solve.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Audio Classification\n",
    "\n",
    "## Preprocessing\n",
    "\n",
    "* All input audio files are processed to equal lengths using zero padding or cutting\n",
    "* Audio downsampled to 22050 from 41000 to reduce complexity\n",
    "* Mel Frequency Cepstral Coefficients (MFCC) used for collecting features\n",
    "* MFCC converted the audio classification task to an image classification task\n",
    "\n",
    "## Training procedure\n",
    "\n",
    "* Transfer learning using pretrained DenseNet for classifying the data. Resnet18, various CNN and RNN models were also tried out, but the results were unsatisfactory.\n",
    "* Optimizer used is Adam with learning rate = 1e-3.\n",
    "* Criterion used is Cross Entropy Loss\n",
    "* Scheduler used StepLR whose parameters are from the PyTorch docs\n",
    "* Training epochs: 20\n",
    "\n",
    "## Final results\n",
    "\n",
    "* Training & validation accuracy: 98% (similar due to small train test split)\n",
    "* Plots have been made towards the end of the report\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pathlib import Path\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import sklearn.metrics as metrics\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "from torch.utils.data import DataLoader, random_split\n",
    "from torch.utils.tensorboard import SummaryWriter\n",
    "from torchvision.datasets import DatasetFolder\n",
    "\n",
    "from models.densenet import DenseNet\n",
    "from utils import audio_loader, train, valid, predict\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Version: 1.9.0+cu111\n",
      "Device: cuda:0\n"
     ]
    }
   ],
   "source": [
    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
    "print(\"Version:\", torch.__version__)\n",
    "print(\"Device:\", device)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "RESULT_DIR = Path('./results')\n",
    "CHECKPOINT_DIR = Path('./checkpoints')\n",
    "RESULT_DIR.mkdir(exist_ok=True)\n",
    "CHECKPOINT_DIR.mkdir(exist_ok=True)\n",
    "\n",
    "# ? hyperparams\n",
    "BATCH_SIZE = 2\n",
    "TRAIN_TEST_SPLIT = 0.8\n",
    "EPOCHS = 20\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input shape: (1, 64, 442)\n",
      "Classes: ['1', '2', '3', '4', '5', '6', '7']\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Python39\\lib\\site-packages\\torchaudio\\functional\\functional.py:432: UserWarning: At least one mel filterbank has all zero values. The value for `n_mels` (128) may be set too high. Or, the value for `n_freqs` (201) may be set too low.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "dataset = DatasetFolder(root='./dataset', loader=audio_loader, extensions='wav')\n",
    "print(f\"Input shape: {tuple(dataset[0][0].shape)}\")\n",
    "\n",
    "classes = dataset.classes\n",
    "print(\"Classes:\", classes)\n",
    "\n",
    "train_size = int(TRAIN_TEST_SPLIT * len(dataset))\n",
    "valid_size = len(dataset) - train_size\n",
    "train_dataset, valid_dataset = random_split(dataset, [train_size, valid_size])\n",
    "\n",
    "# ? dataloader\n",
    "train_loader = DataLoader(train_dataset.dataset, batch_size=BATCH_SIZE, shuffle=True)\n",
    "valid_loader = DataLoader(valid_dataset.dataset, batch_size=BATCH_SIZE, shuffle=True)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model, optimizers and criterion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DenseNet(\n",
      "  (model): DenseNet(\n",
      "    (features): Sequential(\n",
      "      (conv0): Conv2d(1, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3))\n",
      "      (norm0): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      (relu0): ReLU(inplace=True)\n",
      "      (pool0): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
      "      (denseblock1): _DenseBlock(\n",
      "        (denselayer1): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(64, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer2): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(96, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(96, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer3): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(128, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer4): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(160, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(160, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer5): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(192, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer6): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(224, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(224, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "      )\n",
      "      (transition1): _Transition(\n",
      "        (norm): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "        (pool): AvgPool2d(kernel_size=2, stride=2, padding=0)\n",
      "      )\n",
      "      (denseblock2): _DenseBlock(\n",
      "        (denselayer1): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(128, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer2): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(160, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(160, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer3): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(192, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer4): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(224, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(224, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer5): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer6): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(288, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer7): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(320, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(320, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer8): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(352, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(352, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer9): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(384, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer10): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(416, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(416, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer11): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(448, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(448, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer12): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(480, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(480, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "      )\n",
      "      (transition2): _Transition(\n",
      "        (norm): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "        (pool): AvgPool2d(kernel_size=2, stride=2, padding=0)\n",
      "      )\n",
      "      (denseblock3): _DenseBlock(\n",
      "        (denselayer1): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer2): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(288, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(288, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer3): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(320, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(320, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer4): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(352, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(352, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer5): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(384, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(384, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer6): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(416, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(416, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer7): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(448, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(448, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer8): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(480, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(480, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer9): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer10): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(544, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(544, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer11): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(576, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(576, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer12): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(608, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(608, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer13): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(640, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(640, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer14): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(672, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(672, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer15): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(704, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(704, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer16): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(736, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(736, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer17): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(768, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(768, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer18): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(800, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(800, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer19): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(832, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(832, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer20): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(864, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(864, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer21): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(896, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer22): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(928, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(928, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer23): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(960, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(960, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer24): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(992, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(992, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer25): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1024, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer26): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1056, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1056, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer27): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1088, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1088, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer28): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1120, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1120, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer29): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1152, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1152, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer30): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1184, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1184, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer31): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1216, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1216, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer32): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1248, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1248, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer33): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1280, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1280, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer34): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1312, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1312, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer35): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1344, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1344, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer36): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1376, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1376, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer37): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1408, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1408, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer38): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1440, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1440, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer39): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1472, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1472, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer40): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1504, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1504, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer41): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1536, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1536, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer42): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1568, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1568, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer43): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1600, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1600, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer44): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1632, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1632, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer45): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1664, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1664, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer46): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1696, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1696, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer47): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1728, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1728, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer48): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1760, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1760, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "      )\n",
      "      (transition3): _Transition(\n",
      "        (norm): BatchNorm2d(1792, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv): Conv2d(1792, 896, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "        (pool): AvgPool2d(kernel_size=2, stride=2, padding=0)\n",
      "      )\n",
      "      (denseblock4): _DenseBlock(\n",
      "        (denselayer1): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(896, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(896, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer2): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(928, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(928, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer3): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(960, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(960, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer4): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(992, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(992, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer5): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1024, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer6): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1056, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1056, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer7): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1088, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1088, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer8): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1120, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1120, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer9): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1152, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1152, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer10): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1184, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1184, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer11): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1216, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1216, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer12): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1248, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1248, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer13): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1280, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1280, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer14): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1312, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1312, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer15): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1344, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1344, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer16): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1376, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1376, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer17): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1408, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1408, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer18): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1440, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1440, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer19): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1472, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1472, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer20): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1504, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1504, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer21): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1536, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1536, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer22): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1568, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1568, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer23): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1600, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1600, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer24): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1632, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1632, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer25): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1664, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1664, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer26): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1696, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1696, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer27): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1728, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1728, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer28): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1760, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1760, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer29): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1792, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1792, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer30): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1824, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1824, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer31): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1856, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1856, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "        (denselayer32): _DenseLayer(\n",
      "          (norm1): BatchNorm2d(1888, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu1): ReLU(inplace=True)\n",
      "          (conv1): Conv2d(1888, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n",
      "          (norm2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (relu2): ReLU(inplace=True)\n",
      "          (conv2): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        )\n",
      "      )\n",
      "      (norm5): BatchNorm2d(1920, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    )\n",
      "    (classifier): Linear(in_features=1920, out_features=7, bias=True)\n",
      "  )\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "model = DenseNet(len(classes))\n",
    "model.to(device)\n",
    "print(model)\n",
    "\n",
    "optimizer = optim.Adam(model.parameters(), lr=0.0001, weight_decay=1e-3)\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=EPOCHS//3, gamma=0.1)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training and validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Python39\\lib\\site-packages\\torchaudio\\functional\\functional.py:432: UserWarning: At least one mel filterbank has all zero values. The value for `n_mels` (128) may be set too high. Or, the value for `n_freqs` (201) may be set too low.\n",
      "  warnings.warn(\n",
      "C:\\Python39\\lib\\site-packages\\torch\\nn\\functional.py:718: UserWarning: Named tensors and all their associated APIs are an experimental feature and subject to change. Please do not use them for anything important until they are released as stable. (Triggered internally at  ..\\c10/core/TensorImpl.h:1156.)\n",
      "  return torch.max_pool2d(input, kernel_size, stride, padding, dilation, ceil_mode)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[Epoch: 1/20]\tTraining (Loss, Accuracy): (0.993889, 12.61%)\tValidation (Loss, Accuracy): (1.501447, 45.38%)\t\n",
      "[Epoch: 2/20]\tTraining (Loss, Accuracy): (0.859027, 34.45%)\tValidation (Loss, Accuracy): (1.183174, 66.39%)\t\n",
      "[Epoch: 3/20]\tTraining (Loss, Accuracy): (0.788107, 40.34%)\tValidation (Loss, Accuracy): (0.792361, 80.67%)\t\n",
      "[Epoch: 4/20]\tTraining (Loss, Accuracy): (0.656254, 57.98%)\tValidation (Loss, Accuracy): (0.624768, 84.87%)\t\n",
      "[Epoch: 5/20]\tTraining (Loss, Accuracy): (0.614129, 63.03%)\tValidation (Loss, Accuracy): (0.522614, 83.19%)\t\n",
      "[Epoch: 6/20]\tTraining (Loss, Accuracy): (0.505348, 74.79%)\tValidation (Loss, Accuracy): (0.255128, 96.64%)\t\n",
      "[Epoch: 7/20]\tTraining (Loss, Accuracy): (0.396607, 86.55%)\tValidation (Loss, Accuracy): (0.264759, 94.96%)\t\n",
      "[Epoch: 8/20]\tTraining (Loss, Accuracy): (0.363856, 90.76%)\tValidation (Loss, Accuracy): (0.231102, 94.96%)\t\n",
      "[Epoch: 9/20]\tTraining (Loss, Accuracy): (0.408310, 88.24%)\tValidation (Loss, Accuracy): (0.211544, 97.48%)\t\n",
      "[Epoch: 10/20]\tTraining (Loss, Accuracy): (0.385463, 85.71%)\tValidation (Loss, Accuracy): (0.218906, 92.44%)\t\n",
      "[Epoch: 11/20]\tTraining (Loss, Accuracy): (0.358045, 92.44%)\tValidation (Loss, Accuracy): (0.122465, 98.32%)\t\n",
      "[Epoch: 12/20]\tTraining (Loss, Accuracy): (0.416564, 82.35%)\tValidation (Loss, Accuracy): (0.125603, 97.48%)\t\n",
      "[Epoch: 13/20]\tTraining (Loss, Accuracy): (0.360460, 86.55%)\tValidation (Loss, Accuracy): (0.204327, 94.96%)\t\n",
      "[Epoch: 14/20]\tTraining (Loss, Accuracy): (0.354643, 87.39%)\tValidation (Loss, Accuracy): (0.142796, 95.80%)\t\n",
      "[Epoch: 15/20]\tTraining (Loss, Accuracy): (0.346821, 85.71%)\tValidation (Loss, Accuracy): (0.147830, 96.64%)\t\n",
      "[Epoch: 16/20]\tTraining (Loss, Accuracy): (0.362370, 89.08%)\tValidation (Loss, Accuracy): (0.136607, 98.32%)\t\n",
      "[Epoch: 17/20]\tTraining (Loss, Accuracy): (0.366617, 84.87%)\tValidation (Loss, Accuracy): (0.126234, 97.48%)\t\n",
      "[Epoch: 18/20]\tTraining (Loss, Accuracy): (0.331830, 87.39%)\tValidation (Loss, Accuracy): (0.105573, 99.16%)\t\n",
      "[Epoch: 19/20]\tTraining (Loss, Accuracy): (0.313768, 90.76%)\tValidation (Loss, Accuracy): (0.115187, 96.64%)\t\n",
      "[Epoch: 20/20]\tTraining (Loss, Accuracy): (0.380253, 84.87%)\tValidation (Loss, Accuracy): (0.112794, 97.48%)\t\n"
     ]
    }
   ],
   "source": [
    "train_losses, train_accuracies = [], []\n",
    "valid_losses, valid_accuracies = [], []\n",
    "\n",
    "writer = SummaryWriter(comment=\"audio_classification_densenet\")\n",
    "\n",
    "for epoch in range(1, EPOCHS + 1):\n",
    "    train_loss, train_accuracy = train(\n",
    "        model, train_loader, criterion=criterion, optimizer=optimizer, device=device\n",
    "    )\n",
    "    valid_loss, valid_accuracy = valid(\n",
    "        model, valid_loader, criterion=criterion, device=device\n",
    "    )\n",
    "\n",
    "    scheduler.step()\n",
    "\n",
    "    writer.add_scalar('training loss/train_loss', train_loss, epoch)\n",
    "    writer.add_scalar('validation loss/valid_loss', valid_loss, epoch)\n",
    "\n",
    "    writer.add_scalar('training accuracy/train_acc', train_accuracy, epoch)\n",
    "    writer.add_scalar('validation accuracy/valid_acc', valid_accuracy, epoch)\n",
    "\n",
    "    writer.add_scalar('learning rate/lr', optimizer.param_groups[0]['lr'], epoch)\n",
    "\n",
    "    train_losses.append(train_loss), train_accuracies.append(train_accuracy)\n",
    "    valid_losses.append(valid_loss), valid_accuracies.append(valid_accuracy)\n",
    "\n",
    "    print(\n",
    "        f'[Epoch: {epoch}/{EPOCHS}]\\t'\n",
    "        f'Training (Loss, Accuracy): ({train_loss:.6f}, {train_accuracy:.2f}%)\\t'\n",
    "        f'Validation (Loss, Accuracy): ({valid_loss:.6f}, {valid_accuracy:.2f}%)\\t'\n",
    "    )\n",
    "\n",
    "    if epoch % 5 == 0:\n",
    "        torch.save(model.state_dict(), CHECKPOINT_DIR / f'densenet_epoch_{epoch}.pth')\n",
    "\n",
    "writer.close()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1224x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1224x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(17, 5), constrained_layout=True)\n",
    "plt.plot(train_losses, label='Training Loss')\n",
    "plt.plot(valid_losses, label='Validation Loss')\n",
    "plt.legend()\n",
    "plt.xlabel('Epochs', fontsize=16)\n",
    "plt.ylabel('Loss', fontsize=16)\n",
    "plt.title('Loss vs Epochs', fontsize=20)\n",
    "plt.savefig(RESULT_DIR / 'train_loss_vs_epochs.png', facecolor=\"w\", transparent=False)\n",
    "plt.show()\n",
    "\n",
    "plt.figure(figsize=(17, 5), constrained_layout=True)\n",
    "plt.plot(train_accuracies, label='Training Accuracy')\n",
    "plt.plot(valid_accuracies, label='Validation Accuracy')\n",
    "plt.legend()\n",
    "plt.xlabel('Epochs', fontsize=16)\n",
    "plt.ylabel('Accuracy', fontsize=16)\n",
    "plt.title('Accuracy vs Epochs', fontsize=20)\n",
    "plt.savefig(RESULT_DIR / 'train_accuracy_vs_epochs.png', facecolor=\"w\", transparent=False)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Test Dataset\n",
    "\n",
    "* Load model\n",
    "* Classify test data\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "classes = ['1', '2', '3', '4', '5', '6', '7']\n",
    "MODEL_PATH = CHECKPOINT_DIR / 'densenet_epoch_20.pth'\n",
    "\n",
    "def make_predictions(audio_paths, labels):\n",
    "    model = DenseNet(len(classes)).to(device)\n",
    "    model.load_state_dict(torch.load(MODEL_PATH))\n",
    "\n",
    "    predictions = []\n",
    "\n",
    "    for audio_path, label in zip(audio_paths, labels):\n",
    "        input_tensor = audio_loader(audio_path)\n",
    "        predicted, accuracy = predict(model, input_tensor, classes, device)\n",
    "\n",
    "        print(f\"label: {label}, predicted: {predicted} ({accuracy:.2f} %)\")\n",
    "        predictions.append(predicted)\n",
    "\n",
    "    accuracy = metrics.accuracy_score(labels, predictions) * 100\n",
    "\n",
    "    precision = metrics.precision_score(labels, predictions, average='weighted', zero_division=0)\n",
    "    recall = metrics.recall_score(labels, predictions, average='weighted', zero_division=0)\n",
    "    f1 = metrics.f1_score(labels, predictions, average='weighted', zero_division=0)\n",
    "\n",
    "    return accuracy, precision, recall, f1\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Example 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Python39\\lib\\site-packages\\torchaudio\\functional\\functional.py:432: UserWarning: At least one mel filterbank has all zero values. The value for `n_mels` (128) may be set too high. Or, the value for `n_freqs` (201) may be set too low.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "label: 2, predicted: 2 (99.31 %)\n",
      "label: 6, predicted: 6 (76.93 %)\n",
      "Accuracy: 100.000, Precision: 1.000, Recall: 1.000, F1: 1.000\n"
     ]
    }
   ],
   "source": [
    "audio_paths = ['./dataset/2/430816.wav', './dataset/6/112564.wav']\n",
    "labels = ['2', '6']\n",
    "\n",
    "accuracy, precision, recall, f1 = make_predictions(audio_paths, labels)\n",
    "print(f\"Accuracy: {accuracy:.3f}, Precision: {precision:.3f}, Recall: {recall:.3f}, F1: {f1:.3f}\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Example 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Python39\\lib\\site-packages\\torchaudio\\functional\\functional.py:432: UserWarning: At least one mel filterbank has all zero values. The value for `n_mels` (128) may be set too high. Or, the value for `n_freqs` (201) may be set too low.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "label: 1, predicted: 1 (99.98 %)\n",
      "label: 1, predicted: 1 (100.00 %)\n",
      "label: 1, predicted: 1 (99.97 %)\n",
      "label: 1, predicted: 1 (99.59 %)\n",
      "label: 1, predicted: 1 (99.94 %)\n",
      "label: 1, predicted: 1 (99.21 %)\n",
      "label: 1, predicted: 1 (99.97 %)\n",
      "label: 1, predicted: 1 (98.21 %)\n",
      "label: 1, predicted: 1 (99.73 %)\n",
      "label: 1, predicted: 1 (99.62 %)\n",
      "label: 1, predicted: 1 (100.00 %)\n",
      "label: 1, predicted: 1 (98.52 %)\n",
      "label: 1, predicted: 1 (99.97 %)\n",
      "label: 1, predicted: 1 (99.99 %)\n",
      "label: 1, predicted: 1 (99.97 %)\n",
      "label: 1, predicted: 1 (99.79 %)\n",
      "label: 1, predicted: 1 (99.93 %)\n",
      "label: 2, predicted: 2 (44.87 %)\n",
      "label: 2, predicted: 2 (75.67 %)\n",
      "label: 2, predicted: 2 (85.10 %)\n",
      "label: 2, predicted: 2 (96.15 %)\n",
      "label: 2, predicted: 2 (92.49 %)\n",
      "label: 2, predicted: 2 (84.43 %)\n",
      "label: 2, predicted: 2 (93.05 %)\n",
      "label: 2, predicted: 2 (81.09 %)\n",
      "label: 2, predicted: 2 (73.45 %)\n",
      "label: 2, predicted: 2 (96.26 %)\n",
      "label: 2, predicted: 2 (91.72 %)\n",
      "label: 2, predicted: 2 (81.60 %)\n",
      "label: 2, predicted: 5 (73.23 %)\n",
      "label: 2, predicted: 2 (88.96 %)\n",
      "label: 2, predicted: 2 (99.31 %)\n",
      "label: 2, predicted: 2 (94.40 %)\n",
      "label: 2, predicted: 2 (93.21 %)\n",
      "label: 3, predicted: 5 (55.53 %)\n",
      "label: 3, predicted: 3 (98.65 %)\n",
      "label: 3, predicted: 3 (94.24 %)\n",
      "label: 3, predicted: 3 (89.25 %)\n",
      "label: 3, predicted: 3 (98.46 %)\n",
      "label: 3, predicted: 3 (98.28 %)\n",
      "label: 3, predicted: 3 (99.31 %)\n",
      "label: 3, predicted: 3 (66.92 %)\n",
      "label: 3, predicted: 3 (84.20 %)\n",
      "label: 3, predicted: 3 (74.71 %)\n",
      "label: 3, predicted: 3 (99.44 %)\n",
      "label: 3, predicted: 3 (98.74 %)\n",
      "label: 3, predicted: 3 (98.41 %)\n",
      "label: 3, predicted: 3 (92.75 %)\n",
      "label: 3, predicted: 3 (99.16 %)\n",
      "label: 3, predicted: 3 (87.51 %)\n",
      "label: 3, predicted: 3 (98.99 %)\n",
      "label: 4, predicted: 4 (83.89 %)\n",
      "label: 4, predicted: 4 (82.06 %)\n",
      "label: 4, predicted: 4 (89.24 %)\n",
      "label: 4, predicted: 4 (90.22 %)\n",
      "label: 4, predicted: 4 (71.74 %)\n",
      "label: 4, predicted: 4 (79.78 %)\n",
      "label: 4, predicted: 4 (89.85 %)\n",
      "label: 4, predicted: 4 (91.06 %)\n",
      "label: 4, predicted: 4 (85.59 %)\n",
      "label: 4, predicted: 4 (98.43 %)\n",
      "label: 4, predicted: 4 (80.03 %)\n",
      "label: 4, predicted: 4 (97.47 %)\n",
      "label: 4, predicted: 4 (99.02 %)\n",
      "label: 4, predicted: 4 (83.95 %)\n",
      "label: 4, predicted: 4 (97.62 %)\n",
      "label: 4, predicted: 4 (60.56 %)\n",
      "label: 4, predicted: 4 (99.36 %)\n",
      "label: 5, predicted: 5 (94.44 %)\n",
      "label: 5, predicted: 5 (99.93 %)\n",
      "label: 5, predicted: 5 (97.64 %)\n",
      "label: 5, predicted: 5 (97.94 %)\n",
      "label: 5, predicted: 5 (97.99 %)\n",
      "label: 5, predicted: 5 (97.62 %)\n",
      "label: 5, predicted: 5 (98.27 %)\n",
      "label: 5, predicted: 5 (99.29 %)\n",
      "label: 5, predicted: 5 (97.35 %)\n",
      "label: 5, predicted: 5 (98.01 %)\n",
      "label: 5, predicted: 5 (95.75 %)\n",
      "label: 5, predicted: 5 (98.06 %)\n",
      "label: 5, predicted: 5 (99.93 %)\n",
      "label: 5, predicted: 5 (99.56 %)\n",
      "label: 5, predicted: 5 (99.33 %)\n",
      "label: 5, predicted: 5 (97.41 %)\n",
      "label: 5, predicted: 5 (99.30 %)\n",
      "label: 6, predicted: 6 (76.93 %)\n",
      "label: 6, predicted: 6 (53.70 %)\n",
      "label: 6, predicted: 6 (93.67 %)\n",
      "label: 6, predicted: 6 (99.36 %)\n",
      "label: 6, predicted: 6 (61.12 %)\n",
      "label: 6, predicted: 6 (44.24 %)\n",
      "label: 6, predicted: 6 (97.19 %)\n",
      "label: 6, predicted: 6 (84.34 %)\n",
      "label: 6, predicted: 6 (99.23 %)\n",
      "label: 6, predicted: 6 (93.38 %)\n",
      "label: 6, predicted: 6 (99.90 %)\n",
      "label: 6, predicted: 6 (99.10 %)\n",
      "label: 6, predicted: 6 (96.21 %)\n",
      "label: 6, predicted: 6 (98.74 %)\n",
      "label: 6, predicted: 6 (90.92 %)\n",
      "label: 6, predicted: 6 (89.62 %)\n",
      "label: 6, predicted: 6 (97.82 %)\n",
      "label: 7, predicted: 7 (94.71 %)\n",
      "label: 7, predicted: 7 (99.69 %)\n",
      "label: 7, predicted: 7 (99.26 %)\n",
      "label: 7, predicted: 7 (98.40 %)\n",
      "label: 7, predicted: 7 (99.02 %)\n",
      "label: 7, predicted: 5 (43.84 %)\n",
      "label: 7, predicted: 7 (99.09 %)\n",
      "label: 7, predicted: 7 (97.95 %)\n",
      "label: 7, predicted: 7 (99.77 %)\n",
      "label: 7, predicted: 7 (95.72 %)\n",
      "label: 7, predicted: 7 (97.35 %)\n",
      "label: 7, predicted: 7 (99.19 %)\n",
      "label: 7, predicted: 7 (97.90 %)\n",
      "label: 7, predicted: 7 (95.99 %)\n",
      "label: 7, predicted: 7 (95.86 %)\n",
      "label: 7, predicted: 7 (97.68 %)\n",
      "label: 7, predicted: 7 (90.81 %)\n",
      "Accuracy: 97.479, Precision: 0.979, Recall: 0.975, F1: 0.975\n"
     ]
    }
   ],
   "source": [
    "audio_paths = list(Path('./dataset').glob('**/*.wav'))\n",
    "labels = [path.parts[-2] for path in audio_paths]\n",
    "\n",
    "accuracy, precision, recall, f1 = make_predictions(audio_paths, labels)\n",
    "print(f\"Accuracy: {accuracy:.3f}, Precision: {precision:.3f}, Recall: {recall:.3f}, F1: {f1:.3f}\")\n"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "63fd5069d213b44bf678585dea6b12cceca9941eaf7f819626cde1f2670de90d"
  },
  "kernelspec": {
   "display_name": "Python 3.9.6 64-bit",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}