train_taho_deepfake_classification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 764,
     "status": "ok",
     "timestamp": 1710261740054,
     "user": {
      "displayName": "김성호",
      "userId": "00239474051960738712"
     },
     "user_tz": -540
    },
    "id": "ZREHZZy-3-YU",
    "outputId": "66eb5f4d-7ba9-4191-ad1d-de60d8a23c01"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/hyoju/바탕화면/RealForensics/data/Forensics\n"
     ]
    }
   ],
   "source": [
    "%cd /home/hyoju/바탕화면/RealForensics/data/Forensics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 75655,
     "status": "ok",
     "timestamp": 1710261815708,
     "user": {
      "displayName": "김성호",
      "userId": "00239474051960738712"
     },
     "user_tz": -540
    },
    "id": "w_EyEKDl3Mki",
    "outputId": "346ba824-cb7b-43a2-ab5e-0c84e93767b6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded 720 items from ./csv_files/train_fake.csv\n",
      "Loaded 720 items from ./csv_files/train_real.csv\n",
      "Loaded 140 items from ./csv_files/val_fake.csv\n",
      "Loaded 140 items from ./csv_files/val_real.csv\n",
      "Loaded 140 items from ./csv_files/test_fake.csv\n",
      "Loaded 140 items from ./csv_files/test_real.csv\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import os\n",
    "import cv2\n",
    "import numpy as np\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from torchvision.transforms import Compose, Resize, ToTensor\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def load_dataset(file_path, root_dir, label_dirs, frames_limit):\n",
    "    df = pd.read_csv(file_path, delim_whitespace=True, names=['filename', 'label'])\n",
    "    dataset = []\n",
    "    for idx, row in df.iterrows():\n",
    "        video_name = row['filename']\n",
    "        label = row['label']\n",
    "        file_loaded = False  # 파일 로딩 성공 여부를 추적하는 플래그\n",
    "        for label_dir in label_dirs:\n",
    "            full_path = os.path.join(root_dir, label_dir, 'c23', 'cropped_faces', video_name)\n",
    "            if os.path.exists(full_path):\n",
    "                dataset.append((full_path, int(label), frames_limit))\n",
    "                file_loaded = True\n",
    "                break\n",
    "        if not file_loaded:\n",
    "            print(f\"Warning: {video_name} not found in any specified directories.\")\n",
    "    print(f\"Loaded {len(dataset)} items from {file_path}\")\n",
    "    return dataset\n",
    "\n",
    "# 'root_dir'를 현재 작업 디렉토리(.)로 설정\n",
    "root_dir = '.'\n",
    "csv_dir = os.path.join(root_dir, 'csv_files')\n",
    "label_dirs = ['Deepfakes', 'Face2Face', 'FaceShifter', 'FaceSwap', 'NeuralTextures', 'Real']\n",
    "\n",
    "frames_limit=90\n",
    "\n",
    "train_label_dirs = ['Deepfakes', 'Face2Face', 'FaceSwap', 'NeuralTextures', 'Real']  # FaceShifter 제외\n",
    "train_dataset = sum([load_dataset(os.path.join(csv_dir, f'train_{label}.csv'), root_dir, train_label_dirs, frames_limit=frames_limit) for label in ['fake', 'real']], [])\n",
    "\n",
    "val_label_dirs = ['FaceShifter', 'Real']  # FaceShifter와 Real만 포함\n",
    "val_dataset = load_dataset(os.path.join(csv_dir, 'val_fake.csv'), root_dir, val_label_dirs, frames_limit=frames_limit) + \\\n",
    "               load_dataset(os.path.join(csv_dir, 'val_real.csv'), root_dir, val_label_dirs, frames_limit=frames_limit)\n",
    "\n",
    "test_label_dirs = ['FaceShifter', 'Real']  # FaceShifter와 Real만 포함\n",
    "test_dataset = load_dataset(os.path.join(csv_dir, 'test_fake.csv'), root_dir, test_label_dirs, frames_limit=frames_limit) + \\\n",
    "               load_dataset(os.path.join(csv_dir, 'test_real.csv'), root_dir, test_label_dirs, frames_limit=frames_limit)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "executionInfo": {
     "elapsed": 2,
     "status": "ok",
     "timestamp": 1710262105491,
     "user": {
      "displayName": "김성호",
      "userId": "00239474051960738712"
     },
     "user_tz": -540
    },
    "id": "zua1h2in3eVI"
   },
   "outputs": [],
   "source": [
    "from torchvision.transforms import RandomCrop, Resize, RandomHorizontalFlip, RandomGrayscale, ToTensor, Compose\n",
    "from torch.utils.data import DataLoader, Dataset, WeightedRandomSampler\n",
    "import cv2\n",
    "import numpy as np\n",
    "import os\n",
    "import pandas as pd\n",
    "from PIL import Image\n",
    "\n",
    "class VideoDataset(Dataset):\n",
    "    def __init__(self, dataset, transform=None):\n",
    "        self.dataset = dataset\n",
    "        self.transform = transform\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.dataset)\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        video_path, label, frames_limit = self.dataset[idx]\n",
    "        cap = cv2.VideoCapture(video_path)\n",
    "        frames = []\n",
    "        frame_count = 0\n",
    "        while True:\n",
    "            ret, frame = cap.read()\n",
    "            if not ret or frame_count == frames_limit:\n",
    "                break\n",
    "            if self.transform:\n",
    "                # frame을 PIL Image로 변환 후 transform\n",
    "                frame = Image.fromarray(frame)\n",
    "                frame = self.transform(frame)\n",
    "            frames.append(np.array(frame))  # transform 후 다시 numpy 배열로 변환\n",
    "            frame_count += 1\n",
    "        cap.release()\n",
    "\n",
    "        # 프레임 수가 frames_limit보다 적은 경우, 마지막 프레임을 반복하여 추가\n",
    "        while len(frames) < frames_limit:\n",
    "            frames.append(frames[-1])\n",
    "\n",
    "        # 첫 frames_limit 개수의 프레임만 사용\n",
    "        frames = frames[:frames_limit]\n",
    "        frames = np.stack(frames)\n",
    "        #print(f\"Before transpose frames shape: {frames.shape}\")  # (60, 3, 112, 112)\n",
    "        frames = np.transpose(frames, (1, 0, 2, 3))  # 채널 차원을 맞추기 위한 조정 (C, D, H, W)\n",
    "        #print(f\"Final frames shape: {frames.shape}\")  # (60, 3, 112, 112)\n",
    "        return frames, label\n",
    "\n",
    "# 새로운 변환 정의\n",
    "transform = Compose([\n",
    "    RandomCrop(140),\n",
    "    Resize((112, 112)),\n",
    "    #Resize((28, 28)),\n",
    "    RandomHorizontalFlip(p=0.5),\n",
    "    RandomGrayscale(p=0.5),\n",
    "    ToTensor(),\n",
    "])\n",
    "\n",
    "def make_sampler(dataset, num_samples=144, fake_weight=16, real_weight=128):\n",
    "    targets = [label for _, label, _ in dataset] # dataset은 (video_path, label, frames_limit) 튜플의 리스트\n",
    "    weights = [fake_weight if label == 1 else real_weight for label in targets]\n",
    "    sampler = WeightedRandomSampler(weights, num_samples=num_samples, replacement=True)\n",
    "    return sampler\n",
    "\n",
    "batch_size = 32\n",
    "\n",
    "\n",
    "train_sampler = make_sampler(train_dataset)\n",
    "val_sampler = make_sampler(val_dataset, num_samples=len(val_dataset))\n",
    "test_sampler = make_sampler(test_dataset, num_samples=len(test_dataset))\n",
    "\n",
    "train_loader = DataLoader(VideoDataset(train_dataset, transform=transform), batch_size=batch_size, sampler=train_sampler)\n",
    "val_loader = DataLoader(VideoDataset(val_dataset, transform=transform), batch_size=len(val_dataset), sampler=val_sampler)\n",
    "test_loader = DataLoader(VideoDataset(test_dataset, transform=transform), batch_size=len(test_dataset), sampler=test_sampler)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "id": "wcrk6xRF34De"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'\\nimport torch\\nimport torch.nn as nn\\nimport torch.nn.functional as F\\n\\n# 모델 정의\\nclass Simple3DCNN(nn.Module):\\n    def __init__(self):\\n        super(Simple3DCNN, self).__init__()\\n        self.conv1 = nn.Conv3d(3, 8, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\\n        self.pool = nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2))\\n        self.conv2 = nn.Conv3d(8, 16, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\\n        self.fc1 = nn.Linear(16*15*28*28, 512)\\n        self.fc2 = nn.Linear(512, 2)\\n\\n    def forward(self, x):\\n        #print(f\"Initial input shape: {x.shape}\") # 4, 3, 60, 112, 112\\n        x = self.pool(F.relu(self.conv1(x)))\\n        #print(f\"After conv1 shape: {x.shape}\") # 4, 8, 30, 56, 56\\n        x = self.pool(F.relu(self.conv2(x)))\\n        #print(f\"After conv2 shape: {x.shape}\") # 4, 16, 15, 28, 28\\n        x = x.view(x.size(0), -1)  # x.size(0)은 배치 크기, Flatten the tensor\\n        #print(f\"Before FC layers shape: {x.shape}\") # 4, 16*15*28*28\\n        x = F.relu(self.fc1(x))\\n        #print(f\"After fc1 shape: {x.shape}\")\\n        x = self.fc2(x)\\n        #print(f\"Output shape: {x.shape}\")\\n        return x\\n'"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "# 모델 정의\n",
    "class Simple3DCNN(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(Simple3DCNN, self).__init__()\n",
    "        self.conv1 = nn.Conv3d(3, 8, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\n",
    "        self.pool = nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2))\n",
    "        self.conv2 = nn.Conv3d(8, 16, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\n",
    "        self.fc1 = nn.Linear(16*15*28*28, 512)\n",
    "        self.fc2 = nn.Linear(512, 2)\n",
    "\n",
    "    def forward(self, x):\n",
    "        #print(f\"Initial input shape: {x.shape}\") # 4, 3, 60, 112, 112\n",
    "        x = self.pool(F.relu(self.conv1(x)))\n",
    "        #print(f\"After conv1 shape: {x.shape}\") # 4, 8, 30, 56, 56\n",
    "        x = self.pool(F.relu(self.conv2(x)))\n",
    "        #print(f\"After conv2 shape: {x.shape}\") # 4, 16, 15, 28, 28\n",
    "        x = x.view(x.size(0), -1)  # x.size(0)은 배치 크기, Flatten the tensor\n",
    "        #print(f\"Before FC layers shape: {x.shape}\") # 4, 16*15*28*28\n",
    "        x = F.relu(self.fc1(x))\n",
    "        #print(f\"After fc1 shape: {x.shape}\")\n",
    "        x = self.fc2(x)\n",
    "        #print(f\"Output shape: {x.shape}\")\n",
    "        return x\n",
    "'''\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "executionInfo": {
     "elapsed": 379,
     "status": "ok",
     "timestamp": 1710262112053,
     "user": {
      "displayName": "김성호",
      "userId": "00239474051960738712"
     },
     "user_tz": -540
    },
    "id": "LYsFcA7JtZrb"
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "class ConvLSTMCell(nn.Module):\n",
    "    def __init__(self, input_dim, hidden_dim, kernel_size, bias):\n",
    "        \"\"\"\n",
    "        Initialize ConvLSTM cell.\n",
    "\n",
    "        Parameters:\n",
    "            input_dim (int) - Number of channels of input tensor.\n",
    "            hidden_dim (int) - Number of channels of hidden state.\n",
    "            kernel_size (int or (int, int)) - Size of the convolutional kernel.\n",
    "            bias (bool) - Whether or not to add the bias.\n",
    "        \"\"\"\n",
    "        super(ConvLSTMCell, self).__init__()\n",
    "\n",
    "        self.input_dim = input_dim\n",
    "        self.hidden_dim = hidden_dim\n",
    "\n",
    "        self.kernel_size = kernel_size\n",
    "        self.padding = (kernel_size[0] // 2, kernel_size[1] // 2)\n",
    "        self.bias = bias\n",
    "\n",
    "        self.conv = nn.Conv2d(in_channels=self.input_dim + self.hidden_dim,\n",
    "                              out_channels=4 * self.hidden_dim,  # for input_gate, forget_gate, current_state, output_gate\n",
    "                              kernel_size=self.kernel_size,\n",
    "                              padding=self.padding,\n",
    "                              bias=self.bias)\n",
    "\n",
    "    def forward(self, input_tensor, cur_state):\n",
    "        h_cur, c_cur = cur_state\n",
    "\n",
    "        combined = torch.cat([input_tensor, h_cur], dim=1)  # concatenate along channel axis\n",
    "\n",
    "        combined_conv = self.conv(combined)\n",
    "        cc_i, cc_f, cc_o, cc_g = torch.split(combined_conv, self.hidden_dim, dim=1)\n",
    "        i = torch.sigmoid(cc_i)\n",
    "        f = torch.sigmoid(cc_f)\n",
    "        o = torch.sigmoid(cc_o)\n",
    "        g = torch.tanh(cc_g)\n",
    "\n",
    "        c_next = f * c_cur + i * g\n",
    "        h_next = o * torch.tanh(c_next)\n",
    "\n",
    "        return h_next, c_next\n",
    "\n",
    "    def init_hidden(self, batch_size, image_size):\n",
    "        height, width = image_size\n",
    "        return (torch.zeros(batch_size, self.hidden_dim, height, width, device=self.conv.weight.device),\n",
    "                torch.zeros(batch_size, self.hidden_dim, height, width, device=self.conv.weight.device))\n",
    "\n",
    "\n",
    "class SimpleConvLSTMModel(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(SimpleConvLSTMModel, self).__init__()\n",
    "\n",
    "        # ConvLSTM 셀\n",
    "        self.convLSTM1 = ConvLSTMCell(input_dim=3, hidden_dim=32, kernel_size=(3, 3), bias=True)\n",
    "        self.convLSTM2 = ConvLSTMCell(input_dim=32, hidden_dim=64, kernel_size=(3, 3), bias=True)\n",
    "\n",
    "        # average pooling으로 시간적 차원 축소\n",
    "        self.avgPool = nn.AdaptiveAvgPool3d((1, None, None))\n",
    "\n",
    "        # fully connected layer\n",
    "        self.fc1 = nn.Linear(3*60*112*112, 512)\n",
    "        self.fc2 = nn.Linear(512, 2)\n",
    "\n",
    "\n",
    "    def forward(self, x):\n",
    "        # x: (batch_size, channels, depth, height, width)\n",
    "        batch_size, _, depth, height, width = x.size()\n",
    "        #print(f\"Initial input shape: {x.shape}\")\n",
    "\n",
    "        # 초기 상태\n",
    "        h1, c1 = self.convLSTM1.init_hidden(batch_size, (height, width))\n",
    "        #print(f\"After conv1 shape: {x.shape}\")\n",
    "        h2, c2 = self.convLSTM2.init_hidden(batch_size, (height, width))\n",
    "        #print(f\"After conv2 shape: {x.shape}\")\n",
    "\n",
    "        # 모든 시간 스텝을 순회\n",
    "        for t in range(depth):\n",
    "            # 현재 시간 스텝의 프레임을 가져옴\n",
    "            current_frame = x[:, :, t, :, :]\n",
    "            # ConvLSTM 셀을 통과\n",
    "            h1, c1 = self.convLSTM1(current_frame, (h1, c1))\n",
    "            h2, c2 = self.convLSTM2(h1, (h2, c2))\n",
    "\n",
    "        # 시간 차원에 대해 평균 풀링\n",
    "        pooled = self.avgPool(h2.unsqueeze(2))  # (batch_size, channels, 1, height, width)\n",
    "        #print(f\"After pooling shape: {x.shape}\")\n",
    "\n",
    "        # 텐서 평탄화 전의 출력 크기를 동적으로 계산\n",
    "        flat_size = x.size(1) * x.size(2) * x.size(3)\n",
    "        x = x.view(x.size(0), -1)  # 배치 크기를 유지하면서 평탄화\n",
    "        #print(f\"After flattening shape: {x.shape}\")\n",
    "        # 동적으로 계산된 크기를 사용하여 fc 레이어 적용\n",
    "        x = self.fc1(x)\n",
    "        #print(f\"After FC1 layers shape: {x.shape}\")\n",
    "        x = self.fc2(x)\n",
    "        return x\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hBTCc1I_RKfR"
   },
   "source": [
    "### conv + lstm layer\n",
    "- default: conv4 lstm1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "executionInfo": {
     "elapsed": 2,
     "status": "ok",
     "timestamp": 1710265881408,
     "user": {
      "displayName": "김성호",
      "userId": "00239474051960738712"
     },
     "user_tz": -540
    },
    "id": "-TejvNMu3wuF"
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "# CNN\n",
    "class Simple3DCNN(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(Simple3DCNN, self).__init__()\n",
    "        self.conv1 = nn.Conv3d(3, 8, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\n",
    "        self.pool = nn.MaxPool3d(kernel_size=(1, 2, 2), stride=(1, 2, 2))\n",
    "        self.conv2 = nn.Conv3d(8, 16, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\n",
    "        self.pool = nn.MaxPool3d(kernel_size=(1, 2, 2), stride=(1, 2, 2))\n",
    "        self.conv3 = nn.Conv3d(16, 32, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\n",
    "        self.pool = nn.MaxPool3d(kernel_size=(1, 2, 2), stride=(1, 2, 2))\n",
    "        self.conv4 = nn.Conv3d(32, 64, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\n",
    "\n",
    "        self.lstm = nn.LSTM(input_size=64*7*7, hidden_size=128, num_layers=1, batch_first=True)\n",
    "\n",
    "        #self.fc1 = nn.Linear(64*3*3*3, 512)\n",
    "        self.fc = nn.Linear(128, 2)\n",
    "\n",
    "    def forward(self, x):\n",
    "\n",
    "        #print(f\"Initial input shape: {x.shape}\") # 4, 3, 60, 56, 56\n",
    "\n",
    "        x = self.pool(F.relu(self.conv1(x)))\n",
    "        #print(f\"After conv1 shape: {x.shape}\") # 4, 8, 30, 28, 28\n",
    "\n",
    "        x = self.pool(F.relu(self.conv2(x)))\n",
    "        #print(f\"After conv2 shape: {x.shape}\") # 4, 16, 15, 14, 14\n",
    "\n",
    "        x = self.pool(F.relu(self.conv3(x)))\n",
    "        #print(f\"After conv3 shape: {x.shape}\") # 4, 32, 7, 7, 7\n",
    "\n",
    "        x = self.pool(F.relu(self.conv4(x)))\n",
    "        #print(f\"After conv4 shape: {x.shape}\") # 4, 64, 3, 3, 3\n",
    "\n",
    "        # LSTM을 위한 준비: (batch_size, seq_len, features)\n",
    "        x = x.view(x.size(0), x.size(2), -1)  # x.size(0)은 배치 크기, Flatten the tensor\n",
    "        #print(f\"Before LSTM layers shape: {x.shape}\") # 4, 32*3*7*7\n",
    "\n",
    "        # LSTM을 통과\n",
    "        lstm_out, (h_n, c_n) = self.lstm(x)\n",
    "        #print(f\"After LSTM layers shape: {x.shape}\")\n",
    "        x = lstm_out[:, -1, :]  # 시퀀스의 마지막 출력만 사용\n",
    "        #print(f\"시퀀스의 마지막 출력만 사용 이후 shape: {x.shape}\")\n",
    "        #x = F.relu(self.fc1(x))\n",
    "        #print(f\"After fc1 shape: {x.shape}\")\n",
    "        x = self.fc(x)\n",
    "        #print(f\"Output shape: {x.shape}\")\n",
    "        return x\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- conv3, lstm1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "# CNN\n",
    "class Simple3DCNN(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(Simple3DCNN, self).__init__()\n",
    "        self.conv1 = nn.Conv3d(3, 16, kernel_size=(3, 7, 7), stride=(1, 1, 1), padding=(1, 3, 3))\n",
    "        self.pool = nn.MaxPool3d(kernel_size=(1, 2, 2), stride=(1, 2, 2))\n",
    "        self.conv2 = nn.Conv3d(16, 32, kernel_size=(3, 5, 5), stride=(1, 1, 1), padding=(1, 2, 2))\n",
    "        self.pool = nn.MaxPool3d(kernel_size=(1, 2, 2), stride=(1, 2, 2))\n",
    "        self.conv3 = nn.Conv3d(32, 64, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\n",
    "        #self.pool = nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2))\n",
    "        #self.conv4 = nn.Conv3d(32, 64, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\n",
    "\n",
    "        self.lstm = nn.LSTM(input_size=64*14*14, hidden_size=128, num_layers=1, batch_first=True)\n",
    "\n",
    "        #self.fc1 = nn.Linear(64*3*3*3, 512)\n",
    "        self.fc = nn.Linear(128, 2)\n",
    "\n",
    "    def forward(self, x):\n",
    "\n",
    "        print(f\"Initial input shape: {x.shape}\") # 4, 3, 60, 56, 56\n",
    "\n",
    "        x = self.pool(F.relu(self.conv1(x)))\n",
    "        print(f\"After conv1 shape: {x.shape}\") # 4, 8, 30, 28, 28\n",
    "\n",
    "        x = self.pool(F.relu(self.conv2(x)))\n",
    "        print(f\"After conv2 shape: {x.shape}\") # 4, 16, 15, 14, 14\n",
    "\n",
    "        x = self.pool(F.relu(self.conv3(x)))\n",
    "        print(f\"After conv3 shape: {x.shape}\") # 4, 32, 7, 7, 7\n",
    "\n",
    "        #x = self.pool(F.relu(self.conv4(x)))\n",
    "        #print(f\"After conv4 shape: {x.shape}\") # 4, 64, 3, 3, 3\n",
    "\n",
    "        # LSTM을 위한 준비: (batch_size, seq_len, features)\n",
    "        x = x.view(x.size(0), x.size(2), -1)  # x.size(0)은 배치 크기, Flatten the tensor\n",
    "        print(f\"Before LSTM layers shape: {x.shape}\") # 4, 32*3*7*7\n",
    "\n",
    "        # LSTM을 통과\n",
    "        lstm_out, (h_n, c_n) = self.lstm(x)\n",
    "        print(f\"After LSTM layers shape: {x.shape}\")\n",
    "        x = lstm_out[:, -1, :]  # 시퀀스의 마지막 출력만 사용\n",
    "        #print(f\"시퀀스의 마지막 출력만 사용 이후 shape: {x.shape}\")\n",
    "        #x = F.relu(self.fc1(x))\n",
    "        print(f\"After fc1 shape: {x.shape}\")\n",
    "        x = self.fc(x)\n",
    "        print(f\"Output shape: {x.shape}\")\n",
    "        return x\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### CSN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "class CSNBottleneck(nn.Module):\n",
    "    expansion = 4\n",
    "\n",
    "    def __init__(self, in_planes, planes, stride=1):\n",
    "        super(CSNBottleneck, self).__init__()\n",
    "        # Assuming stride is a tuple for temporal, height, width dimensions: (T, H, W)\n",
    "        temporal_stride, height_stride, width_stride = stride\n",
    "\n",
    "        self.conv1 = nn.Conv3d(in_planes, planes, kernel_size=1, bias=False)\n",
    "        self.bn1 = nn.BatchNorm3d(planes)\n",
    "        # Using the temporal_stride for the middle conv layer to preserve temporal dimension\n",
    "        self.conv2 = nn.Conv3d(planes, planes, kernel_size=3, stride=(temporal_stride, height_stride, width_stride), padding=1, bias=False)\n",
    "        self.bn2 = nn.BatchNorm3d(planes)\n",
    "        self.conv3 = nn.Conv3d(planes, planes * self.expansion, kernel_size=1, bias=False)\n",
    "        self.bn3 = nn.BatchNorm3d(planes * self.expansion)\n",
    "\n",
    "        self.shortcut = nn.Sequential()\n",
    "        if stride != (1, 1, 1) or in_planes != planes * self.expansion:\n",
    "            self.shortcut = nn.Sequential(\n",
    "                nn.Conv3d(in_planes, planes * self.expansion, kernel_size=1, stride=stride, bias=False),\n",
    "                nn.BatchNorm3d(planes * self.expansion)\n",
    "            )\n",
    "\n",
    "    def forward(self, x):\n",
    "        out = F.relu(self.bn1(self.conv1(x)))\n",
    "        out = F.relu(self.bn2(self.conv2(out)))\n",
    "        out = self.bn3(self.conv3(out))\n",
    "        out += self.shortcut(x)\n",
    "        out = F.relu(out)\n",
    "        return out\n",
    "\n",
    "class CSNResNet(nn.Module):\n",
    "    def __init__(self, block, num_blocks, num_classes=2):\n",
    "        super(CSNResNet, self).__init__()\n",
    "        self.in_planes = 64\n",
    "\n",
    "        self.conv1 = nn.Conv3d(3, 64, kernel_size=(7, 7, 7), stride=(1, 2, 2), padding=(3, 3, 3), bias=False)\n",
    "        self.bn1 = nn.BatchNorm3d(64)\n",
    "        self.pool1 = nn.MaxPool3d(kernel_size=(1, 3, 3), stride=(1, 2, 2), padding=(0, 1, 1))\n",
    "        \n",
    "        self.layer1 = self._make_layer(block, 64, num_blocks[0], stride=(1, 1, 1))\n",
    "        self.layer2 = self._make_layer(block, 128, num_blocks[1], stride=(1, 2, 2))\n",
    "        self.layer3 = self._make_layer(block, 256, num_blocks[2], stride=(1, 2, 2))\n",
    "        self.layer4 = self._make_layer(block, 512, num_blocks[3], stride=(1, 2, 2))\n",
    "\n",
    "        self.pool2 = nn.AdaptiveAvgPool3d((1, 1, 1))\n",
    "        self.fc = nn.Linear(512 * block.expansion, num_classes)\n",
    "\n",
    "    def _make_layer(self, block, planes, num_blocks, stride):\n",
    "        strides = [stride] + [(1, 1, 1)] * (num_blocks - 1)\n",
    "        layers = []\n",
    "        for stride in strides:\n",
    "            layers.append(block(self.in_planes, planes, stride))\n",
    "            self.in_planes = planes * block.expansion\n",
    "        return nn.Sequential(*layers)\n",
    "\n",
    "    def forward(self, x):\n",
    "        out = F.relu(self.bn1(self.conv1(x)))\n",
    "        out = self.pool1(out)\n",
    "        out = self.layer1(out)\n",
    "        out = self.layer2(out)\n",
    "        out = self.layer3(out)\n",
    "        out = self.layer4(out)\n",
    "        out = self.pool2(out)\n",
    "        out = out.view(out.size(0), -1)\n",
    "        out = self.fc(out)\n",
    "        return out\n",
    "\n",
    "def CSNResNet101():\n",
    "    # (3, 4, 23, 3) are the number of blocks in each layer\n",
    "    return CSNResNet(CSNBottleneck, [3, 4, 23, 3])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "#### cudnn initializa 오류 발생 시에만 run ####\n",
    "def force_cudnn_initialization():\n",
    "    s = 32\n",
    "    dev = torch.device('cuda')\n",
    "    torch.nn.functional.conv2d(torch.zeros(s, s, s, s, device=dev), torch.zeros(s, s, s, s, device=dev))\n",
    "\n",
    "force_cudnn_initialization()'''"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### EfficientViT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/daiv/바탕화면/efficientvit\n"
     ]
    }
   ],
   "source": [
    "cd /home/daiv/바탕화면/efficientvit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "from efficientvit.models.efficientvit.backbone import EfficientViTBackbone\n",
    "from efficientvit.models.utils import build_kwargs_from_config\n",
    "\n",
    "class ClsHead(nn.Module):\n",
    "    def __init__(self, in_channels, width_list, n_classes=2, dropout=0.0, norm=\"bn2d\", act_func=\"hswish\", fid=\"stage_final\"):\n",
    "        super().__init__()\n",
    "        self.in_channels = in_channels\n",
    "        self.width_list = width_list\n",
    "        self.n_classes = n_classes\n",
    "        self.dropout = dropout\n",
    "        self.norm = norm\n",
    "        self.act_func = act_func\n",
    "        self.fid = fid\n",
    "        \n",
    "        self.layers = nn.Sequential(\n",
    "            nn.Conv2d(in_channels, width_list[0], kernel_size=1, bias=False),\n",
    "            nn.BatchNorm2d(width_list[0]),\n",
    "            nn.ReLU(inplace=True),\n",
    "            nn.AdaptiveAvgPool2d(1),\n",
    "            nn.Flatten(),\n",
    "            nn.Dropout(dropout),\n",
    "            nn.Linear(width_list[0], width_list[1]),\n",
    "            nn.ReLU(inplace=True),\n",
    "            nn.Linear(width_list[1], n_classes)\n",
    "        )\n",
    "    \n",
    "    def forward(self, x):\n",
    "        x = self.layers(x)\n",
    "        return x\n",
    "\n",
    "class EfficientViTCls(nn.Module):\n",
    "    def __init__(self, backbone_cfg, head_cfg):\n",
    "        super().__init__()\n",
    "        self.backbone = EfficientViTBackbone(**build_kwargs_from_config(backbone_cfg, EfficientViTBackbone))\n",
    "        self.head = ClsHead(**build_kwargs_from_config(head_cfg, ClsHead))\n",
    "    \n",
    "    def forward(self, x):\n",
    "        x = self.backbone(x)\n",
    "        # Adapt backbone output to match ClsHead input\n",
    "        # This part may require further modification depending on the output of the backbone\n",
    "        x = x['stage_final']  # Assuming that the final output of the backbone is under 'stage_final'\n",
    "        x = self.head(x)\n",
    "        return x\n",
    "\n",
    "class EfficientViTForVideo(nn.Module):\n",
    "    def __init__(self, backbone_cfg, head_cfg, num_frames):\n",
    "        super().__init__()\n",
    "        self.efficientvit_cls = EfficientViTCls(backbone_cfg, head_cfg)\n",
    "        self.num_frames = num_frames\n",
    "    \n",
    "    def forward(self, x):\n",
    "        # x: [batch_size, channels, time, height, width]\n",
    "        batch_size, channels, time, height, width = x.size()\n",
    "        x = x.view(batch_size * time, channels, height, width)  # 5D to 4D\n",
    "        \n",
    "        # 모델에 각 프레임을 독립적으로 전달\n",
    "        predictions = self.efficientvit_cls(x)\n",
    "        \n",
    "        # 각 프레임의 예측 결과를 평균냄\n",
    "        predictions = predictions.view(batch_size, time, -1).mean(1)\n",
    "        return predictions\n",
    "\n",
    "# Example configuration\n",
    "backbone_cfg = {\n",
    "    'width_list': [64, 128, 256, 512, 1024],\n",
    "    'depth_list': [1, 2, 2, 8, 8],\n",
    "    'dim': 64  # This parameter will depend on EfficientViTBackbone and may need to be adjusted\n",
    "}\n",
    "\n",
    "head_cfg = {\n",
    "    'in_channels': 1024,  # This should match the last channel size of the backbone\n",
    "    'width_list': [1024, 1280],\n",
    "    'n_classes': 2,\n",
    "    'dropout': 0.5,\n",
    "    'norm': 'bn2d',\n",
    "    'act_func': 'relu'\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/hyoju/바탕화면/RealForensics/data/Forensics\n"
     ]
    }
   ],
   "source": [
    "%cd /home/hyoju/바탕화면/RealForensics/data/Forensics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initial input shape: torch.Size([1, 3, 25, 112, 112])\n",
      "After conv1 shape: torch.Size([1, 16, 25, 56, 56])\n",
      "After conv2 shape: torch.Size([1, 32, 25, 28, 28])\n",
      "After conv3 shape: torch.Size([1, 64, 25, 14, 14])\n",
      "Before LSTM layers shape: torch.Size([1, 25, 12544])\n",
      "After LSTM layers shape: torch.Size([1, 25, 12544])\n",
      "After fc1 shape: torch.Size([1, 128])\n",
      "Output shape: torch.Size([1, 2])\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "==========================================================================================\n",
       "Layer (type:depth-idx)                   Output Shape              Param #\n",
       "==========================================================================================\n",
       "Simple3DCNN                              [1, 2]                    --\n",
       "├─Conv3d: 1-1                            [1, 16, 25, 112, 112]     7,072\n",
       "├─MaxPool3d: 1-2                         [1, 16, 25, 56, 56]       --\n",
       "├─Conv3d: 1-3                            [1, 32, 25, 56, 56]       38,432\n",
       "├─MaxPool3d: 1-4                         [1, 32, 25, 28, 28]       --\n",
       "├─Conv3d: 1-5                            [1, 64, 25, 28, 28]       55,360\n",
       "├─MaxPool3d: 1-6                         [1, 64, 25, 14, 14]       --\n",
       "├─LSTM: 1-7                              [1, 25, 128]              6,489,088\n",
       "├─Linear: 1-8                            [1, 2]                    258\n",
       "==========================================================================================\n",
       "Total params: 6,590,210\n",
       "Trainable params: 6,590,210\n",
       "Non-trainable params: 0\n",
       "Total mult-adds (G): 6.48\n",
       "==========================================================================================\n",
       "Input size (MB): 3.76\n",
       "Forward/backward pass size (MB): 70.27\n",
       "Params size (MB): 26.36\n",
       "Estimated Total Size (MB): 100.40\n",
       "=========================================================================================="
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from torchinfo import summary\n",
    "\n",
    "# 모델 인스턴스 생성\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "model = Simple3DCNN().to(device)\n",
    "#model = CSNResNet101().to(device)\n",
    "#model = EfficientViTForVideo(backbone_cfg, head_cfg,frames_limit).to(device)\n",
    "\n",
    "# 요약 정보 출력\n",
    "summary(model, input_size=(1, 3, 25, 112, 112))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "HNbSOSUiRIZ4"
   },
   "source": [
    "- conv3 + lstm1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "executionInfo": {
     "elapsed": 461,
     "status": "ok",
     "timestamp": 1710279545674,
     "user": {
      "displayName": "김성호",
      "userId": "00239474051960738712"
     },
     "user_tz": -540
    },
    "id": "FQGxHGdRRHG-"
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "# CNN\n",
    "class LSTM3DCNN(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(LSTM3DCNN, self).__init__()\n",
    "        self.conv1 = nn.Conv3d(3, 8, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\n",
    "        self.pool = nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2))\n",
    "\n",
    "        self.conv2 = nn.Conv3d(8, 16, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\n",
    "        self.pool = nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2))\n",
    "        self.dropout1 = nn.Dropout(0.2)\n",
    "\n",
    "        self.conv3 = nn.Conv3d(16, 32, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\n",
    "        self.dropout2 = nn.Dropout(0.25)\n",
    "        #self.pool = nn.MaxPool3d(kernel_size=(2, 2, 2), stride=(2, 2, 2))\n",
    "        #self.conv4 = nn.Conv3d(32, 64, kernel_size=(3, 3, 3), stride=(1, 1, 1), padding=(1, 1, 1))\n",
    "\n",
    "        self.lstm1 = nn.LSTM(input_size=32*7*7, hidden_size=512, num_layers=1, batch_first=True)\n",
    "        self.lstm2 = nn.LSTM(input_size=512, hidden_size=256, num_layers=1, batch_first=True)\n",
    "\n",
    "        self.fc1 = nn.Linear(256, 32)\n",
    "        self.dropout3 = nn.Dropout(0.2)\n",
    "        self.fc2 = nn.Linear(32, 2)\n",
    "\n",
    "    def forward(self, x):\n",
    "\n",
    "        #print(f\"Initial input shape: {x.shape}\") # 4, 3, 60, 56, 56\n",
    "\n",
    "        x = self.pool(F.relu(self.conv1(x)))\n",
    "        #print(f\"After conv1 shape: {x.shape}\") # 4, 8, 30, 28, 28\n",
    "\n",
    "        x = self.pool(F.relu(self.conv2(x)))\n",
    "        x = self.dropout1(x)\n",
    "        #print(f\"After conv2 shape: {x.shape}\") # 4, 16, 15, 14, 14\n",
    "\n",
    "        x = self.pool(F.relu(self.conv3(x)))\n",
    "        x = self.dropout2(x)\n",
    "        #print(f\"After conv3 shape: {x.shape}\") # 4, 32, 7, 7, 7\n",
    "\n",
    "        #x = self.pool(F.relu(self.conv4(x)))\n",
    "        #print(f\"After conv4 shape: {x.shape}\") # 4, 64, 3, 3, 3\n",
    "\n",
    "        # LSTM을 위한 준비: (batch_size, seq_len, features)\n",
    "        x = x.view(x.size(0), x.size(2), -1)  # x.size(0)은 배치 크기, Flatten the tensor\n",
    "        #print(f\"Before LSTM layers shape: {x.shape}\") # 4, 32*3*7*7\n",
    "\n",
    "        # LSTM을 통과\n",
    "        x, _ = self.lstm1(x)\n",
    "        x, _ = self.lstm2(x)\n",
    "        #print(f\"After LSTM layers shape: {x.shape}\")\n",
    "        x = x[:, -1, :]  # 시퀀스의 마지막 출력만 사용\n",
    "        #print(f\"시퀀스의 마지막 출력만 사용 이후 shape: {x.shape}\")\n",
    "        #x = F.relu(self.fc1(x))\n",
    "        #print(f\"After fc1 shape: {x.shape}\")\n",
    "        x = self.fc1(x)\n",
    "        x = self.dropout3(x)\n",
    "        x = self.fc2(x)\n",
    "        #print(f\"Output shape: {x.shape}\")\n",
    "        return x\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([144, 3, 90, 112, 112]) torch.Size([144])\n"
     ]
    }
   ],
   "source": [
    "#train_loader의 구성을 알고 싶다면\n",
    "for x, y in train_loader:\n",
    "    print(x.shape, y.shape)\n",
    "    break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 39712,
     "status": "ok",
     "timestamp": 1710276630887,
     "user": {
      "displayName": "김성호",
      "userId": "00239474051960738712"
     },
     "user_tz": -540
    },
    "id": "0tnXFcwH3t0M",
    "outputId": "5da1a37b-8590-42be-ee61-d4ed93606995"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Batch size: torch.Size([144, 3, 60, 56, 56])\n",
      "Input batch dimension after permutation: torch.Size([144, 3, 60, 56, 56])\n",
      "Before LSTM layers shape: torch.Size([144, 7, 1568])\n",
      "After LSTM layers shape: torch.Size([144, 7, 256])\n",
      "시퀀스의 마지막 출력만 사용 이후 shape: torch.Size([144, 256])\n",
      "Output shape: torch.Size([144, 2])\n",
      "Output shape: torch.Size([144, 2]), Labels shape: torch.Size([144])\n"
     ]
    }
   ],
   "source": [
    "# 디버깅: shape 확인\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "model = LSTM3DCNN().to(device)\n",
    "#model = SimpleConvLSTMModel().to(device)\n",
    "for videos, labels in train_loader:\n",
    "    print(f\"Batch size: {videos.shape}\")\n",
    "    videos = videos.to(device)\n",
    "    print(f'Input batch dimension after permutation: {videos.shape}')\n",
    "    labels = labels.to(device)\n",
    "    outputs = model(videos)\n",
    "    print(f\"Output shape: {outputs.shape}, Labels shape: {labels.shape}\")\n",
    "    break  # 첫 번째 배치만 확인하고 중단"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 464,
     "status": "ok",
     "timestamp": 1710279548231,
     "user": {
      "displayName": "김성호",
      "userId": "00239474051960738712"
     },
     "user_tz": -540
    },
    "id": "8hU-rty3enEf",
    "outputId": "cc05e116-d3e7-4c6c-835c-f1c668bd77d9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True\n",
      "11.8\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "print(torch.cuda.is_available())\n",
    "print(torch.version.cuda)\n",
    "\n",
    "torch.cuda.empty_cache()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tue Mar 19 23:48:40 2024       \n",
      "+-----------------------------------------------------------------------------+\n",
      "| NVIDIA-SMI 525.147.05   Driver Version: 525.147.05   CUDA Version: 12.0     |\n",
      "|-------------------------------+----------------------+----------------------+\n",
      "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
      "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
      "|                               |                      |               MIG M. |\n",
      "|===============================+======================+======================|\n",
      "|   0  NVIDIA GeForce ...  Off  | 00000000:01:00.0 Off |                  N/A |\n",
      "| 31%   39C    P0   114W / 420W |      2MiB / 24576MiB |      0%      Default |\n",
      "|                               |                      |                  N/A |\n",
      "+-------------------------------+----------------------+----------------------+\n",
      "                                                                               \n",
      "+-----------------------------------------------------------------------------+\n",
      "| Processes:                                                                  |\n",
      "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
      "|        ID   ID                                                   Usage      |\n",
      "|=============================================================================|\n",
      "|  No running processes found                                                 |\n",
      "+-----------------------------------------------------------------------------+\n"
     ]
    }
   ],
   "source": [
    "!nvidia-smi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "background_save": true,
     "base_uri": "https://localhost:8080/",
     "height": 603
    },
    "id": "j-kdPwhM3gQt"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\u001b[34m\u001b[1mwandb\u001b[0m: \u001b[33mWARNING\u001b[0m Path /home/hyoju/바탕화면/RealForensics/data/Forensics/wandb/ wasn't writable, using system temp directory.\n",
      "Failed to detect the name of this notebook, you can set it manually with the WANDB_NOTEBOOK_NAME environment variable to enable code saving.\n",
      "wandb: WARNING Path /home/hyoju/바탕화면/RealForensics/data/Forensics/wandb/ wasn't writable, using system temp directory\n",
      "\u001b[34m\u001b[1mwandb\u001b[0m: \u001b[33mWARNING\u001b[0m Path /home/hyoju/바탕화면/RealForensics/data/Forensics/wandb/ wasn't writable, using system temp directory\n",
      "\u001b[34m\u001b[1mwandb\u001b[0m: Currently logged in as: \u001b[33mwoony\u001b[0m (\u001b[33msamjang\u001b[0m). Use \u001b[1m`wandb login --relogin`\u001b[0m to force relogin\n",
      "\u001b[34m\u001b[1mwandb\u001b[0m: \u001b[33mWARNING\u001b[0m If you're specifying your api key in code, ensure this code is not shared publicly.\n",
      "\u001b[34m\u001b[1mwandb\u001b[0m: \u001b[33mWARNING\u001b[0m Consider setting the WANDB_API_KEY environment variable, or running `wandb login` from the command line.\n",
      "\u001b[34m\u001b[1mwandb\u001b[0m: Appending key for api.wandb.ai to your netrc file: /home/daiv/.netrc\n",
      "\u001b[34m\u001b[1mwandb\u001b[0m: Currently logged in as: \u001b[33mtaho\u001b[0m (\u001b[33msamjang\u001b[0m). Use \u001b[1m`wandb login --relogin`\u001b[0m to force relogin\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "Tracking run with wandb version 0.16.4"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Run data is saved locally in <code>/tmp/wandb/run-20240319_235138-n57eqyoq</code>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "Syncing run <strong><a href='https://wandb.ai/samjang/deepfake_classification/runs/n57eqyoq' target=\"_blank\">conv4_lstm1_1e-3(10,0.5)_120_144_112</a></strong> to <a href='https://wandb.ai/samjang/deepfake_classification' target=\"_blank\">Weights & Biases</a> (<a href='https://wandb.me/run' target=\"_blank\">docs</a>)<br/>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View project at <a href='https://wandb.ai/samjang/deepfake_classification' target=\"_blank\">https://wandb.ai/samjang/deepfake_classification</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       " View run at <a href='https://wandb.ai/samjang/deepfake_classification/runs/n57eqyoq' target=\"_blank\">https://wandb.ai/samjang/deepfake_classification/runs/n57eqyoq</a>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "ename": "OutOfMemoryError",
     "evalue": "CUDA out of memory. Tried to allocate 6.46 GiB. GPU 0 has a total capacity of 23.69 GiB of which 4.27 GiB is free. Including non-PyTorch memory, this process has 19.42 GiB memory in use. Of the allocated memory 11.34 GiB is allocated by PyTorch, and 6.51 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation.  See documentation for Memory Management  (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mOutOfMemoryError\u001b[0m                          Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[11], line 74\u001b[0m\n\u001b[1;32m     72\u001b[0m outputs \u001b[38;5;241m=\u001b[39m model(videos)\n\u001b[1;32m     73\u001b[0m loss \u001b[38;5;241m=\u001b[39m criterion(outputs, labels)\n\u001b[0;32m---> 74\u001b[0m \u001b[43mloss\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mbackward\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     75\u001b[0m optimizer\u001b[38;5;241m.\u001b[39mstep()\n\u001b[1;32m     77\u001b[0m train_loss \u001b[38;5;241m+\u001b[39m\u001b[38;5;241m=\u001b[39m loss\u001b[38;5;241m.\u001b[39mitem() \u001b[38;5;241m*\u001b[39m videos\u001b[38;5;241m.\u001b[39msize(\u001b[38;5;241m0\u001b[39m)\n",
      "File \u001b[0;32m~/anaconda3/envs/daiv/lib/python3.8/site-packages/torch/_tensor.py:522\u001b[0m, in \u001b[0;36mTensor.backward\u001b[0;34m(self, gradient, retain_graph, create_graph, inputs)\u001b[0m\n\u001b[1;32m    512\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m has_torch_function_unary(\u001b[38;5;28mself\u001b[39m):\n\u001b[1;32m    513\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m handle_torch_function(\n\u001b[1;32m    514\u001b[0m         Tensor\u001b[38;5;241m.\u001b[39mbackward,\n\u001b[1;32m    515\u001b[0m         (\u001b[38;5;28mself\u001b[39m,),\n\u001b[0;32m   (...)\u001b[0m\n\u001b[1;32m    520\u001b[0m         inputs\u001b[38;5;241m=\u001b[39minputs,\n\u001b[1;32m    521\u001b[0m     )\n\u001b[0;32m--> 522\u001b[0m \u001b[43mtorch\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mautograd\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mbackward\u001b[49m\u001b[43m(\u001b[49m\n\u001b[1;32m    523\u001b[0m \u001b[43m    \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mgradient\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mretain_graph\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mcreate_graph\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43minputs\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43minputs\u001b[49m\n\u001b[1;32m    524\u001b[0m \u001b[43m\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/anaconda3/envs/daiv/lib/python3.8/site-packages/torch/autograd/__init__.py:266\u001b[0m, in \u001b[0;36mbackward\u001b[0;34m(tensors, grad_tensors, retain_graph, create_graph, grad_variables, inputs)\u001b[0m\n\u001b[1;32m    261\u001b[0m     retain_graph \u001b[38;5;241m=\u001b[39m create_graph\n\u001b[1;32m    263\u001b[0m \u001b[38;5;66;03m# The reason we repeat the same comment below is that\u001b[39;00m\n\u001b[1;32m    264\u001b[0m \u001b[38;5;66;03m# some Python versions print out the first line of a multi-line function\u001b[39;00m\n\u001b[1;32m    265\u001b[0m \u001b[38;5;66;03m# calls in the traceback and some print out the last line\u001b[39;00m\n\u001b[0;32m--> 266\u001b[0m \u001b[43mVariable\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_execution_engine\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mrun_backward\u001b[49m\u001b[43m(\u001b[49m\u001b[43m  \u001b[49m\u001b[38;5;66;43;03m# Calls into the C++ engine to run the backward pass\u001b[39;49;00m\n\u001b[1;32m    267\u001b[0m \u001b[43m    \u001b[49m\u001b[43mtensors\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    268\u001b[0m \u001b[43m    \u001b[49m\u001b[43mgrad_tensors_\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    269\u001b[0m \u001b[43m    \u001b[49m\u001b[43mretain_graph\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    270\u001b[0m \u001b[43m    \u001b[49m\u001b[43mcreate_graph\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    271\u001b[0m \u001b[43m    \u001b[49m\u001b[43minputs\u001b[49m\u001b[43m,\u001b[49m\n\u001b[1;32m    272\u001b[0m \u001b[43m    \u001b[49m\u001b[43mallow_unreachable\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mTrue\u001b[39;49;00m\u001b[43m,\u001b[49m\n\u001b[1;32m    273\u001b[0m \u001b[43m    \u001b[49m\u001b[43maccumulate_grad\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mTrue\u001b[39;49;00m\u001b[43m,\u001b[49m\n\u001b[1;32m    274\u001b[0m \u001b[43m\u001b[49m\u001b[43m)\u001b[49m\n",
      "\u001b[0;31mOutOfMemoryError\u001b[0m: CUDA out of memory. Tried to allocate 6.46 GiB. GPU 0 has a total capacity of 23.69 GiB of which 4.27 GiB is free. Including non-PyTorch memory, this process has 19.42 GiB memory in use. Of the allocated memory 11.34 GiB is allocated by PyTorch, and 6.51 GiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation.  See documentation for Memory Management  (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)"
     ]
    }
   ],
   "source": [
    "import torch.optim as optim\n",
    "from adamp import AdamP\n",
    "from torch.optim.lr_scheduler import LambdaLR, StepLR\n",
    "import wandb\n",
    "\n",
    "learning_rate=1e-3\n",
    "num_epochs = 100\n",
    "train_losses = []\n",
    "val_losses = []\n",
    "train_accuracies = []\n",
    "val_accuracies = []\n",
    "\n",
    "# training setting\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "model = Simple3DCNN().to(device)\n",
    "#model = SimpleConvLSTMModel().to(device)\n",
    "#model = LSTM3DCNN().to(device)\n",
    "#model = CSNResNet101().to(device)\n",
    "#model = EfficientViTForVideo(backbone_cfg, head_cfg, frames_limit).to(device)\n",
    "\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = AdamP(model.parameters(), lr=learning_rate)\n",
    "#scheduler = LambdaLR(optimizer, lr_lambda=lambda epoch: 0.9 ** (epoch // 5))\n",
    "scheduler = StepLR(optimizer, step_size=10, gamma=0.5)\n",
    " \n",
    "\n",
    "# early stopping\n",
    "min_val_loss = np.inf\n",
    "patience = 20\n",
    "patience_counter = 0\n",
    "\n",
    "'''\n",
    "wandb.init(\n",
    "    project='deepfake_classification',\n",
    "    entity='samjang',\n",
    "    name='my_custom_run_name',\n",
    "    tags=['tag1', 'tag2'],\n",
    "    notes='my custom run for deepfake classification.'\n",
    ")\n",
    "'''\n",
    "\n",
    "# WandB에 로그인 (API key)\n",
    "wandb.login(key=\"8dd8fb438a7e71a9afd94334dd57390f73db768c\")\n",
    "\n",
    "# wandb 초기화\n",
    "wandb.init(project='deepfake_classification', entity='samjang', name='conv4_lstm1_1e-3(10,0.5)_120_144_112')\n",
    "\n",
    "\n",
    "# wandb config 기록 (Option)\n",
    "wandb.config = {\n",
    "  \"learning_rate\": learning_rate,\n",
    "  \"epochs\": num_epochs,\n",
    "  \"batch_size\": batch_size\n",
    "}\n",
    "\n",
    "for epoch in range(num_epochs):\n",
    "    model.train()\n",
    "    train_loss = 0.0\n",
    "    train_correct = 0\n",
    "    total_train = 0\n",
    "\n",
    "    for videos, labels in train_loader:\n",
    "        #print(f'Input batch dimension: {videos.shape}') # [4, 112, 270, 3, 112] (batch_size, width, depth, channels, height)\n",
    "        #print(f'Labels batch dimension: {labels.shape}')\n",
    "        #break  # 첫 배치만 확인하고 반복 중지\n",
    "\n",
    "        # DataLoader로부터의 출력: (batch_size, depth, height, channels, width)\n",
    "        videos = videos.to(device) # (batch_size, C, D, H, W)로 변경\n",
    "        labels = labels.to(device)\n",
    "\n",
    "        optimizer.zero_grad()\n",
    "        outputs = model(videos)\n",
    "        loss = criterion(outputs, labels)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        train_loss += loss.item() * videos.size(0)\n",
    "        train_correct += (outputs.argmax(1) == labels).sum().item()\n",
    "        total_train += labels.size(0)\n",
    "    \n",
    "    scheduler.step()  # 학습률 업데이트\n",
    "    wandb.log({\"train_loss\": train_loss / total_train, \"train_accuracy\": train_correct / total_train})\n",
    "    train_losses.append(train_loss / total_train)\n",
    "    train_accuracies.append(train_correct / total_train)\n",
    "    \n",
    "    model.eval()\n",
    "    val_loss = 0.0\n",
    "    val_correct = 0\n",
    "    total_val = 0\n",
    "\n",
    "    with torch.no_grad():\n",
    "        for videos, labels in val_loader:\n",
    "            videos = videos.to(device)\n",
    "            labels = labels.to(device)\n",
    "\n",
    "            outputs = model(videos)\n",
    "            loss = criterion(outputs, labels)\n",
    "            val_loss += loss.item() * videos.size(0)\n",
    "            val_correct += (outputs.argmax(1) == labels).sum().item()\n",
    "            total_val += labels.size(0)\n",
    "\n",
    "    # early stopping condition\n",
    "    if val_loss < min_val_loss:\n",
    "        min_val_loss = val_loss\n",
    "        patience_counter = 0\n",
    "    else:\n",
    "        patience_counter += 1\n",
    "        if patience_counter >= patience:\n",
    "            print(f'Early stopping triggered after {epoch + 1} epochs.')\n",
    "            break\n",
    "\n",
    "    wandb.log({\"val_loss\": val_loss / total_val, \"val_accuracy\": val_correct / total_val})\n",
    "    val_losses.append(val_loss / total_val)\n",
    "    val_accuracies.append(val_correct / total_val)\n",
    "\n",
    "    # loss, acc, lr(if changed by scheduler)\n",
    "    print(f'Epoch {epoch+1}, Train Loss: {train_losses[-1]:.4f}, Train Accuracy: {train_accuracies[-1]:.4f}, Val Loss: {val_losses[-1]:.4f}, Val Accuracy: {val_accuracies[-1]:.4f}') #, LR: {scheduler.get_last_lr()[0]}'\n",
    "\n",
    "wandb.finish()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 438
    },
    "executionInfo": {
     "elapsed": 1195,
     "status": "ok",
     "timestamp": 1710248938586,
     "user": {
      "displayName": "김성호",
      "userId": "00239474051960738712"
     },
     "user_tz": -540
    },
    "id": "ePqLdBbe3gIk",
    "outputId": "41e10ef0-7682-4f53-f1ee-1317a0363a36"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# loss, accuracy visualization\n",
    "plt.figure(figsize=(12, 6))\n",
    "\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(train_losses, label='Train Loss')\n",
    "plt.plot(val_losses, label='Validation Loss')\n",
    "plt.title('Training and Validation Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.plot(train_accuracies, label='Train Accuracy')\n",
    "plt.plot(val_accuracies, label='Validation Accuracy')\n",
    "plt.title('Training and Validation Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "e5fbAcccSv-_"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "gpuType": "T4",
   "name": "",
   "version": ""
  },
  "kernelspec": {
   "display_name": "daiv",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.18"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "12549c6d70624e63a7f361b367586f81": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "2056fe00f1154d6885b5adee13efe3c4": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_ebeefdca2df3494c9bfa1b941ba2c5bc",
      "max": 1,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_d698da1f1d074b4aa145dd6807c448fd",
      "value": 1
     }
    },
    "361aff98e1dc439c938e26aec8aaa924": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "VBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "VBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "VBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_f9e05efd6b8b4d85965d716ab6bdc4aa",
       "IPY_MODEL_2056fe00f1154d6885b5adee13efe3c4"
      ],
      "layout": "IPY_MODEL_12549c6d70624e63a7f361b367586f81"
     }
    },
    "8dfafb8fed1a4fb78ac0d9b10e94668f": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "9fca4ad217864c3d96313173dbd0065c": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "d698da1f1d074b4aa145dd6807c448fd": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "ebeefdca2df3494c9bfa1b941ba2c5bc": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "f9e05efd6b8b4d85965d716ab6bdc4aa": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "LabelModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "LabelModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "LabelView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_8dfafb8fed1a4fb78ac0d9b10e94668f",
      "placeholder": "​",
      "style": "IPY_MODEL_9fca4ad217864c3d96313173dbd0065c",
      "value": "0.011 MB of 0.011 MB uploaded\r"
     }
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}