfnLvqpG/vBmMZ87MdPnJxb2JwFNGCISiSdZPKGqizspUgp4P2YqG9jbzfajqOoiVc1X1fyMjND7gIxx25ghiYwfmsiLXgmjrV3ZfqCuz3Z4d+brc0eTPzyVH/9zA7sr6wNyjj8v30FLWzvfnTc+IO/vtkCOkhLgUaBIVe/totgS4H+c0VIzgSpVLQNeB84RkVSns/scZ5sxxgUXTh7Gqp2HKD3k+UNbeqie5rb2fpUwwsOE3105BQW+v7jQ75em6ppaeXLFTuadMLTP9+scq0C2MGYB1wFnishaZzlfRBaKyEKnzCvANqAY+BPwDQBVrQTuAlY6y8+dbcYYF1x40jAAXnJaGR+PkBrcPy5JHZGTFsd3zh3HsuIDvFxY1vMBvfBMwW6qG1v5Sj+cttxXAbv1UFWX0XlfhHcZBW7sYt9jwGMBCM0Y00u5g+KYkpPCkrV7WXj6KLY592CMTO8/LYwjrp05nGcKdnPXS5uYO24wCX64A7utXXnsve1MH57KtNzQvQPA7vQ2xvjkwsnD2FRWTXF5LSUVtQyKjyI1PsrtsHotPEz4xSUnUF7TxH1v+OdRr69v3Mfuyga+OnuEX96vr7KEYYzxyQUnZSICL63f64yQ6n+tiyOm5qYyf0Yujy/fQVFZx8Gbvfend7cxfFAcZ08c6ofo+i5LGMYYnwxJiuGUEWksWbfXM+lgP+u/6OiOc8eRHBvJj/+54bimQl+1s5I1uw5zw2kjCO+HT9HrDUsYxhifXTh5GNsq6jhY19yvWxgAKXFRfO+88RTsPMTzq0uP+X3+9M52kmMjuWx6th+j65ssYRhjfHbeCZlEOP9F9/eEAXDZtGzyh6fyq1c/oqKmqdfH7zxYx+ub9nHtzFziovr/9OU9sYRhjPFZWnwUpznP+w6FhBEWJtx1yQnUNrVywQPvsry4+2eYe1NVHn57G5FhYVx/al7gguxDLGEYY3pl4emj+NyJmWSlxrodil9MyExi8dc/Q0J0BNc8uoJfv/oRLW3dPz+jYEcllz/8Pk99uIvL8rMZnBQTpGjdJYGeiCuY8vPztaCgwO0wjDH9UH1zK3e9VMRTH+5icnYyv79qKnnpn+7Y37q/ht+8vpk3Nu0nIzGab501lsvzs4kM77//e4vIKlXN96msJQxjjPnEq4VlfG9xIVUNLUSGCzGR4cRGhhMbFc7uynrioiJYePpIvnzaiJDot+hNwuj/tTXGGD8678RMpuSmsHj1HuqaWmloaaOxpY2G5jbOPzGTr84eSVo/vGHRHyxhGGNMB5nJsdx4xmi3w+hz+u+FN2OMMUFlCcMYY4xPLGEYY4zxiSUMY4wxPrGEYYwxxieWMIwxxvjEEoYxxhifWMIwxhjjk5CaGkREKoCdQDJQ5bWru9dH1o98TQd8n7LyaB3P1Zv9vYm74+uO9YDjq0tP9eiuTGfbfYndez1U6uG9bvUIbj28X/elenQXpy+v/f07MkZVk30qqaohtwCLfH19ZN3ra4E/z92b/b2Ju6d6HG9deqpHd2U62+5L7KFYD3/9bFk9el+Pbn6eXK1Hb+Lu4XMI2u/IkSVUL0m92IvXL3ZRxl/n7s3+3sTd8XWw69Fdmc62+xK793qo1MPXGHpi9eh9Pbxf96V6dLavr/+uAyF2ScofRKRAfZy5sa8LlbpYPfoWq0ffE6y6hGoL43gscjsAPwqVulg9+harR98TlLpYC8MYY4xPrIVhjDHGJyGdMETkMREpF5ENx3DsdBEpFJFiEblfRMRr3zdFZLOIbBSR3/g36k5j8Xs9ROSnIrJHRNY6y/n+j7zTeALymTj7bxcRFZF0/0XcZSyB+EzuEpH1zuexVESG+T/yo2IJRD3+V0Q+curygoik+D/yo2IJRD0ud37H20UkoP0DxxN/F+93vYhsdZbrvbZ3+zvUo2MditUfFmAOMA3YcAzHfgicCgjwKnCes/0M4N9AtPN6cD+tx0+B20PhM3H25QCv47kPJ70/1gNI8ipzM/BwP63HOUCEs/7/gP/XT+sxARgHvAXk98X4ndjyOmxLA7Y5X1Od9dTu6urrEtItDFV9B6j03iYio0TkNRFZJSLvisj4jseJSCaeX9731fNd/itwibP768CvVbXJOUd5YGsRsHq4IoB1+R1wBxCUTrlA1ENVq72KxhOEugSoHktVtdUp+gGQHdhaBKweRaq6OdCxH0/8XTgXeENVK1X1EPAGMM8ffw9COmF0YRHwTVWdDtwOPNhJmSyg1Ot1qbMNYCwwW0RWiMjbInJyQKPt2vHWA+Am57LBYyKSGrhQe3RcdRGRi4A9qrou0IH24Lg/ExG5W0R2A9cAdwYw1u7442friC/j+U/WDf6shxt8ib8zWcBur9dH6nTcdR1Qz/QWkQTgM8CzXpfuojsr2sm2I//tReBp5s0ETgaeEZGRTsYOCj/V4yHgLuf1XcBv8fxyB9Xx1kVE4oAf4rkM4ho/fSao6g+BH4rI94GbgJ/4OdRu+aseznv9EGgFnvBnjL7wZz3c0F38IvIl4BZn22jgFRFpBrar6qV0XafjruuAShh4WlSHVXWK90YRCQdWOS+X4Plj6t2Mzgb2OuulwGInQXwoIu145nGpCGTgHRx3PVR1v9dxfwJeCmTA3TjeuowCRgDrnF+sbGC1iMxQ1X0Bjt2bP362vD0JvEyQEwZ+qofT0XoB8Nlg/jPlxd+fR7B1Gj+Aqj4OPA4gIm8BX1TVHV5FSoG5Xq+z8fR1lHK8dQ1kR05fWIA8vDqSgOXA5c66AJO7OG4lnlbEkc6h853tC4GfO+tj8TT9pB/WI9OrzLeAp/vrZ9KhzA6C0OkdoM9kjFeZbwLP9dN6zAM2ARnB+pkK5M8VQej0Ptb46brTezueKyGpznqaL3XtMcZgfqDBXoCngDKgBU92vQHPf6OvAeucH+o7uzg2H9gAlAB/4JObHKOAvzv7VgNn9tN6/A0oBNbj+U8rM9D1CFRdOpTZQXBGSQXiM3ne2b4ez/w+Wf20HsV4/pFa6yzBGO0ViHpc6rxXE7AfeL2vxU8nCcPZ/mXncygGvtSb36HuFrvT2xhjjE8G4igpY4wxx8AShjHGGJ9YwjDGGOMTSxjGGGN8YgnDGGOMTyxhmJAmIrVBPt8jIjLRT+/VJp6ZazeIyIs9zfoqIiki8g1/nNuYztiwWhPSRKRWVRP8+H4R+snEegHlHbuI/AXYoqp3d1M+D3hJVU8IRnxm4LEWhhlwRCRDRJ4XkZXOMsvZPkNElovIGufrOGf7F0XkWRF5EVgqInNF5C0ReU48z3144shzBZzt+c56rTOZ4DoR+UBEhjjbRzmvV4rIz31sBb3PJ5MtJojIf0RktXiebXCxU+bXwCinVfK/TtnvOOdZLyI/8+O30QxAljDMQPR74HeqejLwBeARZ/tHwBxVnYpnpthfeh1zKnC9qp7pvJ4K3ApMBEYCszo5TzzwgapOBt4Bvup1/t875+9xLh9n/qPP4rkjH6ARuFRVp+F5PstvnYT1PaBEVaeo6ndE5BxgDDADmAJMF5E5PZ3PmK4MtMkHjQE4C5joNQtokogkAsnAX0RkDJ5ZPCO9jnlDVb2fV/ChqpYCiMhaPPMALetwnmY+mdRxFXC2s34qnzyH4Engni7ijPV671V4nmsAnnmAfun88W/H0/IY0snx5zjLGud1Ap4E8k4X5zOmW5YwzEAUBpyqqg3eG0XkAeBNVb3U6Q94y2t3XYf3aPJab6Pz36UW/aSTsKsy3WlQ1Skikown8dwI3I/nWRkZwHRVbRGRHUBMJ8cL8CtV/b9enteYTtklKTMQLcXzrAkAROTIFNLJwB5n/YsBPP8HeC6FAVzVU2FVrcLzyNbbRSQST5zlTrI4AxjuFK0BEr0OfR34svNsBUQkS0QG+6kOZgCyhGFCXZyIlHot38bzxzff6QjehGfKeoDfAL8SkfeA8ADGdCvwbRH5EMgEqno6QFXX4Jm19Co8DyTKF5ECPK2Nj5wyB4H3nGG4/6uqS/Fc8npfRAqB5/h0QjGmV2xYrTFB5jwlsEFVVUSuAuar6sU9HWeM26wPw5jgmw78wRnZdBgXHo1rzLGwFoYxxhifWB+GMcYYn1jCMMYY4xNLGMYYY3xiCcMYY4xPLGEYY4zxiSUMY4wxPvn/8slB3SCiFXUAAAAASUVORK5CYII=\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "learn.fit_one_cycle(12, max_lr=5e-3)",
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<IPython.core.display.HTML object>",
            "text/html": "Total time: 01:56 <p><table style='width:375px; margin-bottom:10px'>\n  <tr>\n    <th>epoch</th>\n    <th>train_loss</th>\n    <th>valid_loss</th>\n    <th>accuracy</th>\n    <th>time</th>\n  </tr>\n  <tr>\n    <th>1</th>\n    <th>0.205190</th>\n    <th>0.341176</th>\n    <th>0.902000</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>2</th>\n    <th>0.144425</th>\n    <th>0.330120</th>\n    <th>0.906600</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>3</th>\n    <th>0.128548</th>\n    <th>0.159783</th>\n    <th>0.957800</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>4</th>\n    <th>0.091207</th>\n    <th>0.085214</th>\n    <th>0.975900</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>5</th>\n    <th>0.064069</th>\n    <th>0.109873</th>\n    <th>0.969100</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>6</th>\n    <th>0.058363</th>\n    <th>0.099653</th>\n    <th>0.972000</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>7</th>\n    <th>0.051490</th>\n    <th>0.049899</th>\n    <th>0.985400</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>8</th>\n    <th>0.034745</th>\n    <th>0.032275</th>\n    <th>0.990500</th>\n    <th>00:10</th>\n  </tr>\n  <tr>\n    <th>9</th>\n    <th>0.025288</th>\n    <th>0.026926</th>\n    <th>0.992100</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>10</th>\n    <th>0.021510</th>\n    <th>0.021373</th>\n    <th>0.992700</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>11</th>\n    <th>0.012670</th>\n    <th>0.016168</th>\n    <th>0.995300</th>\n    <th>00:10</th>\n  </tr>\n  <tr>\n    <th>12</th>\n    <th>0.013594</th>\n    <th>0.016143</th>\n    <th>0.995000</th>\n    <th>00:09</th>\n  </tr>\n</table>\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "terp = learn.interpret()\nterp.plot_top_losses(9, figsize=(7,7))",
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 504x504 with 9 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "We do squeeze out slightly better performance with AdaptiveAvgPool, but the semantic confusion here, I think, is undesirable. "
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "### Swapping out the last layer only _BEST RESULTS_"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "In the original Darknet19, we have a size 1 convolution into an average pool layer at the end. Here, we try fastai's `conv_layer` of size 3 conv -> relu -> bn. "
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "model = nn.Sequential(\n    dn_conv(1, 8),\n    dn_maxp(),\n    dn_conv(8, 16),\n    dn_maxp(),\n    dn_trip(16, 32),\n    dn_maxp(),\n    dn_trip(32, 64),\n    dn_maxp(),\n    dn_trip(64, 128),\n    dn_maxp(),\n    dn_trip(128, 256),\n    dn_conv(256, 128, size=1),\n    dn_conv(128, 256),\n    conv_layer(256, 10),\n    Flatten(),\n    nn.Linear(360, 10)\n)",
      "execution_count": 116,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "learn = Learner(data, model, loss_func = nn.CrossEntropyLoss(), metrics=accuracy)",
      "execution_count": 122,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "print(learn.summary())",
      "execution_count": 118,
      "outputs": [
        {
          "text": "======================================================================\nLayer (type)         Output Shape         Param #    Trainable \n======================================================================\nConv2d               [1, 8, 28, 28]       72         True      \n______________________________________________________________________\nBatchNorm2d          [1, 8, 28, 28]       16         True      \n______________________________________________________________________\nLeakyReLU            [1, 8, 28, 28]       0          False     \n______________________________________________________________________\nMaxPool2d            [1, 8, 14, 14]       0          False     \n______________________________________________________________________\nConv2d               [1, 16, 14, 14]      1,152      True      \n______________________________________________________________________\nBatchNorm2d          [1, 16, 14, 14]      32         True      \n______________________________________________________________________\nLeakyReLU            [1, 16, 14, 14]      0          False     \n______________________________________________________________________\nMaxPool2d            [1, 16, 7, 7]        0          False     \n______________________________________________________________________\nConv2d               [1, 32, 7, 7]        4,608      True      \n______________________________________________________________________\nBatchNorm2d          [1, 32, 7, 7]        64         True      \n______________________________________________________________________\nLeakyReLU            [1, 32, 7, 7]        0          False     \n______________________________________________________________________\nConv2d               [1, 16, 9, 9]        512        True      \n______________________________________________________________________\nBatchNorm2d          [1, 16, 9, 9]        32         True      \n______________________________________________________________________\nLeakyReLU            [1, 16, 9, 9]        0          False     \n______________________________________________________________________\nConv2d               [1, 32, 9, 9]        4,608      True      \n______________________________________________________________________\nBatchNorm2d          [1, 32, 9, 9]        64         True      \n______________________________________________________________________\nLeakyReLU            [1, 32, 9, 9]        0          False     \n______________________________________________________________________\nMaxPool2d            [1, 32, 4, 4]        0          False     \n______________________________________________________________________\nConv2d               [1, 64, 4, 4]        18,432     True      \n______________________________________________________________________\nBatchNorm2d          [1, 64, 4, 4]        128        True      \n______________________________________________________________________\nLeakyReLU            [1, 64, 4, 4]        0          False     \n______________________________________________________________________\nConv2d               [1, 32, 6, 6]        2,048      True      \n______________________________________________________________________\nBatchNorm2d          [1, 32, 6, 6]        64         True      \n______________________________________________________________________\nLeakyReLU            [1, 32, 6, 6]        0          False     \n______________________________________________________________________\nConv2d               [1, 64, 6, 6]        18,432     True      \n______________________________________________________________________\nBatchNorm2d          [1, 64, 6, 6]        128        True      \n______________________________________________________________________\nLeakyReLU            [1, 64, 6, 6]        0          False     \n______________________________________________________________________\nMaxPool2d            [1, 64, 3, 3]        0          False     \n______________________________________________________________________\nConv2d               [1, 128, 3, 3]       73,728     True      \n______________________________________________________________________\nBatchNorm2d          [1, 128, 3, 3]       256        True      \n______________________________________________________________________\nLeakyReLU            [1, 128, 3, 3]       0          False     \n______________________________________________________________________\nConv2d               [1, 64, 5, 5]        8,192      True      \n______________________________________________________________________\nBatchNorm2d          [1, 64, 5, 5]        128        True      \n______________________________________________________________________\nLeakyReLU            [1, 64, 5, 5]        0          False     \n______________________________________________________________________\nConv2d               [1, 128, 5, 5]       73,728     True      \n______________________________________________________________________\nBatchNorm2d          [1, 128, 5, 5]       256        True      \n______________________________________________________________________\nLeakyReLU            [1, 128, 5, 5]       0          False     \n______________________________________________________________________\nMaxPool2d            [1, 128, 2, 2]       0          False     \n______________________________________________________________________\nConv2d               [1, 256, 2, 2]       294,912    True      \n______________________________________________________________________\nBatchNorm2d          [1, 256, 2, 2]       512        True      \n______________________________________________________________________\nLeakyReLU            [1, 256, 2, 2]       0          False     \n______________________________________________________________________\nConv2d               [1, 128, 4, 4]       32,768     True      \n______________________________________________________________________\nBatchNorm2d          [1, 128, 4, 4]       256        True      \n______________________________________________________________________\nLeakyReLU            [1, 128, 4, 4]       0          False     \n______________________________________________________________________\nConv2d               [1, 256, 4, 4]       294,912    True      \n______________________________________________________________________\nBatchNorm2d          [1, 256, 4, 4]       512        True      \n______________________________________________________________________\nLeakyReLU            [1, 256, 4, 4]       0          False     \n______________________________________________________________________\nConv2d               [1, 128, 6, 6]       32,768     True      \n______________________________________________________________________\nBatchNorm2d          [1, 128, 6, 6]       256        True      \n______________________________________________________________________\nLeakyReLU            [1, 128, 6, 6]       0          False     \n______________________________________________________________________\nConv2d               [1, 256, 6, 6]       294,912    True      \n______________________________________________________________________\nBatchNorm2d          [1, 256, 6, 6]       512        True      \n______________________________________________________________________\nLeakyReLU            [1, 256, 6, 6]       0          False     \n______________________________________________________________________\nConv2d               [1, 10, 6, 6]        23,040     True      \n______________________________________________________________________\nReLU                 [1, 10, 6, 6]        0          False     \n______________________________________________________________________\nBatchNorm2d          [1, 10, 6, 6]        20         True      \n______________________________________________________________________\nFlatten              [1, 360]             0          False     \n______________________________________________________________________\nLinear               [1, 10]              3,610      True      \n______________________________________________________________________\n\nTotal params: 1,185,670\nTotal trainable params: 1,185,670\nTotal non-trainable params: 0\n\n",
          "output_type": "stream",
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "learn.lr_find()\nlearn.recorder.plot()",
      "execution_count": 119,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<IPython.core.display.HTML object>",
            "text/html": ""
          },
          "metadata": {}
        },
        {
          "text": "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n",
          "output_type": "stream",
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 432x288 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "learn.fit_one_cycle(12, max_lr=1e-2)",
      "execution_count": 120,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<IPython.core.display.HTML object>",
            "text/html": "Total time: 01:57 <p><table style='width:375px; margin-bottom:10px'>\n  <tr>\n    <th>epoch</th>\n    <th>train_loss</th>\n    <th>valid_loss</th>\n    <th>accuracy</th>\n    <th>time</th>\n  </tr>\n  <tr>\n    <th>1</th>\n    <th>0.194087</th>\n    <th>0.374787</th>\n    <th>0.898900</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>2</th>\n    <th>0.163486</th>\n    <th>0.290694</th>\n    <th>0.926600</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>3</th>\n    <th>0.128151</th>\n    <th>0.261153</th>\n    <th>0.937100</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>4</th>\n    <th>0.099309</th>\n    <th>0.378880</th>\n    <th>0.898200</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>5</th>\n    <th>0.080600</th>\n    <th>0.070875</th>\n    <th>0.980800</th>\n    <th>00:10</th>\n  </tr>\n  <tr>\n    <th>6</th>\n    <th>0.058267</th>\n    <th>0.066629</th>\n    <th>0.982100</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>7</th>\n    <th>0.049646</th>\n    <th>0.084363</th>\n    <th>0.982100</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>8</th>\n    <th>0.040703</th>\n    <th>0.020266</th>\n    <th>0.993300</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>9</th>\n    <th>0.030180</th>\n    <th>0.027555</th>\n    <th>0.991200</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>10</th>\n    <th>0.025278</th>\n    <th>0.015600</th>\n    <th>0.994700</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>11</th>\n    <th>0.019026</th>\n    <th>0.015453</th>\n    <th>0.995500</th>\n    <th>00:09</th>\n  </tr>\n  <tr>\n    <th>12</th>\n    <th>0.016353</th>\n    <th>0.015098</th>\n    <th>0.996000</th>\n    <th>00:09</th>\n  </tr>\n</table>\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "terp = learn.interpret()\nterp.plot_top_losses(9, figsize=(7,7))",
      "execution_count": 121,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 504x504 with 9 Axes>",
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ],
  "metadata": {
    "language_info": {
      "version": "3.7.1",
      "codemirror_mode": {
        "version": 3,
        "name": "ipython"
      },
      "name": "python",
      "mimetype": "text/x-python",
      "pygments_lexer": "ipython3",
      "file_extension": ".py",
      "nbconvert_exporter": "python"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3",
      "language": "python"
    },
    "gist": {
      "id": "",
      "data": {
        "description": "Implementing Darknet19 from scratch using fast.ai - MNIST",
        "public": true
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 2
}