Computer Vision/05_EfficientNet_and_Custom_Weights.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "05_EfficientNet_and_Custom_Weights.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9uvA77XNx1lC",
        "colab_type": "text"
      },
      "source": [
        "# 05 EfficientNet and Custom Pretrained Models\n",
        "\n",
        "This notebook will cover:\n",
        "* Using a `PyTorch` model\n",
        "* Using pre-trained weights for transfer learning\n",
        "* Setting up a `cnn_learner` style `Learner`"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vpdy_FjlyjRo",
        "colab_type": "text"
      },
      "source": [
        "## The Problem:\n",
        "\n",
        "The problem today will be a familiar one, `PETs`, as we are going to focus on the `Learner` instead"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XDlpoIblxsnZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#Run once per session\n",
        "import os\n",
        "!pip install -q feather-format kornia pyarrow wandb nbdev fastprogress fastai2 fastcore --upgrade \n",
        "!pip install torch==1.3.1\n",
        "!pip install torchvision==0.4.2\n",
        "!pip install Pillow==6.2.1 --upgrade\n",
        "os._exit(00)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MuJmE-Qky00E",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from fastai2.vision.all import *"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dBoVN4tJzlGN",
        "colab_type": "text"
      },
      "source": [
        "Let's make our usual dataloaders real quick"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kgBOdALSzjq_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "path = untar_data(URLs.PETS)/'images'\n",
        "fnames = get_image_files(path)\n",
        "pat = r'/([^/]+)_\\d+.*'"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wUjpDlhizyoF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "batch_tfms = [*aug_transforms(size=224, max_warp=0), Normalize.from_stats(*imagenet_stats)]\n",
        "item_tfms = RandomResizedCrop(460, min_scale=0.75, ratio=(1.,1.))\n",
        "bs=64"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aHQCyG0-z5J2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pets = DataBlock(blocks=(ImageBlock, CategoryBlock),\n",
        "                 get_items=get_image_files,\n",
        "                 splitter=RandomSplitter(),\n",
        "                 get_y=RegexLabeller(pat = r'/([^/]+)_\\d+.*'),\n",
        "                 item_tfms=item_tfms,\n",
        "                 batch_tfms=batch_tfms)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mWapcmrzz7RF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "dls = pets.dataloaders(path, bs=bs)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q9JE4JGBz9Ov",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 414
        },
        "outputId": "fa4e296e-0dd6-4ca1-d7ca-5897a90586d7"
      },
      "source": [
        "dls.show_batch(max_n=9, figsize=(6,7))"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWEAAAGNCAYAAADAecAuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nvkL+P4is1b9AkvImAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x504 with 9 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mC_aOWTs0EwY",
        "colab_type": "text"
      },
      "source": [
        "Now let's focus on our `EfficentNet` model. We'll be working out of Ross Wightman's repository [here](https://github.com/rwightman/pytorch-image-models). Included in this repository is tons of pretrained models for almost every major model in Computer Vision. All were for 224x224 training and validation size. Let's install it"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IfllhYgc0EBo",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 207
        },
        "outputId": "7aee3ae9-f525-4307-fc70-d41137c576b9"
      },
      "source": [
        "!pip install timm"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting timm\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/db/61/57823841acf1b5d138e3ae141269e970a21f6730d7ab9a5de673d02ee522/timm-0.1.16-py3-none-any.whl (141kB)\n",
            "\r\u001b[K     |██▎                             | 10kB 24.0MB/s eta 0:00:01\r\u001b[K     |████▋                           | 20kB 2.8MB/s eta 0:00:01\r\u001b[K     |███████                         | 30kB 3.3MB/s eta 0:00:01\r\u001b[K     |█████████▎                      | 40kB 2.9MB/s eta 0:00:01\r\u001b[K     |███████████▋                    | 51kB 3.3MB/s eta 0:00:01\r\u001b[K     |█████████████▉                  | 61kB 3.9MB/s eta 0:00:01\r\u001b[K     |████████████████▏               | 71kB 4.2MB/s eta 0:00:01\r\u001b[K     |██████████████████▌             | 81kB 4.0MB/s eta 0:00:01\r\u001b[K     |████████████████████▉           | 92kB 4.5MB/s eta 0:00:01\r\u001b[K     |███████████████████████▏        | 102kB 4.6MB/s eta 0:00:01\r\u001b[K     |█████████████████████████▍      | 112kB 4.6MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▊    | 122kB 4.6MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████  | 133kB 4.6MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 143kB 4.6MB/s \n",
            "\u001b[?25hRequirement already satisfied: torchvision in /usr/local/lib/python3.6/dist-packages (from timm) (0.4.2)\n",
            "Requirement already satisfied: torch>=1.0 in /usr/local/lib/python3.6/dist-packages (from timm) (1.3.1)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.6/dist-packages (from torchvision->timm) (1.12.0)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from torchvision->timm) (1.17.5)\n",
            "Requirement already satisfied: pillow>=4.1.1 in /usr/local/lib/python3.6/dist-packages (from torchvision->timm) (6.2.1)\n",
            "Installing collected packages: timm\n",
            "Successfully installed timm-0.1.16\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8rYtnlvw0tC7",
        "colab_type": "text"
      },
      "source": [
        "Now we can then use his weights one of two ways. First we'll show the direct way to load it in, then we'll load in the weights ourselves"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EVpUH4sz0qR0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from timm import create_model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WuLEyHQT01fq",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 54
        },
        "outputId": "221f96b2-a424-4d10-ddfd-251b0866bc6f"
      },
      "source": [
        "net = create_model('efficientnet_b3a', pretrained=True)"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading: \"https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b3_ra-a5e2fbc7.pth\" to /root/.cache/torch/checkpoints/efficientnet_b3_ra-a5e2fbc7.pth\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gUWQz3Q-08nu",
        "colab_type": "text"
      },
      "source": [
        "Now let's take a look at our downloaded model, so we know how to modify it for transfer learning. With fastai models we can do something like so:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TXqbYrPt1PIH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "8ae3910a-2bdf-47be-ff1e-5646839c805e"
      },
      "source": [
        "learn = cnn_learner(dls, resnet18)"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading: \"https://download.pytorch.org/models/resnet18-5c106cde.pth\" to /root/.cache/torch/checkpoints/resnet18-5c106cde.pth\n",
            "100%|██████████| 44.7M/44.7M [00:00<00:00, 222MB/s]\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dXtxLZMk1RTi",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 255
        },
        "outputId": "9960a119-fff7-4c7f-f577-fcc1a1c74381"
      },
      "source": [
        "learn.model[-1]"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Sequential(\n",
              "  (0): AdaptiveConcatPool2d(\n",
              "    (ap): AdaptiveAvgPool2d(output_size=1)\n",
              "    (mp): AdaptiveMaxPool2d(output_size=1)\n",
              "  )\n",
              "  (1): full: False\n",
              "  (2): BatchNorm1d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (3): Dropout(p=0.25, inplace=False)\n",
              "  (4): Linear(in_features=1024, out_features=512, bias=False)\n",
              "  (5): ReLU(inplace=True)\n",
              "  (6): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (7): Dropout(p=0.5, inplace=False)\n",
              "  (8): Linear(in_features=512, out_features=37, bias=False)\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "umAcv_Yy1SrK",
        "colab_type": "text"
      },
      "source": [
        "And we see this head of our model! Let's see if we can do this for our `EfficientNet`"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qnOicnbk07Gd",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 163
        },
        "outputId": "1a85ff93-2f80-4edb-c1d1-a6645d7258c5"
      },
      "source": [
        "net[-1]"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "error",
          "ename": "TypeError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-20-9c0a096ee31d>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mnet\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;31mTypeError\u001b[0m: 'EfficientNet' object does not support indexing"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "h7Tv8QzG1XMb",
        "colab_type": "text"
      },
      "source": [
        "No! Why?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l92TYaOg1ZMa",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "f40f13fd-a67c-4747-a8a0-de0919645063"
      },
      "source": [
        "len(learn.model)"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h9EnxuXL1hXa",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 163
        },
        "outputId": "c81c0f0a-20e3-4196-dd17-e4435870b23c"
      },
      "source": [
        "len(net)"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "error",
          "ename": "TypeError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-23-fefbadd730e7>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnet\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;31mTypeError\u001b[0m: object of type 'EfficientNet' has no len()"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SVxdwdv01jTz",
        "colab_type": "text"
      },
      "source": [
        "We can see that our `fastai2` model was **split** into two different layer groups:\n",
        "\n",
        "* Group 1: Our encoder, which is everything but the last layer of our original model\n",
        "* Group 2: Our head, which is a `fastai2` version of a `Linear` layer plus a few extra bits"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Xn0u9TB51igS",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 255
        },
        "outputId": "3de3cd73-ae53-4acc-d3e5-bd4f548d90a4"
      },
      "source": [
        "create_head(2048, 10)"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Sequential(\n",
              "  (0): AdaptiveConcatPool2d(\n",
              "    (ap): AdaptiveAvgPool2d(output_size=1)\n",
              "    (mp): AdaptiveMaxPool2d(output_size=1)\n",
              "  )\n",
              "  (1): full: False\n",
              "  (2): BatchNorm1d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (3): Dropout(p=0.25, inplace=False)\n",
              "  (4): Linear(in_features=2048, out_features=512, bias=False)\n",
              "  (5): ReLU(inplace=True)\n",
              "  (6): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (7): Dropout(p=0.5, inplace=False)\n",
              "  (8): Linear(in_features=512, out_features=10, bias=False)\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GFkZpDDb16JQ",
        "colab_type": "text"
      },
      "source": [
        "How do we do this for our model? Let's take a look at it:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1Gwb7t5a1zDH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "4c2f6109-9312-495f-d2ac-00c5e0ff87b9"
      },
      "source": [
        "net"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "EfficientNet(\n",
              "  (conv_stem): Conv2d(3, 40, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
              "  (bn1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (act1): Swish()\n",
              "  (blocks): Sequential(\n",
              "    (0): Sequential(\n",
              "      (0): DepthwiseSeparableConv(\n",
              "        (conv_dw): Conv2d(40, 40, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=40, bias=False)\n",
              "        (bn1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(40, 10, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(10, 40, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pw): Conv2d(40, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn2): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Identity()\n",
              "      )\n",
              "      (1): DepthwiseSeparableConv(\n",
              "        (conv_dw): Conv2d(24, 24, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=24, bias=False)\n",
              "        (bn1): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(24, 6, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(6, 24, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pw): Conv2d(24, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn2): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Identity()\n",
              "      )\n",
              "    )\n",
              "    (1): Sequential(\n",
              "      (0): InvertedResidual(\n",
              "        (conv_pw): Conv2d(24, 144, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(144, 144, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=144, bias=False)\n",
              "        (bn2): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(144, 6, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(6, 144, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(144, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (1): InvertedResidual(\n",
              "        (conv_pw): Conv2d(32, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=192, bias=False)\n",
              "        (bn2): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(192, 8, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(8, 192, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(192, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (2): InvertedResidual(\n",
              "        (conv_pw): Conv2d(32, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=192, bias=False)\n",
              "        (bn2): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(192, 8, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(8, 192, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(192, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (2): Sequential(\n",
              "      (0): InvertedResidual(\n",
              "        (conv_pw): Conv2d(32, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(192, 192, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), groups=192, bias=False)\n",
              "        (bn2): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(192, 8, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(8, 192, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(192, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (1): InvertedResidual(\n",
              "        (conv_pw): Conv2d(48, 288, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(288, 288, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=288, bias=False)\n",
              "        (bn2): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(288, 12, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(12, 288, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(288, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (2): InvertedResidual(\n",
              "        (conv_pw): Conv2d(48, 288, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(288, 288, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=288, bias=False)\n",
              "        (bn2): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(288, 12, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(12, 288, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(288, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (3): Sequential(\n",
              "      (0): InvertedResidual(\n",
              "        (conv_pw): Conv2d(48, 288, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(288, 288, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=288, bias=False)\n",
              "        (bn2): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(288, 12, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(12, 288, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(288, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (1): InvertedResidual(\n",
              "        (conv_pw): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(576, 576, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=576, bias=False)\n",
              "        (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(576, 24, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(24, 576, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (2): InvertedResidual(\n",
              "        (conv_pw): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(576, 576, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=576, bias=False)\n",
              "        (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(576, 24, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(24, 576, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (3): InvertedResidual(\n",
              "        (conv_pw): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(576, 576, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=576, bias=False)\n",
              "        (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(576, 24, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(24, 576, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (4): InvertedResidual(\n",
              "        (conv_pw): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(576, 576, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=576, bias=False)\n",
              "        (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(576, 24, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(24, 576, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (4): Sequential(\n",
              "      (0): InvertedResidual(\n",
              "        (conv_pw): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(576, 576, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=576, bias=False)\n",
              "        (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(576, 24, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(24, 576, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(576, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(136, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (1): InvertedResidual(\n",
              "        (conv_pw): Conv2d(136, 816, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(816, 816, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=816, bias=False)\n",
              "        (bn2): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(816, 34, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(34, 816, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(816, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(136, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (2): InvertedResidual(\n",
              "        (conv_pw): Conv2d(136, 816, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(816, 816, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=816, bias=False)\n",
              "        (bn2): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(816, 34, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(34, 816, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(816, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(136, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (3): InvertedResidual(\n",
              "        (conv_pw): Conv2d(136, 816, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(816, 816, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=816, bias=False)\n",
              "        (bn2): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(816, 34, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(34, 816, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(816, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(136, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (4): InvertedResidual(\n",
              "        (conv_pw): Conv2d(136, 816, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(816, 816, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=816, bias=False)\n",
              "        (bn2): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(816, 34, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(34, 816, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(816, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(136, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (5): Sequential(\n",
              "      (0): InvertedResidual(\n",
              "        (conv_pw): Conv2d(136, 816, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(816, 816, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), groups=816, bias=False)\n",
              "        (bn2): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(816, 34, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(34, 816, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(816, 232, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(232, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (1): InvertedResidual(\n",
              "        (conv_pw): Conv2d(232, 1392, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(1392, 1392, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=1392, bias=False)\n",
              "        (bn2): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(1392, 58, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(58, 1392, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(1392, 232, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(232, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (2): InvertedResidual(\n",
              "        (conv_pw): Conv2d(232, 1392, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(1392, 1392, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=1392, bias=False)\n",
              "        (bn2): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(1392, 58, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(58, 1392, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(1392, 232, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(232, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (3): InvertedResidual(\n",
              "        (conv_pw): Conv2d(232, 1392, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(1392, 1392, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=1392, bias=False)\n",
              "        (bn2): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(1392, 58, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(58, 1392, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(1392, 232, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(232, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (4): InvertedResidual(\n",
              "        (conv_pw): Conv2d(232, 1392, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(1392, 1392, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=1392, bias=False)\n",
              "        (bn2): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(1392, 58, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(58, 1392, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(1392, 232, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(232, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (5): InvertedResidual(\n",
              "        (conv_pw): Conv2d(232, 1392, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(1392, 1392, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=1392, bias=False)\n",
              "        (bn2): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(1392, 58, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(58, 1392, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(1392, 232, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(232, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (6): Sequential(\n",
              "      (0): InvertedResidual(\n",
              "        (conv_pw): Conv2d(232, 1392, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(1392, 1392, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=1392, bias=False)\n",
              "        (bn2): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(1392, 58, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(58, 1392, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(1392, 384, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (1): InvertedResidual(\n",
              "        (conv_pw): Conv2d(384, 2304, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(2304, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(2304, 2304, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=2304, bias=False)\n",
              "        (bn2): BatchNorm2d(2304, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(2304, 96, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(96, 2304, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(2304, 384, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "  )\n",
              "  (conv_head): Conv2d(384, 1536, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "  (bn2): BatchNorm2d(1536, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (act2): Swish()\n",
              "  (global_pool): SelectAdaptivePool2d (output_size=1, pool_type=avg)\n",
              "  (classifier): Linear(in_features=1536, out_features=1000, bias=True)\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mwWI4TRu19jo",
        "colab_type": "text"
      },
      "source": [
        "We can see that our `Pooling` layer and our `Linear` layer is the last two layers of our model. Let's pop those off"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7ccPsb1p2ZoH",
        "colab_type": "text"
      },
      "source": [
        "Now if we use the original `fastai2` `create_body` function, we'll get an error:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vmsqRiIR1833",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 306
        },
        "outputId": "2040f592-93d2-42ce-a042-7ffebe150340"
      },
      "source": [
        "body = create_body(net, pretrained=False, cut=-1)"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "error",
          "ename": "TypeError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-29-d13ee58c7af6>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mbody\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcreate_body\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnet\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpretrained\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcut\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/fastai2/vision/learner.py\u001b[0m in \u001b[0;36mcreate_body\u001b[0;34m(arch, pretrained, cut)\u001b[0m\n\u001b[1;32m     25\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mcreate_body\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0march\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpretrained\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcut\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     26\u001b[0m     \u001b[0;34m\"Cut off the body of a typically pretrained `arch` as determined by `cut`\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 27\u001b[0;31m     \u001b[0mmodel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0march\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpretrained\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mpretrained\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     28\u001b[0m     \u001b[0;31m#cut = ifnone(cut, cnn_config(arch)['cut'])\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     29\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mcut\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    539\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    540\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 541\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    542\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mhook\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    543\u001b[0m             \u001b[0mhook_result\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhook\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mresult\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mTypeError\u001b[0m: forward() got an unexpected keyword argument 'pretrained'"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zA82u3JM2eDX",
        "colab_type": "text"
      },
      "source": [
        "Why? Let's take a look"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zHJTsBkj2IRC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def create_body(arch, pretrained=True, cut=None):\n",
        "    \"Cut off the body of a typically pretrained `arch` as determined by `cut`\"\n",
        "    model = arch(pretrained=pretrained)\n",
        "    #cut = ifnone(cut, cnn_config(arch)['cut'])\n",
        "    if cut is None:\n",
        "        ll = list(enumerate(model.children()))\n",
        "        cut = next(i for i,o in reversed(ll) if has_pool_type(o))\n",
        "    if   isinstance(cut, int): return nn.Sequential(*list(model.children())[:cut])\n",
        "    elif callable(cut): return cut(model)\n",
        "    else: raise NamedError(\"cut must be either integer or a function\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dHT0Cjad2lky",
        "colab_type": "text"
      },
      "source": [
        "We can see that arch needs to be a **generator**. Let's try to make a function to help us with specifically his library"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k-leIsOU2pJ_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def create_timm_body(arch:str, pretrained=True, cut=None):\n",
        "  model = create_model(arch, pretrained=pretrained)\n",
        "  if cut is None:\n",
        "    ll = list(enumerate(model.children()))\n",
        "    cut = next(i for i,o in reversed(ll) if has_pool_type(o))\n",
        "  if isinstance(cut, int): return nn.Sequential(*list(model.children())[:cut])\n",
        "  elif callable(cut): return cut(model)\n",
        "  else: raise NamedError(\"cut must be either integer or function\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hLLq5Gjm3YYa",
        "colab_type": "text"
      },
      "source": [
        "Let's try it out!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mFze6_w93ZtL",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "body = create_timm_body('efficientnet_b3a', pretrained=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H2I2lIhp3Xqg",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "54bf82e3-12e9-4a04-e106-26ea9f5ad495"
      },
      "source": [
        "len(body)"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "7"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DSWsuOBj3hTy",
        "colab_type": "text"
      },
      "source": [
        "Now we can see that we have seven seperate splits"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ijFdYP1k3gwZ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "8d0fbe97-b199-405b-c328-8d586e1dff1a"
      },
      "source": [
        "body"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Sequential(\n",
              "  (0): Conv2d(3, 40, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
              "  (1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (2): Swish()\n",
              "  (3): Sequential(\n",
              "    (0): Sequential(\n",
              "      (0): DepthwiseSeparableConv(\n",
              "        (conv_dw): Conv2d(40, 40, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=40, bias=False)\n",
              "        (bn1): BatchNorm2d(40, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(40, 10, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(10, 40, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pw): Conv2d(40, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn2): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Identity()\n",
              "      )\n",
              "      (1): DepthwiseSeparableConv(\n",
              "        (conv_dw): Conv2d(24, 24, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=24, bias=False)\n",
              "        (bn1): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(24, 6, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(6, 24, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pw): Conv2d(24, 24, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn2): BatchNorm2d(24, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Identity()\n",
              "      )\n",
              "    )\n",
              "    (1): Sequential(\n",
              "      (0): InvertedResidual(\n",
              "        (conv_pw): Conv2d(24, 144, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(144, 144, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=144, bias=False)\n",
              "        (bn2): BatchNorm2d(144, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(144, 6, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(6, 144, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(144, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (1): InvertedResidual(\n",
              "        (conv_pw): Conv2d(32, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=192, bias=False)\n",
              "        (bn2): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(192, 8, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(8, 192, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(192, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (2): InvertedResidual(\n",
              "        (conv_pw): Conv2d(32, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=192, bias=False)\n",
              "        (bn2): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(192, 8, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(8, 192, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(192, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (2): Sequential(\n",
              "      (0): InvertedResidual(\n",
              "        (conv_pw): Conv2d(32, 192, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(192, 192, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), groups=192, bias=False)\n",
              "        (bn2): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(192, 8, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(8, 192, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(192, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (1): InvertedResidual(\n",
              "        (conv_pw): Conv2d(48, 288, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(288, 288, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=288, bias=False)\n",
              "        (bn2): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(288, 12, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(12, 288, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(288, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (2): InvertedResidual(\n",
              "        (conv_pw): Conv2d(48, 288, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(288, 288, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=288, bias=False)\n",
              "        (bn2): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(288, 12, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(12, 288, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(288, 48, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (3): Sequential(\n",
              "      (0): InvertedResidual(\n",
              "        (conv_pw): Conv2d(48, 288, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(288, 288, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=288, bias=False)\n",
              "        (bn2): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(288, 12, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(12, 288, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(288, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (1): InvertedResidual(\n",
              "        (conv_pw): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(576, 576, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=576, bias=False)\n",
              "        (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(576, 24, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(24, 576, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (2): InvertedResidual(\n",
              "        (conv_pw): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(576, 576, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=576, bias=False)\n",
              "        (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(576, 24, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(24, 576, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (3): InvertedResidual(\n",
              "        (conv_pw): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(576, 576, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=576, bias=False)\n",
              "        (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(576, 24, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(24, 576, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (4): InvertedResidual(\n",
              "        (conv_pw): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(576, 576, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=576, bias=False)\n",
              "        (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(576, 24, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(24, 576, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(576, 96, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (4): Sequential(\n",
              "      (0): InvertedResidual(\n",
              "        (conv_pw): Conv2d(96, 576, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(576, 576, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=576, bias=False)\n",
              "        (bn2): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(576, 24, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(24, 576, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(576, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(136, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (1): InvertedResidual(\n",
              "        (conv_pw): Conv2d(136, 816, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(816, 816, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=816, bias=False)\n",
              "        (bn2): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(816, 34, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(34, 816, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(816, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(136, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (2): InvertedResidual(\n",
              "        (conv_pw): Conv2d(136, 816, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(816, 816, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=816, bias=False)\n",
              "        (bn2): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(816, 34, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(34, 816, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(816, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(136, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (3): InvertedResidual(\n",
              "        (conv_pw): Conv2d(136, 816, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(816, 816, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=816, bias=False)\n",
              "        (bn2): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(816, 34, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(34, 816, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(816, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(136, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (4): InvertedResidual(\n",
              "        (conv_pw): Conv2d(136, 816, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(816, 816, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=816, bias=False)\n",
              "        (bn2): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(816, 34, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(34, 816, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(816, 136, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(136, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (5): Sequential(\n",
              "      (0): InvertedResidual(\n",
              "        (conv_pw): Conv2d(136, 816, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(816, 816, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2), groups=816, bias=False)\n",
              "        (bn2): BatchNorm2d(816, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(816, 34, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(34, 816, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(816, 232, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(232, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (1): InvertedResidual(\n",
              "        (conv_pw): Conv2d(232, 1392, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(1392, 1392, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=1392, bias=False)\n",
              "        (bn2): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(1392, 58, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(58, 1392, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(1392, 232, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(232, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (2): InvertedResidual(\n",
              "        (conv_pw): Conv2d(232, 1392, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(1392, 1392, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=1392, bias=False)\n",
              "        (bn2): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(1392, 58, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(58, 1392, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(1392, 232, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(232, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (3): InvertedResidual(\n",
              "        (conv_pw): Conv2d(232, 1392, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(1392, 1392, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=1392, bias=False)\n",
              "        (bn2): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(1392, 58, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(58, 1392, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(1392, 232, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(232, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (4): InvertedResidual(\n",
              "        (conv_pw): Conv2d(232, 1392, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(1392, 1392, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=1392, bias=False)\n",
              "        (bn2): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(1392, 58, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(58, 1392, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(1392, 232, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(232, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (5): InvertedResidual(\n",
              "        (conv_pw): Conv2d(232, 1392, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(1392, 1392, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2), groups=1392, bias=False)\n",
              "        (bn2): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(1392, 58, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(58, 1392, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(1392, 232, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(232, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "    (6): Sequential(\n",
              "      (0): InvertedResidual(\n",
              "        (conv_pw): Conv2d(232, 1392, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(1392, 1392, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=1392, bias=False)\n",
              "        (bn2): BatchNorm2d(1392, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(1392, 58, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(58, 1392, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(1392, 384, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "      (1): InvertedResidual(\n",
              "        (conv_pw): Conv2d(384, 2304, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn1): BatchNorm2d(2304, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act1): Swish()\n",
              "        (conv_dw): Conv2d(2304, 2304, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=2304, bias=False)\n",
              "        (bn2): BatchNorm2d(2304, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "        (act2): Swish()\n",
              "        (se): SqueezeExcite(\n",
              "          (avg_pool): AdaptiveAvgPool2d(output_size=1)\n",
              "          (conv_reduce): Conv2d(2304, 96, kernel_size=(1, 1), stride=(1, 1))\n",
              "          (act1): Swish()\n",
              "          (conv_expand): Conv2d(96, 2304, kernel_size=(1, 1), stride=(1, 1))\n",
              "        )\n",
              "        (conv_pwl): Conv2d(2304, 384, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "        (bn3): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "      )\n",
              "    )\n",
              "  )\n",
              "  (4): Conv2d(384, 1536, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
              "  (5): BatchNorm2d(1536, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (6): Swish()\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wdbWezt33nji",
        "colab_type": "text"
      },
      "source": [
        "But we've popped off the last layers we need! Let's move onto our head of the model. We know the input should be `3072` (we can see this in the last linear layer of the original model). We need it 2x it because of our `AdaptiveConcatPooling` We want it to have an output to our classes. But what if we dont' know that?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "s7Dl7GUf4IwK",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "73e2aac8-418e-4eb6-b4d0-52ecd3f81a33"
      },
      "source": [
        "nf = num_features_model(nn.Sequential(*body.children())) * (2); nf"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3072"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RtKo2gZI3z2H",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "head = create_head(nf, dls.c)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8cs1079U34Rq",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 255
        },
        "outputId": "7819c9a7-05cf-4b04-fef8-96e1c4b3192d"
      },
      "source": [
        "head"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Sequential(\n",
              "  (0): AdaptiveConcatPool2d(\n",
              "    (ap): AdaptiveAvgPool2d(output_size=1)\n",
              "    (mp): AdaptiveMaxPool2d(output_size=1)\n",
              "  )\n",
              "  (1): full: False\n",
              "  (2): BatchNorm1d(3072, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (3): Dropout(p=0.25, inplace=False)\n",
              "  (4): Linear(in_features=3072, out_features=512, bias=False)\n",
              "  (5): ReLU(inplace=True)\n",
              "  (6): BatchNorm1d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
              "  (7): Dropout(p=0.5, inplace=False)\n",
              "  (8): Linear(in_features=512, out_features=37, bias=False)\n",
              ")"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S5E3DkE637zF",
        "colab_type": "text"
      },
      "source": [
        "Now finally we need to wrap it together"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cuX7gbp337FZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = nn.Sequential(body, head)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PlNm4nvs4TbD",
        "colab_type": "text"
      },
      "source": [
        "And then we initialize our new head of our model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S_rWhCb14TD6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "apply_init(model[1], nn.init.kaiming_normal_)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bvRWb8OD4ZmY",
        "colab_type": "text"
      },
      "source": [
        "Now we have our two layer-long model! What's next?"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FJOo0dVz4YGn",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "db176919-e2f5-4357-daba-b3bf7d01f3a3"
      },
      "source": [
        "len(model)"
      ],
      "execution_count": 71,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 71
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kHY0Rc484dhC",
        "colab_type": "text"
      },
      "source": [
        "Let's try making a `Learner`"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nRo7QWwO4ZEn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "learn = Learner(dls, model, loss_func=LabelSmoothingCrossEntropy())"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KhYpDsVB4i8g",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "b0944269-69ed-46e3-eef4-90da3c0cddda"
      },
      "source": [
        "learn.summary()"
      ],
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Sequential (Input shape: ['64 x 3 x 224 x 224'])\n",
              "================================================================\n",
              "Layer (type)         Output Shape         Param #    Trainable \n",
              "================================================================\n",
              "Conv2d               64 x 40 x 112 x 112  1,080      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 40 x 112 x 112  80         True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 40 x 112 x 112  0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 40 x 112 x 112  360        True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 40 x 112 x 112  80         True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 40 x 112 x 112  0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 40 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 10 x 1 x 1      410        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 10 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 40 x 1 x 1      440        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 112 x 112  960        True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 24 x 112 x 112  48         True      \n",
              "________________________________________________________________\n",
              "Identity             64 x 24 x 112 x 112  0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 112 x 112  216        True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 24 x 112 x 112  48         True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 24 x 112 x 112  0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 24 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 6 x 1 x 1       150        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 6 x 1 x 1       0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 1 x 1      168        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 112 x 112  576        True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 24 x 112 x 112  48         True      \n",
              "________________________________________________________________\n",
              "Identity             64 x 24 x 112 x 112  0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 144 x 112 x 112 3,456      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 144 x 112 x 112 288        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 144 x 112 x 112 0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 144 x 56 x 56   1,296      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 144 x 56 x 56   288        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 144 x 56 x 56   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 144 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 6 x 1 x 1       870        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 6 x 1 x 1       0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 144 x 1 x 1     1,008      True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 32 x 56 x 56    4,608      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 32 x 56 x 56    64         True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 56 x 56   6,144      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 192 x 56 x 56   384        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 192 x 56 x 56   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 56 x 56   1,728      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 192 x 56 x 56   384        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 192 x 56 x 56   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 192 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 8 x 1 x 1       1,544      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 8 x 1 x 1       0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 1 x 1     1,728      True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 32 x 56 x 56    6,144      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 32 x 56 x 56    64         True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 56 x 56   6,144      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 192 x 56 x 56   384        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 192 x 56 x 56   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 56 x 56   1,728      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 192 x 56 x 56   384        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 192 x 56 x 56   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 192 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 8 x 1 x 1       1,544      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 8 x 1 x 1       0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 1 x 1     1,728      True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 32 x 56 x 56    6,144      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 32 x 56 x 56    64         True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 56 x 56   6,144      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 192 x 56 x 56   384        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 192 x 56 x 56   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 28 x 28   4,800      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 192 x 28 x 28   384        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 192 x 28 x 28   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 192 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 8 x 1 x 1       1,544      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 8 x 1 x 1       0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 1 x 1     1,728      True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 48 x 28 x 28    9,216      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 48 x 28 x 28    96         True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 28 x 28   13,824     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 288 x 28 x 28   576        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 288 x 28 x 28   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 28 x 28   7,200      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 288 x 28 x 28   576        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 288 x 28 x 28   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 288 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 12 x 1 x 1      3,468      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 12 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 1 x 1     3,744      True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 48 x 28 x 28    13,824     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 48 x 28 x 28    96         True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 28 x 28   13,824     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 288 x 28 x 28   576        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 288 x 28 x 28   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 28 x 28   7,200      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 288 x 28 x 28   576        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 288 x 28 x 28   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 288 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 12 x 1 x 1      3,468      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 12 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 1 x 1     3,744      True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 48 x 28 x 28    13,824     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 48 x 28 x 28    96         True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 28 x 28   13,824     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 288 x 28 x 28   576        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 288 x 28 x 28   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 14 x 14   2,592      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 288 x 14 x 14   576        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 288 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 288 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 12 x 1 x 1      3,468      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 12 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 1 x 1     3,744      True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 96 x 14 x 14    27,648     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 96 x 14 x 14    192        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   55,296     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   5,184      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 576 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 1 x 1      13,848     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 24 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 1 x 1     14,400     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 96 x 14 x 14    55,296     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 96 x 14 x 14    192        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   55,296     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   5,184      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 576 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 1 x 1      13,848     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 24 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 1 x 1     14,400     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 96 x 14 x 14    55,296     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 96 x 14 x 14    192        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   55,296     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   5,184      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 576 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 1 x 1      13,848     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 24 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 1 x 1     14,400     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 96 x 14 x 14    55,296     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 96 x 14 x 14    192        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   55,296     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   5,184      True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 576 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 1 x 1      13,848     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 24 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 1 x 1     14,400     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 96 x 14 x 14    55,296     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 96 x 14 x 14    192        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   55,296     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   14,400     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 576 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 1 x 1      13,848     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 24 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 1 x 1     14,400     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 136 x 14 x 14   78,336     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 136 x 14 x 14   272        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   110,976    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   20,400     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 816 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 34 x 1 x 1      27,778     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 34 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 1 x 1     28,560     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 136 x 14 x 14   110,976    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 136 x 14 x 14   272        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   110,976    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   20,400     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 816 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 34 x 1 x 1      27,778     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 34 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 1 x 1     28,560     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 136 x 14 x 14   110,976    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 136 x 14 x 14   272        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   110,976    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   20,400     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 816 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 34 x 1 x 1      27,778     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 34 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 1 x 1     28,560     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 136 x 14 x 14   110,976    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 136 x 14 x 14   272        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   110,976    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   20,400     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 816 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 34 x 1 x 1      27,778     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 34 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 1 x 1     28,560     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 136 x 14 x 14   110,976    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 136 x 14 x 14   272        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   110,976    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 7 x 7     20,400     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 7 x 7     1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 7 x 7     0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 816 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 34 x 1 x 1      27,778     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 34 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 1 x 1     28,560     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 232 x 7 x 7     189,312    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 232 x 7 x 7     464        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    322,944    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    34,800     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1392 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 58 x 1 x 1      80,794     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 58 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 1 x 1    82,128     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 232 x 7 x 7     322,944    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 232 x 7 x 7     464        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    322,944    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    34,800     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1392 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 58 x 1 x 1      80,794     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 58 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 1 x 1    82,128     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 232 x 7 x 7     322,944    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 232 x 7 x 7     464        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    322,944    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    34,800     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1392 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 58 x 1 x 1      80,794     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 58 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 1 x 1    82,128     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 232 x 7 x 7     322,944    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 232 x 7 x 7     464        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    322,944    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    34,800     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1392 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 58 x 1 x 1      80,794     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 58 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 1 x 1    82,128     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 232 x 7 x 7     322,944    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 232 x 7 x 7     464        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    322,944    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    34,800     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1392 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 58 x 1 x 1      80,794     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 58 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 1 x 1    82,128     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 232 x 7 x 7     322,944    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 232 x 7 x 7     464        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    322,944    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    12,528     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1392 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 58 x 1 x 1      80,794     True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 58 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 1 x 1    82,128     True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 384 x 7 x 7     534,528    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 384 x 7 x 7     768        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 2304 x 7 x 7    884,736    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 2304 x 7 x 7    4,608      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 2304 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 2304 x 7 x 7    20,736     True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 2304 x 7 x 7    4,608      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 2304 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 2304 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 96 x 1 x 1      221,280    True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 96 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 2304 x 1 x 1    223,488    True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 384 x 7 x 7     884,736    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 384 x 7 x 7     768        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1536 x 7 x 7    589,824    True      \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1536 x 7 x 7    3,072      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1536 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1536 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveMaxPool2d    64 x 1536 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Flatten              64 x 3072            0          False     \n",
              "________________________________________________________________\n",
              "BatchNorm1d          64 x 3072            6,144      True      \n",
              "________________________________________________________________\n",
              "Dropout              64 x 3072            0          False     \n",
              "________________________________________________________________\n",
              "Linear               64 x 512             1,572,864  True      \n",
              "________________________________________________________________\n",
              "ReLU                 64 x 512             0          False     \n",
              "________________________________________________________________\n",
              "BatchNorm1d          64 x 512             1,024      True      \n",
              "________________________________________________________________\n",
              "Dropout              64 x 512             0          False     \n",
              "________________________________________________________________\n",
              "Linear               64 x 37              18,944     True      \n",
              "________________________________________________________________\n",
              "\n",
              "Total params: 12,295,208\n",
              "Total trainable params: 12,295,208\n",
              "Total non-trainable params: 0\n",
              "\n",
              "Optimizer used: <function Adam at 0x7f5fda4b96a8>\n",
              "Loss function: LabelSmoothingCrossEntropy()\n",
              "\n",
              "Callbacks:\n",
              "  - TrainEvalCallback\n",
              "  - Recorder\n",
              "  - ProgressCallback"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ordXWift5MMP",
        "colab_type": "text"
      },
      "source": [
        "Oh no! It isn't frozen, what do we do? We never split the model! Since we have it set to where `model[0]` is the first group and `model[1]` is the second group, we can use the `default_split` splitter. Let's try again"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5TC2WTFm5Vad",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "learn = Learner(dls, model, loss_func=LabelSmoothingCrossEntropy(), \n",
        "                splitter=default_split, metrics=accuracy)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kJZ3fvuZ4kct",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "31338991-1f2a-4b45-9324-ed1c8b991933"
      },
      "source": [
        "learn.freeze()\n",
        "learn.summary()"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Sequential (Input shape: ['64 x 3 x 224 x 224'])\n",
              "================================================================\n",
              "Layer (type)         Output Shape         Param #    Trainable \n",
              "================================================================\n",
              "Conv2d               64 x 40 x 112 x 112  1,080      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 40 x 112 x 112  80         True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 40 x 112 x 112  0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 40 x 112 x 112  360        False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 40 x 112 x 112  80         True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 40 x 112 x 112  0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 40 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 10 x 1 x 1      410        False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 10 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 40 x 1 x 1      440        False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 112 x 112  960        False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 24 x 112 x 112  48         True      \n",
              "________________________________________________________________\n",
              "Identity             64 x 24 x 112 x 112  0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 112 x 112  216        False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 24 x 112 x 112  48         True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 24 x 112 x 112  0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 24 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 6 x 1 x 1       150        False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 6 x 1 x 1       0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 1 x 1      168        False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 112 x 112  576        False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 24 x 112 x 112  48         True      \n",
              "________________________________________________________________\n",
              "Identity             64 x 24 x 112 x 112  0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 144 x 112 x 112 3,456      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 144 x 112 x 112 288        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 144 x 112 x 112 0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 144 x 56 x 56   1,296      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 144 x 56 x 56   288        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 144 x 56 x 56   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 144 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 6 x 1 x 1       870        False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 6 x 1 x 1       0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 144 x 1 x 1     1,008      False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 32 x 56 x 56    4,608      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 32 x 56 x 56    64         True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 56 x 56   6,144      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 192 x 56 x 56   384        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 192 x 56 x 56   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 56 x 56   1,728      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 192 x 56 x 56   384        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 192 x 56 x 56   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 192 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 8 x 1 x 1       1,544      False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 8 x 1 x 1       0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 1 x 1     1,728      False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 32 x 56 x 56    6,144      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 32 x 56 x 56    64         True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 56 x 56   6,144      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 192 x 56 x 56   384        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 192 x 56 x 56   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 56 x 56   1,728      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 192 x 56 x 56   384        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 192 x 56 x 56   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 192 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 8 x 1 x 1       1,544      False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 8 x 1 x 1       0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 1 x 1     1,728      False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 32 x 56 x 56    6,144      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 32 x 56 x 56    64         True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 56 x 56   6,144      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 192 x 56 x 56   384        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 192 x 56 x 56   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 28 x 28   4,800      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 192 x 28 x 28   384        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 192 x 28 x 28   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 192 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 8 x 1 x 1       1,544      False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 8 x 1 x 1       0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 192 x 1 x 1     1,728      False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 48 x 28 x 28    9,216      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 48 x 28 x 28    96         True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 28 x 28   13,824     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 288 x 28 x 28   576        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 288 x 28 x 28   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 28 x 28   7,200      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 288 x 28 x 28   576        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 288 x 28 x 28   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 288 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 12 x 1 x 1      3,468      False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 12 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 1 x 1     3,744      False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 48 x 28 x 28    13,824     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 48 x 28 x 28    96         True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 28 x 28   13,824     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 288 x 28 x 28   576        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 288 x 28 x 28   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 28 x 28   7,200      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 288 x 28 x 28   576        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 288 x 28 x 28   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 288 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 12 x 1 x 1      3,468      False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 12 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 1 x 1     3,744      False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 48 x 28 x 28    13,824     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 48 x 28 x 28    96         True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 28 x 28   13,824     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 288 x 28 x 28   576        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 288 x 28 x 28   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 14 x 14   2,592      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 288 x 14 x 14   576        True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 288 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 288 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 12 x 1 x 1      3,468      False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 12 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 288 x 1 x 1     3,744      False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 96 x 14 x 14    27,648     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 96 x 14 x 14    192        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   55,296     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   5,184      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 576 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 1 x 1      13,848     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 24 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 1 x 1     14,400     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 96 x 14 x 14    55,296     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 96 x 14 x 14    192        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   55,296     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   5,184      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 576 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 1 x 1      13,848     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 24 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 1 x 1     14,400     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 96 x 14 x 14    55,296     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 96 x 14 x 14    192        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   55,296     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   5,184      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 576 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 1 x 1      13,848     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 24 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 1 x 1     14,400     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 96 x 14 x 14    55,296     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 96 x 14 x 14    192        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   55,296     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   5,184      False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 576 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 1 x 1      13,848     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 24 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 1 x 1     14,400     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 96 x 14 x 14    55,296     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 96 x 14 x 14    192        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   55,296     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 14 x 14   14,400     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 576 x 14 x 14   1,152      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 576 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 576 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 24 x 1 x 1      13,848     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 24 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 576 x 1 x 1     14,400     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 136 x 14 x 14   78,336     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 136 x 14 x 14   272        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   110,976    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   20,400     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 816 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 34 x 1 x 1      27,778     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 34 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 1 x 1     28,560     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 136 x 14 x 14   110,976    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 136 x 14 x 14   272        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   110,976    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   20,400     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 816 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 34 x 1 x 1      27,778     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 34 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 1 x 1     28,560     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 136 x 14 x 14   110,976    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 136 x 14 x 14   272        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   110,976    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   20,400     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 816 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 34 x 1 x 1      27,778     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 34 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 1 x 1     28,560     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 136 x 14 x 14   110,976    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 136 x 14 x 14   272        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   110,976    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   20,400     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 816 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 34 x 1 x 1      27,778     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 34 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 1 x 1     28,560     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 136 x 14 x 14   110,976    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 136 x 14 x 14   272        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 14 x 14   110,976    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 14 x 14   1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 14 x 14   0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 7 x 7     20,400     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 816 x 7 x 7     1,632      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 816 x 7 x 7     0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 816 x 1 x 1     0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 34 x 1 x 1      27,778     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 34 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 816 x 1 x 1     28,560     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 232 x 7 x 7     189,312    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 232 x 7 x 7     464        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    322,944    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    34,800     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1392 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 58 x 1 x 1      80,794     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 58 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 1 x 1    82,128     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 232 x 7 x 7     322,944    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 232 x 7 x 7     464        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    322,944    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    34,800     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1392 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 58 x 1 x 1      80,794     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 58 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 1 x 1    82,128     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 232 x 7 x 7     322,944    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 232 x 7 x 7     464        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    322,944    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    34,800     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1392 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 58 x 1 x 1      80,794     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 58 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 1 x 1    82,128     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 232 x 7 x 7     322,944    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 232 x 7 x 7     464        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    322,944    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    34,800     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1392 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 58 x 1 x 1      80,794     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 58 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 1 x 1    82,128     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 232 x 7 x 7     322,944    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 232 x 7 x 7     464        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    322,944    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    34,800     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1392 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 58 x 1 x 1      80,794     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 58 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 1 x 1    82,128     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 232 x 7 x 7     322,944    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 232 x 7 x 7     464        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    322,944    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 7 x 7    12,528     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1392 x 7 x 7    2,784      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1392 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1392 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 58 x 1 x 1      80,794     False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 58 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1392 x 1 x 1    82,128     False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 384 x 7 x 7     534,528    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 384 x 7 x 7     768        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 2304 x 7 x 7    884,736    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 2304 x 7 x 7    4,608      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 2304 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 2304 x 7 x 7    20,736     False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 2304 x 7 x 7    4,608      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 2304 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 2304 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 96 x 1 x 1      221,280    False     \n",
              "________________________________________________________________\n",
              "Swish                64 x 96 x 1 x 1      0          False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 2304 x 1 x 1    223,488    False     \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 384 x 7 x 7     884,736    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 384 x 7 x 7     768        True      \n",
              "________________________________________________________________\n",
              "Conv2d               64 x 1536 x 7 x 7    589,824    False     \n",
              "________________________________________________________________\n",
              "BatchNorm2d          64 x 1536 x 7 x 7    3,072      True      \n",
              "________________________________________________________________\n",
              "Swish                64 x 1536 x 7 x 7    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveAvgPool2d    64 x 1536 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "AdaptiveMaxPool2d    64 x 1536 x 1 x 1    0          False     \n",
              "________________________________________________________________\n",
              "Flatten              64 x 3072            0          False     \n",
              "________________________________________________________________\n",
              "BatchNorm1d          64 x 3072            6,144      True      \n",
              "________________________________________________________________\n",
              "Dropout              64 x 3072            0          False     \n",
              "________________________________________________________________\n",
              "Linear               64 x 512             1,572,864  True      \n",
              "________________________________________________________________\n",
              "ReLU                 64 x 512             0          False     \n",
              "________________________________________________________________\n",
              "BatchNorm1d          64 x 512             1,024      True      \n",
              "________________________________________________________________\n",
              "Dropout              64 x 512             0          False     \n",
              "________________________________________________________________\n",
              "Linear               64 x 37              18,944     True      \n",
              "________________________________________________________________\n",
              "\n",
              "Total params: 12,295,208\n",
              "Total trainable params: 1,686,272\n",
              "Total non-trainable params: 10,608,936\n",
              "\n",
              "Optimizer used: <function Adam at 0x7efc8c72c6a8>\n",
              "Loss function: LabelSmoothingCrossEntropy()\n",
              "\n",
              "Model frozen up to parameter group number 1\n",
              "\n",
              "Callbacks:\n",
              "  - TrainEvalCallback\n",
              "  - Recorder\n",
              "  - ProgressCallback"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GAUjWw6e5jCm",
        "colab_type": "text"
      },
      "source": [
        "That looks much better. Let's train!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6Sz46E5N5P4L",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        },
        "outputId": "37433ef8-98e5-4674-ce76-add64544ff0c"
      },
      "source": [
        "learn.lr_find()"
      ],
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              ""
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEKCAYAAADjDHn2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGGN8+v8vBpxDHpw0UgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5J2z-WOg5lHu",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "74c85859-2be4-4c5c-e3d0-9c7c2c6c9538"
      },
      "source": [
        "learn.fit_one_cycle(5, slice(3e-2))"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: left;\">\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",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>1.700641</td>\n",
              "      <td>1.801064</td>\n",
              "      <td>0.770636</td>\n",
              "      <td>01:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>1.277738</td>\n",
              "      <td>1.027878</td>\n",
              "      <td>0.895805</td>\n",
              "      <td>01:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2</td>\n",
              "      <td>1.046595</td>\n",
              "      <td>0.974152</td>\n",
              "      <td>0.914750</td>\n",
              "      <td>01:15</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3</td>\n",
              "      <td>0.935457</td>\n",
              "      <td>0.929803</td>\n",
              "      <td>0.916779</td>\n",
              "      <td>01:16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>4</td>\n",
              "      <td>0.870509</td>\n",
              "      <td>0.899026</td>\n",
              "      <td>0.932341</td>\n",
              "      <td>01:16</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9QKd7Tl96mxq",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "learn.save('stage_1')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3k7rTGv-6hOp",
        "colab_type": "text"
      },
      "source": [
        "Then we can unfreeze it and train a little more"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "F2fY4fHMBXYE",
        "colab_type": "text"
      },
      "source": [
        "## TODO: Unfreeze and train"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PkT9dpMQ6hL7",
        "colab_type": "text"
      },
      "source": [
        "## TODO: Use pretrained pkl file (transfer learn twice)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3BV-y7Qo577g",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}