U_Net.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "view-in-github"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Mostafa-wael/U-Net-in-PyTorch/blob/main/U_Net.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "h6nC3-Nk1dJi"
      },
      "source": [
        "# U-Net\n",
        "This is an implementation of the U-Net model from the paper, [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://papers.labml.ai/paper/1505.04597).\n",
        "\n",
        "U-Net is made up of a contracting path and an expanding path.\n",
        "The contracting route is a series of convolutional layers and pooling layers that gradually diminish the resolution of the feature map.\n",
        "The expansive route is a succession of up-sampling and convolutional layers that gradually enhance the resolution of the feature map.\n",
        "\n",
        "The relevant feature map from the contracting path is concatenated with the current feature map at each step in the expansive path."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OVmp0vMZcoI6"
      },
      "source": [
        "# Implementation 1\n",
        "Adapted from [here](https://nn.labml.ai/unet/index.html)."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 99,
      "metadata": {
        "id": "p12wLDC6118O"
      },
      "outputs": [],
      "source": [
        "import torch\n",
        "import torchvision.transforms.functional\n",
        "from torch import nn"
      ]
    },
    {
      "attachments": {},
      "cell_type": "markdown",
      "metadata": {
        "id": "UoVUg5mp2Scp"
      },
      "source": [
        "### Two $3 \\times 3$ Convolution Layers\n",
        "\n",
        "Each step in the contraction path and expansive path have two $3 \\times 3$\n",
        "convolutional layers followed by ReLU activations.\n",
        "\n",
        "In the U-Net paper they used $0$ padding,\n",
        "but we use $1$ padding so that final feature map is not cropped."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 100,
      "metadata": {
        "id": "LObQQ1Mc18P9"
      },
      "outputs": [],
      "source": [
        "class DoubleConvolution(nn.Module):\n",
        "    def __init__(self, in_channels: int, out_channels: int):\n",
        "        super().__init__()\n",
        "\n",
        "        # First 3x3 convolutional layer\n",
        "        self.first = nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1)\n",
        "        self.act1 = nn.ReLU()\n",
        "        # Second 3x3 convolutional layer\n",
        "        self.second = nn.Conv2d(out_channels, out_channels, kernel_size=3, padding=1)\n",
        "        self.act2 = nn.ReLU()\n",
        "\n",
        "    def forward(self, x: torch.Tensor):\n",
        "        # Apply the two convolution layers and activations\n",
        "        x = self.first(x)\n",
        "        x = self.act1(x)\n",
        "        x = self.second(x)\n",
        "        return self.act2(x)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XkmBUpbP3ez0"
      },
      "source": [
        "  ### Down-sample\n",
        "\n",
        "  Each step in the contracting path down-samples the feature map with\n",
        "  a $2 \\times 2$ max pooling layer."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 101,
      "metadata": {
        "id": "XcFc4j-v2WSq"
      },
      "outputs": [],
      "source": [
        "class DownSample(nn.Module):\n",
        "    def __init__(self):\n",
        "        super().__init__()\n",
        "        # Max pooling layer\n",
        "        self.pool = nn.MaxPool2d(2)\n",
        "\n",
        "    def forward(self, x: torch.Tensor):\n",
        "        return self.pool(x)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yPEUYvnhqMKu"
      },
      "source": [
        "  ### Up-sample\n",
        "\n",
        "  Each step in the expansive path up-samples the feature map with\n",
        "  a $2 \\times 2$ up-convolution."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 102,
      "metadata": {
        "id": "Hx9bPVtu2bzI"
      },
      "outputs": [],
      "source": [
        "class UpSample(nn.Module):\n",
        "    def __init__(self, in_channels: int, out_channels: int):\n",
        "        super().__init__()\n",
        "\n",
        "        # Up-convolution\n",
        "        self.up = nn.ConvTranspose2d(in_channels, out_channels, kernel_size=2, stride=2)\n",
        "\n",
        "    def forward(self, x: torch.Tensor):\n",
        "        return self.up(x)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iCusRi_AqhkJ"
      },
      "source": [
        "  ### Crop and Concatenate the feature map\n",
        "\n",
        "  At every step in the expansive path the corresponding feature map from the contracting path\n",
        "  concatenated with the current feature map.\n",
        "  "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 103,
      "metadata": {
        "id": "YOkAcqwtqcTA"
      },
      "outputs": [],
      "source": [
        "class CropAndConcat(nn.Module):\n",
        "    def forward(self, x: torch.Tensor, contracting_x: torch.Tensor):\n",
        "        # Crop the feature map from the contracting path to the size of the current feature map\n",
        "        contracting_x = torchvision.transforms.functional.center_crop(contracting_x, [x.shape[2], x.shape[3]])\n",
        "        # Concatenate the feature maps\n",
        "        x = torch.cat([x, contracting_x], dim=1)\n",
        "        return x"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GgclCYiNr85j"
      },
      "source": [
        "## U-Net\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 104,
      "metadata": {
        "id": "JDuoqE_6rwhD"
      },
      "outputs": [],
      "source": [
        "class UNet(nn.Module):\n",
        "    def __init__(self, out_channels: int):\n",
        "\n",
        "        super().__init__()\n",
        "        # Double convolution layers for the contracting path.\n",
        "        # The number of features gets doubled at each step starting from 64.\n",
        "        down_conv_sizes = [(3, 64), (64, 128), (128, 256), (256, 512)]\n",
        "        self.down_conv = nn.ModuleList([DoubleConvolution(i, o) for i, o in down_conv_sizes])\n",
        "        \n",
        "        # Down sampling layers for the contracting path\n",
        "        self.down_sample = nn.ModuleList([DownSample() for _ in range(4)])\n",
        "\n",
        "        # The two convolution layers at the lowest resolution (the bottom of the U).\n",
        "        self.middle_conv = DoubleConvolution(512, 1024)\n",
        "\n",
        "        # Up sampling layers for the expansive path.\n",
        "        # The number of features is halved with up-sampling.\n",
        "        upsample_sizes = [(1024, 512), (512, 256), (256, 128), (128, 64)]\n",
        "        \n",
        "        self.up_sample = nn.ModuleList([UpSample(i, o) for i, o in upsample_sizes])\n",
        "        # Double convolution layers for the expansive path.\n",
        "        # Their input is the concatenation of the current feature map and the feature map from the\n",
        "        # contracting path. Therefore, the number of input features is double the number of features\n",
        "        # from up-sampling.\n",
        "        up_conv_sizes = [(1024, 512), (512, 256), (256, 128), (128, 64)]\n",
        "        self.up_conv = nn.ModuleList([DoubleConvolution(i, o) for i, o in up_conv_sizes])\n",
        "        \n",
        "        # Crop and concatenate layers for the expansive path.\n",
        "        self.concat = nn.ModuleList([CropAndConcat() for _ in range(4)])\n",
        "        \n",
        "        # Final 1x1 convolution layer to produce the output\n",
        "        self.final_conv = nn.Conv2d(64, out_channels, kernel_size=1)\n",
        "\n",
        "    def forward(self, x: torch.Tensor):\n",
        "        # To collect the outputs of contracting path for later concatenation with the expansive path.\n",
        "        pass_through = []\n",
        "        # Contracting path\n",
        "        for i in range(len(self.down_conv)):\n",
        "            # Two 3x3 convolutional layers\n",
        "            x = self.down_conv[i](x)\n",
        "            # Collect the output\n",
        "            pass_through.append(x)\n",
        "            # Down-sample\n",
        "            x = self.down_sample[i](x)\n",
        "\n",
        "        # Two 3x3 convolutional layers at the bottom of the U-Net\n",
        "        x = self.middle_conv(x)\n",
        "\n",
        "        # Expansive path\n",
        "        for i in range(len(self.up_conv)):\n",
        "            # Up-sample\n",
        "            x = self.up_sample[i](x)\n",
        "            # Concatenate the output of the contracting path\n",
        "            x = self.concat[i](x, pass_through.pop())\n",
        "            # Two 3x3 convolutional layers\n",
        "            x = self.up_conv[i](x)\n",
        "\n",
        "        # Final 1x1 convolution layer\n",
        "        out = self.final_conv(x)\n",
        "\n",
        "        return out"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 105,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "nOMJDdYZCJzY",
        "outputId": "fb9183ee-ce8e-47e8-812d-69693201345e"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 0/59\n",
            "----------\n",
            "LR 0.0001\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/torch/optim/lr_scheduler.py:139: UserWarning: Detected call of `lr_scheduler.step()` before `optimizer.step()`. In PyTorch 1.1.0 and later, you should call them in the opposite order: `optimizer.step()` before `lr_scheduler.step()`.  Failure to do this will result in PyTorch skipping the first value of the learning rate schedule. See more details at https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate\n",
            "  warnings.warn(\"Detected call of `lr_scheduler.step()` before `optimizer.step()`. \"\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "train: bce: 0.662320, dice: 0.990881, loss: 0.826600\n",
            "val: bce: 0.647937, dice: 0.990587, loss: 0.819262\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 1/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.637467, dice: 0.990848, loss: 0.814157\n",
            "val: bce: 0.617508, dice: 0.990572, loss: 0.804040\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 2/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.600763, dice: 0.990839, loss: 0.795801\n",
            "val: bce: 0.567151, dice: 0.990566, loss: 0.778859\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 3/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.538302, dice: 0.990830, loss: 0.764566\n",
            "val: bce: 0.480114, dice: 0.990556, loss: 0.735335\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 4/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.426923, dice: 0.990837, loss: 0.708880\n",
            "val: bce: 0.312010, dice: 0.990716, loss: 0.651363\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 5/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.193359, dice: 0.992174, loss: 0.592766\n",
            "val: bce: 0.048589, dice: 0.997209, loss: 0.522899\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 6/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.079860, dice: 0.995238, loss: 0.537549\n",
            "val: bce: 0.067354, dice: 0.994900, loss: 0.531127\n",
            "0m 3s\n",
            "Epoch 7/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.049103, dice: 0.995212, loss: 0.522157\n",
            "val: bce: 0.033138, dice: 0.993467, loss: 0.513302\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 8/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.030533, dice: 0.990140, loss: 0.510336\n",
            "val: bce: 0.029462, dice: 0.981788, loss: 0.505625\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 9/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.029601, dice: 0.977688, loss: 0.503645\n",
            "val: bce: 0.029920, dice: 0.967517, loss: 0.498719\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 10/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.028557, dice: 0.963704, loss: 0.496131\n",
            "val: bce: 0.026072, dice: 0.950587, loss: 0.488329\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 11/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.024090, dice: 0.945197, loss: 0.484643\n",
            "val: bce: 0.021548, dice: 0.925430, loss: 0.473489\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 12/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.020365, dice: 0.918849, loss: 0.469607\n",
            "val: bce: 0.019167, dice: 0.893814, loss: 0.456490\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 13/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.018594, dice: 0.887963, loss: 0.453278\n",
            "val: bce: 0.018244, dice: 0.861435, loss: 0.439840\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 14/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.017911, dice: 0.856420, loss: 0.437165\n",
            "val: bce: 0.018216, dice: 0.832258, loss: 0.425237\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 15/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.017775, dice: 0.828933, loss: 0.423354\n",
            "val: bce: 0.018127, dice: 0.809797, loss: 0.413962\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 16/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.017444, dice: 0.807634, loss: 0.412539\n",
            "val: bce: 0.017759, dice: 0.792565, loss: 0.405162\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 17/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.017093, dice: 0.791220, loss: 0.404157\n",
            "val: bce: 0.017602, dice: 0.779690, loss: 0.398646\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 18/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.017042, dice: 0.778146, loss: 0.397594\n",
            "val: bce: 0.017781, dice: 0.767112, loss: 0.392447\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 19/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.017334, dice: 0.763925, loss: 0.390630\n",
            "val: bce: 0.018595, dice: 0.751872, loss: 0.385234\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 20/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.018226, dice: 0.748781, loss: 0.383503\n",
            "val: bce: 0.019718, dice: 0.738290, loss: 0.379004\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 21/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.019597, dice: 0.731618, loss: 0.375608\n",
            "val: bce: 0.021209, dice: 0.717719, loss: 0.369464\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 22/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.020094, dice: 0.707524, loss: 0.363809\n",
            "val: bce: 0.021075, dice: 0.692520, loss: 0.356797\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 23/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.020936, dice: 0.680818, loss: 0.350877\n",
            "val: bce: 0.024012, dice: 0.666725, loss: 0.345369\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 24/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.022187, dice: 0.652586, loss: 0.337386\n",
            "val: bce: 0.022777, dice: 0.644117, loss: 0.333447\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 25/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.021530, dice: 0.627207, loss: 0.324368\n",
            "val: bce: 0.024417, dice: 0.619078, loss: 0.321747\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 26/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.022091, dice: 0.603954, loss: 0.313022\n",
            "val: bce: 0.023080, dice: 0.600102, loss: 0.311591\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 27/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.021367, dice: 0.584787, loss: 0.303077\n",
            "val: bce: 0.022555, dice: 0.579388, loss: 0.300971\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 28/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.020901, dice: 0.566519, loss: 0.293710\n",
            "val: bce: 0.021829, dice: 0.563229, loss: 0.292529\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 29/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020724, dice: 0.556852, loss: 0.288788\n",
            "val: bce: 0.022183, dice: 0.561854, loss: 0.292018\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 30/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.021046, dice: 0.554800, loss: 0.287923\n",
            "val: bce: 0.022362, dice: 0.561453, loss: 0.291908\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 31/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020997, dice: 0.553013, loss: 0.287005\n",
            "val: bce: 0.022071, dice: 0.558772, loss: 0.290422\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 32/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020758, dice: 0.551686, loss: 0.286222\n",
            "val: bce: 0.021958, dice: 0.557287, loss: 0.289622\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 33/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020726, dice: 0.549969, loss: 0.285347\n",
            "val: bce: 0.021984, dice: 0.556337, loss: 0.289160\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 34/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020757, dice: 0.548380, loss: 0.284569\n",
            "val: bce: 0.021981, dice: 0.554266, loss: 0.288124\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 35/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020767, dice: 0.546605, loss: 0.283686\n",
            "val: bce: 0.022035, dice: 0.552651, loss: 0.287343\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 36/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020764, dice: 0.544836, loss: 0.282800\n",
            "val: bce: 0.021911, dice: 0.550968, loss: 0.286440\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 37/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020632, dice: 0.543130, loss: 0.281881\n",
            "val: bce: 0.021729, dice: 0.548917, loss: 0.285323\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 38/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020568, dice: 0.541273, loss: 0.280920\n",
            "val: bce: 0.021808, dice: 0.547272, loss: 0.284540\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 39/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020578, dice: 0.539238, loss: 0.279908\n",
            "val: bce: 0.021733, dice: 0.545498, loss: 0.283616\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 40/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020516, dice: 0.537223, loss: 0.278870\n",
            "val: bce: 0.021576, dice: 0.543391, loss: 0.282484\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 41/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020331, dice: 0.535226, loss: 0.277779\n",
            "val: bce: 0.021469, dice: 0.541415, loss: 0.281442\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 42/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020268, dice: 0.533044, loss: 0.276656\n",
            "val: bce: 0.021354, dice: 0.539425, loss: 0.280389\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 43/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020154, dice: 0.530931, loss: 0.275543\n",
            "val: bce: 0.021286, dice: 0.537243, loss: 0.279265\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 44/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020179, dice: 0.528468, loss: 0.274324\n",
            "val: bce: 0.021249, dice: 0.535134, loss: 0.278192\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 45/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019983, dice: 0.526190, loss: 0.273086\n",
            "val: bce: 0.021016, dice: 0.532550, loss: 0.276783\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 46/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019904, dice: 0.523427, loss: 0.271666\n",
            "val: bce: 0.021011, dice: 0.530188, loss: 0.275600\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 47/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019855, dice: 0.520439, loss: 0.270147\n",
            "val: bce: 0.020976, dice: 0.527636, loss: 0.274306\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 48/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019785, dice: 0.517248, loss: 0.268517\n",
            "val: bce: 0.020813, dice: 0.524562, loss: 0.272688\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 49/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019718, dice: 0.513950, loss: 0.266834\n",
            "val: bce: 0.020763, dice: 0.521428, loss: 0.271096\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 50/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019622, dice: 0.510338, loss: 0.264980\n",
            "val: bce: 0.021009, dice: 0.519034, loss: 0.270022\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 51/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019800, dice: 0.506630, loss: 0.263215\n",
            "val: bce: 0.020422, dice: 0.515593, loss: 0.268007\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 52/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019407, dice: 0.503172, loss: 0.261289\n",
            "val: bce: 0.020999, dice: 0.513502, loss: 0.267250\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 53/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019579, dice: 0.499222, loss: 0.259401\n",
            "val: bce: 0.020487, dice: 0.509055, loss: 0.264771\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 54/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019613, dice: 0.495746, loss: 0.257679\n",
            "val: bce: 0.020637, dice: 0.505755, loss: 0.263196\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 55/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019384, dice: 0.491465, loss: 0.255425\n",
            "val: bce: 0.021074, dice: 0.503822, loss: 0.262448\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 56/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019808, dice: 0.488256, loss: 0.254032\n",
            "val: bce: 0.020282, dice: 0.500821, loss: 0.260552\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 57/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019316, dice: 0.485296, loss: 0.252306\n",
            "val: bce: 0.021410, dice: 0.501713, loss: 0.261561\n",
            "0m 3s\n",
            "Epoch 58/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019779, dice: 0.482008, loss: 0.250894\n",
            "val: bce: 0.020344, dice: 0.496499, loss: 0.258421\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 59/59\n",
            "----------\n",
            "LR 1.0000000000000002e-06\n",
            "train: bce: 0.019188, dice: 0.481858, loss: 0.250523\n",
            "val: bce: 0.020509, dice: 0.495652, loss: 0.258081\n",
            "saving best model\n",
            "0m 3s\n",
            "Best val loss: 0.258081\n",
            "(3, 6, 192, 192)\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1200x1200 with 9 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "import test\n",
        "\n",
        "test.run(UNet)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "diJuIWYWcz2w"
      },
      "source": [
        "# Implementation 2\n",
        "More simple and strightforward "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 106,
      "metadata": {
        "id": "uB3wScRd7Bqp"
      },
      "outputs": [],
      "source": [
        "# First, the necessary modules are imported from the torch and torchvision packages, including the nn module for building neural networks and the pre-trained models provided in torchvision.models. \n",
        "# The relu function is also imported from torch.nn.functional.\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "from torchvision import models\n",
        "from torch.nn.functional import relu\n",
        "\n",
        "\n",
        "# Then, a custom class UNet is defined as a subclass of nn.Module. \n",
        "# The __init__ method initializes the architecture of the U-Net by defining the layers for both the encoder and decoder parts of the network. \n",
        "# The argument n_class specifies the number of classes for the segmentation task.\n",
        "class UNet(nn.Module):\n",
        "    def __init__(self, n_class):\n",
        "        super().__init__()\n",
        "        \n",
        "        # Encoder\n",
        "        # In the encoder, convolutional layers with the Conv2d function are used to extract features from the input image. \n",
        "        # Each block in the encoder consists of two convolutional layers followed by a max-pooling layer, with the exception of the last block which does not include a max-pooling layer.\n",
        "        # -------\n",
        "        # input: 572x572x3\n",
        "        self.e11 = nn.Conv2d(3, 64, kernel_size=3, padding=1) # output: 570x570x64\n",
        "        self.e12 = nn.Conv2d(64, 64, kernel_size=3, padding=1) # output: 568x568x64\n",
        "        self.pool1 = nn.MaxPool2d(kernel_size=2, stride=2) # output: 284x284x64\n",
        "\n",
        "        # input: 284x284x64\n",
        "        self.e21 = nn.Conv2d(64, 128, kernel_size=3, padding=1) # output: 282x282x128\n",
        "        self.e22 = nn.Conv2d(128, 128, kernel_size=3, padding=1) # output: 280x280x128\n",
        "        self.pool2 = nn.MaxPool2d(kernel_size=2, stride=2) # output: 140x140x128\n",
        "\n",
        "        # input: 140x140x128\n",
        "        self.e31 = nn.Conv2d(128, 256, kernel_size=3, padding=1) # output: 138x138x256\n",
        "        self.e32 = nn.Conv2d(256, 256, kernel_size=3, padding=1) # output: 136x136x256\n",
        "        self.pool3 = nn.MaxPool2d(kernel_size=2, stride=2) # output: 68x68x256\n",
        "\n",
        "        # input: 68x68x256\n",
        "        self.e41 = nn.Conv2d(256, 512, kernel_size=3, padding=1) # output: 66x66x512\n",
        "        self.e42 = nn.Conv2d(512, 512, kernel_size=3, padding=1) # output: 64x64x512\n",
        "        self.pool4 = nn.MaxPool2d(kernel_size=2, stride=2) # output: 32x32x512\n",
        "\n",
        "        # input: 32x32x512\n",
        "        self.e51 = nn.Conv2d(512, 1024, kernel_size=3, padding=1) # output: 30x30x1024\n",
        "        self.e52 = nn.Conv2d(1024, 1024, kernel_size=3, padding=1) # output: 28x28x1024\n",
        "\n",
        "\n",
        "        # Decoder\n",
        "        # In the decoder, transpose convolutional layers with the ConvTranspose2d function are used to upsample the feature maps to the original size of the input image. \n",
        "        # Each block in the decoder consists of an upsampling layer, a concatenation with the corresponding encoder feature map, and two convolutional layers.\n",
        "        # -------\n",
        "        self.upconv1 = nn.ConvTranspose2d(1024, 512, kernel_size=2, stride=2)\n",
        "        self.d11 = nn.Conv2d(1024, 512, kernel_size=3, padding=1)\n",
        "        self.d12 = nn.Conv2d(512, 512, kernel_size=3, padding=1)\n",
        "\n",
        "        self.upconv2 = nn.ConvTranspose2d(512, 256, kernel_size=2, stride=2)\n",
        "        self.d21 = nn.Conv2d(512, 256, kernel_size=3, padding=1)\n",
        "        self.d22 = nn.Conv2d(256, 256, kernel_size=3, padding=1)\n",
        "\n",
        "        self.upconv3 = nn.ConvTranspose2d(256, 128, kernel_size=2, stride=2)\n",
        "        self.d31 = nn.Conv2d(256, 128, kernel_size=3, padding=1)\n",
        "        self.d32 = nn.Conv2d(128, 128, kernel_size=3, padding=1)\n",
        "\n",
        "        self.upconv4 = nn.ConvTranspose2d(128, 64, kernel_size=2, stride=2)\n",
        "        self.d41 = nn.Conv2d(128, 64, kernel_size=3, padding=1)\n",
        "        self.d42 = nn.Conv2d(64, 64, kernel_size=3, padding=1)\n",
        "\n",
        "        # Output layer\n",
        "        self.outconv = nn.Conv2d(64, n_class, kernel_size=1)\n",
        "\n",
        "    def forward(self, x):\n",
        "        # Encoder\n",
        "        xe11 = relu(self.e11(x))\n",
        "        xe12 = relu(self.e12(xe11))\n",
        "        xp1 = self.pool1(xe12)\n",
        "\n",
        "        xe21 = relu(self.e21(xp1))\n",
        "        xe22 = relu(self.e22(xe21))\n",
        "        xp2 = self.pool2(xe22)\n",
        "\n",
        "        xe31 = relu(self.e31(xp2))\n",
        "        xe32 = relu(self.e32(xe31))\n",
        "        xp3 = self.pool3(xe32)\n",
        "\n",
        "        xe41 = relu(self.e41(xp3))\n",
        "        xe42 = relu(self.e42(xe41))\n",
        "        xp4 = self.pool4(xe42)\n",
        "\n",
        "        xe51 = relu(self.e51(xp4))\n",
        "        xe52 = relu(self.e52(xe51))\n",
        "        \n",
        "        # Decoder\n",
        "        xu1 = self.upconv1(xe52)\n",
        "        xu11 = torch.cat([xu1, xe42], dim=1)\n",
        "        xd11 = relu(self.d11(xu11))\n",
        "        xd12 = relu(self.d12(xd11))\n",
        "\n",
        "        xu2 = self.upconv2(xd12)\n",
        "        xu22 = torch.cat([xu2, xe32], dim=1)\n",
        "        xd21 = relu(self.d21(xu22))\n",
        "        xd22 = relu(self.d22(xd21))\n",
        "\n",
        "        xu3 = self.upconv3(xd22)\n",
        "        xu33 = torch.cat([xu3, xe22], dim=1)\n",
        "        xd31 = relu(self.d31(xu33))\n",
        "        xd32 = relu(self.d32(xd31))\n",
        "\n",
        "        xu4 = self.upconv4(xd32)\n",
        "        xu44 = torch.cat([xu4, xe12], dim=1)\n",
        "        xd41 = relu(self.d41(xu44))\n",
        "        xd42 = relu(self.d42(xd41))\n",
        "\n",
        "        # Output layer\n",
        "        out = self.outconv(xd42)\n",
        "\n",
        "        return out"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 107,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "wngRGB9yRldf",
        "outputId": "460ed05e-3450-4d10-e649-12a35d1d5bd3"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 0/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.698310, dice: 0.990453, loss: 0.844382\n",
            "val: bce: 0.691170, dice: 0.990315, loss: 0.840742\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 1/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.687076, dice: 0.990414, loss: 0.838745\n",
            "val: bce: 0.679788, dice: 0.990278, loss: 0.835033\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 2/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.673903, dice: 0.990376, loss: 0.832139\n",
            "val: bce: 0.661664, dice: 0.990225, loss: 0.825945\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 3/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.650573, dice: 0.990313, loss: 0.820443\n",
            "val: bce: 0.627470, dice: 0.990147, loss: 0.808808\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 4/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.604367, dice: 0.990247, loss: 0.797307\n",
            "val: bce: 0.551308, dice: 0.990115, loss: 0.770712\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 5/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.483832, dice: 0.990425, loss: 0.737128\n",
            "val: bce: 0.303975, dice: 0.991485, loss: 0.647730\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 6/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.182874, dice: 0.994276, loss: 0.588575\n",
            "val: bce: 0.149654, dice: 0.994618, loss: 0.572136\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 7/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.107580, dice: 0.995280, loss: 0.551430\n",
            "val: bce: 0.051828, dice: 0.996618, loss: 0.524223\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 8/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.046877, dice: 0.995627, loss: 0.521252\n",
            "val: bce: 0.045144, dice: 0.992514, loss: 0.518829\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 9/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.042860, dice: 0.991079, loss: 0.516969\n",
            "val: bce: 0.038446, dice: 0.989054, loss: 0.513750\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 10/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.035965, dice: 0.988257, loss: 0.512111\n",
            "val: bce: 0.033612, dice: 0.985485, loss: 0.509549\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 11/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.032430, dice: 0.982707, loss: 0.507568\n",
            "val: bce: 0.031496, dice: 0.975841, loss: 0.503669\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 12/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.030761, dice: 0.970909, loss: 0.500835\n",
            "val: bce: 0.029791, dice: 0.959907, loss: 0.494849\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 13/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.028150, dice: 0.951323, loss: 0.489736\n",
            "val: bce: 0.025908, dice: 0.931515, loss: 0.478712\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 14/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.024285, dice: 0.915911, loss: 0.470098\n",
            "val: bce: 0.022563, dice: 0.884283, loss: 0.453423\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 15/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.021403, dice: 0.867995, loss: 0.444699\n",
            "val: bce: 0.020308, dice: 0.842618, loss: 0.431463\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 16/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.019569, dice: 0.832432, loss: 0.426001\n",
            "val: bce: 0.019323, dice: 0.813802, loss: 0.416563\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 17/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.019287, dice: 0.804697, loss: 0.411992\n",
            "val: bce: 0.018969, dice: 0.788450, loss: 0.403710\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 18/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.018064, dice: 0.781493, loss: 0.399779\n",
            "val: bce: 0.017532, dice: 0.768390, loss: 0.392961\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 19/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.017455, dice: 0.763310, loss: 0.390382\n",
            "val: bce: 0.017909, dice: 0.751604, loss: 0.384757\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 20/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.017761, dice: 0.747443, loss: 0.382602\n",
            "val: bce: 0.017494, dice: 0.736608, loss: 0.377051\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 21/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.017806, dice: 0.731571, loss: 0.374688\n",
            "val: bce: 0.018396, dice: 0.716606, loss: 0.367501\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 22/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.018422, dice: 0.711303, loss: 0.364863\n",
            "val: bce: 0.018676, dice: 0.693448, loss: 0.356062\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 23/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.018915, dice: 0.688214, loss: 0.353565\n",
            "val: bce: 0.019296, dice: 0.668238, loss: 0.343767\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 24/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.019616, dice: 0.663084, loss: 0.341350\n",
            "val: bce: 0.020214, dice: 0.641933, loss: 0.331073\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 25/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.020721, dice: 0.637093, loss: 0.328907\n",
            "val: bce: 0.021911, dice: 0.617062, loss: 0.319487\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 26/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.022227, dice: 0.613042, loss: 0.317634\n",
            "val: bce: 0.023003, dice: 0.596659, loss: 0.309831\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 27/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.023707, dice: 0.594679, loss: 0.309193\n",
            "val: bce: 0.024857, dice: 0.584441, loss: 0.304649\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 28/59\n",
            "----------\n",
            "LR 0.0001\n",
            "train: bce: 0.023337, dice: 0.583221, loss: 0.303279\n",
            "val: bce: 0.023630, dice: 0.572596, loss: 0.298113\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 29/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.023198, dice: 0.575032, loss: 0.299115\n",
            "val: bce: 0.022744, dice: 0.571992, loss: 0.297368\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 30/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.022127, dice: 0.574703, loss: 0.298415\n",
            "val: bce: 0.022013, dice: 0.571872, loss: 0.296943\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 31/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.021788, dice: 0.573430, loss: 0.297609\n",
            "val: bce: 0.022135, dice: 0.570157, loss: 0.296146\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 32/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.021812, dice: 0.572112, loss: 0.296962\n",
            "val: bce: 0.021747, dice: 0.569313, loss: 0.295530\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 33/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.021222, dice: 0.571267, loss: 0.296244\n",
            "val: bce: 0.021060, dice: 0.569127, loss: 0.295094\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 34/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020705, dice: 0.570699, loss: 0.295702\n",
            "val: bce: 0.020966, dice: 0.567954, loss: 0.294460\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 35/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020775, dice: 0.569232, loss: 0.295004\n",
            "val: bce: 0.021124, dice: 0.566570, loss: 0.293847\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 36/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020774, dice: 0.568056, loss: 0.294415\n",
            "val: bce: 0.020748, dice: 0.565786, loss: 0.293267\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 37/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020196, dice: 0.567352, loss: 0.293774\n",
            "val: bce: 0.020308, dice: 0.565217, loss: 0.292762\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 38/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020144, dice: 0.566025, loss: 0.293084\n",
            "val: bce: 0.020545, dice: 0.563583, loss: 0.292064\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 39/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.020086, dice: 0.564665, loss: 0.292375\n",
            "val: bce: 0.020138, dice: 0.562860, loss: 0.291499\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 40/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019695, dice: 0.563707, loss: 0.291701\n",
            "val: bce: 0.019997, dice: 0.561718, loss: 0.290858\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 41/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019757, dice: 0.562256, loss: 0.291006\n",
            "val: bce: 0.019923, dice: 0.560600, loss: 0.290261\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 42/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019343, dice: 0.561111, loss: 0.290227\n",
            "val: bce: 0.019433, dice: 0.559838, loss: 0.289635\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 43/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019244, dice: 0.559736, loss: 0.289490\n",
            "val: bce: 0.019611, dice: 0.558127, loss: 0.288869\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 44/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.018956, dice: 0.558277, loss: 0.288616\n",
            "val: bce: 0.019119, dice: 0.557454, loss: 0.288286\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 45/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.019015, dice: 0.556734, loss: 0.287875\n",
            "val: bce: 0.019303, dice: 0.555761, loss: 0.287532\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 46/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.018571, dice: 0.555033, loss: 0.286802\n",
            "val: bce: 0.018622, dice: 0.554998, loss: 0.286810\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 47/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.018231, dice: 0.553460, loss: 0.285845\n",
            "val: bce: 0.019040, dice: 0.552749, loss: 0.285894\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 48/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.018362, dice: 0.551553, loss: 0.284957\n",
            "val: bce: 0.018285, dice: 0.551879, loss: 0.285082\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 49/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.018213, dice: 0.549226, loss: 0.283720\n",
            "val: bce: 0.018503, dice: 0.549720, loss: 0.284112\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 50/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.017794, dice: 0.547467, loss: 0.282631\n",
            "val: bce: 0.018461, dice: 0.547937, loss: 0.283199\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 51/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.017672, dice: 0.545303, loss: 0.281488\n",
            "val: bce: 0.017987, dice: 0.546466, loss: 0.282227\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 52/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.017591, dice: 0.542846, loss: 0.280219\n",
            "val: bce: 0.018077, dice: 0.544446, loss: 0.281261\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 53/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.017626, dice: 0.540604, loss: 0.279115\n",
            "val: bce: 0.017947, dice: 0.542779, loss: 0.280363\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 54/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.017322, dice: 0.538755, loss: 0.278039\n",
            "val: bce: 0.018530, dice: 0.540671, loss: 0.279600\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 55/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.017632, dice: 0.536573, loss: 0.277103\n",
            "val: bce: 0.017761, dice: 0.539426, loss: 0.278594\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 56/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.017768, dice: 0.533844, loss: 0.275806\n",
            "val: bce: 0.018287, dice: 0.536916, loss: 0.277601\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 57/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.017448, dice: 0.532097, loss: 0.274773\n",
            "val: bce: 0.018628, dice: 0.535046, loss: 0.276837\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 58/59\n",
            "----------\n",
            "LR 1e-05\n",
            "train: bce: 0.017869, dice: 0.529877, loss: 0.273873\n",
            "val: bce: 0.017645, dice: 0.535386, loss: 0.276516\n",
            "saving best model\n",
            "0m 3s\n",
            "Epoch 59/59\n",
            "----------\n",
            "LR 1.0000000000000002e-06\n",
            "train: bce: 0.017251, dice: 0.528870, loss: 0.273060\n",
            "val: bce: 0.018044, dice: 0.533895, loss: 0.275970\n",
            "saving best model\n",
            "0m 3s\n",
            "Best val loss: 0.275970\n",
            "(3, 6, 192, 192)\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1200x1200 with 9 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "import test\n",
        "\n",
        "test.run(UNet)"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "authorship_tag": "ABX9TyNdVlv6zI8B07irOX0V+OYo",
      "gpuType": "T4",
      "include_colab_link": true,
      "provenance": []
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}