mca_validations.ipynb

{"cells":[{"cell_type":"code","execution_count":1,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":30742,"status":"ok","timestamp":1683126729447,"user":{"displayName":"Tony Zhao","userId":"01794725021251545211"},"user_tz":-480},"id":"sZLlZRk09Efl","outputId":"1939dd68-4d21-45f0-dc1f-b25885dc635b"},"outputs":[],"source":["import pandas as pd\n","import torch\n","from torch import nn\n","import torch.nn.functional as F\n","\n","from sklearn.metrics import accuracy_score\n","from sklearn.metrics import confusion_matrix"]},{"cell_type":"code","execution_count":2,"metadata":{"id":"7xFwg4dAgAHL"},"outputs":[],"source":["# Customized 3D CNN model to adapt to the input shape\n","class cnn_classifier(nn.Module):\n","    def __init__(self):\n","        super().__init__()\n","        self.conv11 = nn.Conv3d(in_channels=1, out_channels=32, kernel_size=3, stride=1, padding=1)\n","        self.conv12 = nn.Conv3d(in_channels=32, out_channels=32, kernel_size=3, stride=1, padding=1)\n","        self.pool1 = nn.MaxPool3d(kernel_size=2, padding=(0, 1, 1))\n","\n","        self.conv21 = nn.Conv3d(in_channels=32, out_channels=64, kernel_size=3, stride=1, padding=1)\n","        self.conv22 = nn.Conv3d(in_channels=64, out_channels=64, kernel_size=3, stride=1, padding=1)\n","        self.pool2 = nn.MaxPool3d(kernel_size=2, padding=0)\n","\n","        self.fc_layer = nn.Linear(64*8*1*1, 2)\n","        self.dropout_layer = nn.Dropout(p=0.3)\n","\n","    def forward(self, xb):\n","        h1 = self.conv11(xb)\n","        h1 = self.conv12(h1)\n","        h1 = self.pool1(h1)\n","        h1 = F.relu(h1)\n","\n","        h2 = self.conv21(h1)\n","        h2 = self.conv22(h2)\n","        h2 = self.pool2(h2)\n","        h2 = F.relu(h2)\n","\n","        # Before the fully connected layer, we need to flatten the output\n","        flatten = h2.view(-1, 64*8*1*1)\n","        out = self.fc_layer(flatten)\n","        return out"]},{"cell_type":"code","execution_count":3,"metadata":{},"outputs":[],"source":["# The default validation type is 'arousal', change to validate others\n","# ['valence', 'arousal', 'dominance', 'liking']\n","validation_type = 'arousal'\n","validation_model = f'{validation_type}_model.pth'\n","validation_x_test = f'{validation_type}_x_test.pth'\n","validation_y_test = f'{validation_type}_y_test.pth'\n","\n","# Load model from the .pth file\n","model_state_dict = torch.load(f'models/{validation_model}', map_location=torch.device('cpu'))\n","model = cnn_classifier()\n","model.load_state_dict(model_state_dict)\n","\n","# Load the x_test and y_test\n","x_test = torch.load(f'models/{validation_x_test}')\n","y_test = torch.load(f'models/{validation_y_test}')"]},{"cell_type":"code","execution_count":4,"metadata":{"id":"53QF2nEt8S4M"},"outputs":[{"name":"stdout","output_type":"stream","text":["Validation Model: valence\n","Model Accuracy: 0.99296875\n","                    Predict: Low Valence  Predict: High Valence\n","True: Low Valence                   1104                      8\n","True: High Valence                    10                   1438\n"]}],"source":["device = 'cuda' if torch.cuda.is_available() else 'cpu'\n","\n","# Predict process\n","with torch.no_grad():\n","    outputs = model(x_test.view(-1, 1, 32, 5, 3).float().to(device))\n","    predicted = outputs.argmax(1)\n","\n","# Convert to numpy array\n","predicted = predicted.cpu().numpy()\n","y_test = y_test.cpu().numpy()\n","\n","# Calculate accuracy and confusion matrix\n","accuracy = accuracy_score(y_test, predicted)\n","conf_matrix = confusion_matrix(y_test, predicted)\n","conf_matrix_df = pd.DataFrame(conf_matrix,\n","                              index=['True: Low Valence', 'True: High Valence'],\n","                              columns=['Predict: Low Valence', 'Predict: High Valence'])\n","\n","print(f\"Validation Model: {validation_type}\")\n","print(f\"Model Accuracy: {accuracy}\")\n","print(conf_matrix_df)"]}],"metadata":{"accelerator":"GPU","colab":{"machine_shape":"hm","provenance":[]},"gpuClass":"premium","kernelspec":{"display_name":"Python 3","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":"3.11.4"}},"nbformat":4,"nbformat_minor":0}