cnn-implementation-for-mnist-dataset.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "41f34c43",
   "metadata": {
    "papermill": {
     "duration": 0.009337,
     "end_time": "2024-08-02T03:24:24.608132",
     "exception": false,
     "start_time": "2024-08-02T03:24:24.598795",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Explanation of the CNN Implementation for MNIST Dataset\n",
    "\n",
    "# 1. Data Analysis and Preprocessing\n",
    "\n",
    "1. **Load the Data:**\n",
    "   - Import the training, test, and sample submission CSV files using `pandas.read_csv`.\n",
    "   - This gives us the `train_df`, `test_df`, and `sample_submission` dataframes.\n",
    "\n",
    "2. **Check for Missing Values:**\n",
    "   - Use `isnull().sum().sum()` on the dataframes to ensure there are no missing values.\n",
    "   - This step ensures data integrity before proceeding with analysis.\n",
    "\n",
    "3. **Visualize Label Distribution:**\n",
    "   - Use `seaborn.countplot` to plot the distribution of digit labels in the training set.\n",
    "   - This helps us understand the balance of the classes in the dataset.\n",
    "\n",
    "4. **Separate Features and Labels:**\n",
    "   - Extract the pixel values (features) and the labels from the training dataframe.\n",
    "   - The features are stored in `X` and the labels in `y`.\n",
    "\n",
    "5. **Normalize Pixel Values:**\n",
    "   - Scale the pixel values to the range [0, 1] by dividing by 255.\n",
    "   - Normalization helps in faster convergence of the model during training.\n",
    "\n",
    "6. **Reshape Data:**\n",
    "   - Reshape the data into the shape required by the neural network, which is `(28, 28, 1)` for each image.\n",
    "   - This shape is necessary for the Conv2D layers in the CNN.\n",
    "\n",
    "7. **Categorical Encoding:**\n",
    "   - Convert the labels to one-hot encoded format using `tensorflow.keras.utils.to_categorical`.\n",
    "   - One-hot encoding is necessary for the softmax output layer of the CNN.\n",
    "\n",
    "8. **Train-Validation Split:**\n",
    "   - Split the training data into training and validation sets using `train_test_split`.\n",
    "   - This allows us to evaluate the model on unseen data during training.\n",
    "\n",
    "# 2. Model Building\n",
    "\n",
    "1. **Build the CNN Model:**\n",
    "   - Define a sequential model using `tensorflow.keras.models.Sequential`.\n",
    "   - Add Conv2D layers for feature extraction, each followed by a MaxPooling2D layer to reduce spatial dimensions.\n",
    "   - Add Dropout layers to prevent overfitting by randomly setting a fraction of input units to 0.\n",
    "   - Flatten the output from the convolutional layers to feed into dense layers.\n",
    "   - Add Dense layers, with the final Dense layer having 10 units (one for each digit) and a softmax activation function for classification.\n",
    "\n",
    "2. **Compile the Model:**\n",
    "   - Compile the model with the Adam optimizer, categorical crossentropy loss, and accuracy as the metric.\n",
    "   - The Adam optimizer is chosen for its efficiency and effectiveness.\n",
    "   - Categorical crossentropy is used because it is a multi-class classification problem.\n",
    "\n",
    "# 3. Model Training and Evaluation\n",
    "\n",
    "1. **Train the Model:**\n",
    "   - Fit the model on the training data using the `fit` method.\n",
    "   - Provide the validation data for evaluation during training.\n",
    "   - Set the number of epochs and batch size.\n",
    "\n",
    "2. **Evaluate the Model:**\n",
    "   - Evaluate the model on the validation data using the `evaluate` method.\n",
    "   - This provides the validation loss and accuracy.\n",
    "\n",
    "3. **Visualize Training Progress:**\n",
    "   - Plot the training and validation accuracy using `matplotlib.pyplot.plot`.\n",
    "   - Plot the training and validation loss.\n",
    "   - These plots help in understanding the model’s learning progress and whether it is overfitting or underfitting.\n",
    "\n",
    "# 4. Generate Predictions and Submission File\n",
    "\n",
    "1. **Predict on Test Data:**\n",
    "   - Use the trained model to predict labels for the test data using the `predict` method.\n",
    "   - Convert the predictions to class labels using `np.argmax`.\n",
    "\n",
    "2. **Create Submission File:**\n",
    "   - Create a DataFrame with ImageId and the predicted labels.\n",
    "   - Save the DataFrame to a CSV file in the required format using `to_csv`.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0f80bc1b",
   "metadata": {
    "papermill": {
     "duration": 0.008429,
     "end_time": "2024-08-02T03:24:24.625446",
     "exception": false,
     "start_time": "2024-08-02T03:24:24.617017",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# 1. Data Analysis and Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "2445231e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:24.645712Z",
     "iopub.status.busy": "2024-08-02T03:24:24.644352Z",
     "iopub.status.idle": "2024-08-02T03:24:43.081053Z",
     "shell.execute_reply": "2024-08-02T03:24:43.079822Z"
    },
    "papermill": {
     "duration": 18.449748,
     "end_time": "2024-08-02T03:24:43.083942",
     "exception": false,
     "start_time": "2024-08-02T03:24:24.634194",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-08-02 03:24:30.187640: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered\n",
      "2024-08-02 03:24:30.187782: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered\n",
      "2024-08-02 03:24:30.368167: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n"
     ]
    }
   ],
   "source": [
    "# Load and preprocess MNIST dataset for QCNN\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "from tensorflow.keras.utils import to_categorical"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "374c7ce2",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:43.104304Z",
     "iopub.status.busy": "2024-08-02T03:24:43.103441Z",
     "iopub.status.idle": "2024-08-02T03:24:49.837437Z",
     "shell.execute_reply": "2024-08-02T03:24:49.836249Z"
    },
    "papermill": {
     "duration": 6.747217,
     "end_time": "2024-08-02T03:24:49.840388",
     "exception": false,
     "start_time": "2024-08-02T03:24:43.093171",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Load the data\n",
    "# It contains 60,000 training images and 10,000 testing images, \n",
    "# each of which is a grayscale image of size 28×28 pixels.\n",
    "train_df = pd.read_csv('/kaggle/input/digit-recognizer/train.csv')\n",
    "test_df = pd.read_csv('/kaggle/input/digit-recognizer/test.csv')\n",
    "sample_submission = pd.read_csv('/kaggle/input/digit-recognizer/sample_submission.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c43f45b1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:49.860997Z",
     "iopub.status.busy": "2024-08-02T03:24:49.860534Z",
     "iopub.status.idle": "2024-08-02T03:24:49.881988Z",
     "shell.execute_reply": "2024-08-02T03:24:49.880683Z"
    },
    "papermill": {
     "duration": 0.03537,
     "end_time": "2024-08-02T03:24:49.885266",
     "exception": false,
     "start_time": "2024-08-02T03:24:49.849896",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   label  pixel0  pixel1  pixel2  pixel3  pixel4  pixel5  pixel6  pixel7  \\\n",
      "0      1       0       0       0       0       0       0       0       0   \n",
      "1      0       0       0       0       0       0       0       0       0   \n",
      "2      1       0       0       0       0       0       0       0       0   \n",
      "3      4       0       0       0       0       0       0       0       0   \n",
      "4      0       0       0       0       0       0       0       0       0   \n",
      "\n",
      "   pixel8  ...  pixel774  pixel775  pixel776  pixel777  pixel778  pixel779  \\\n",
      "0       0  ...         0         0         0         0         0         0   \n",
      "1       0  ...         0         0         0         0         0         0   \n",
      "2       0  ...         0         0         0         0         0         0   \n",
      "3       0  ...         0         0         0         0         0         0   \n",
      "4       0  ...         0         0         0         0         0         0   \n",
      "\n",
      "   pixel780  pixel781  pixel782  pixel783  \n",
      "0         0         0         0         0  \n",
      "1         0         0         0         0  \n",
      "2         0         0         0         0  \n",
      "3         0         0         0         0  \n",
      "4         0         0         0         0  \n",
      "\n",
      "[5 rows x 785 columns]\n"
     ]
    }
   ],
   "source": [
    "# Display the first few rows of the training data\n",
    "print(train_df.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "82076462",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:49.907971Z",
     "iopub.status.busy": "2024-08-02T03:24:49.906417Z",
     "iopub.status.idle": "2024-08-02T03:24:49.963512Z",
     "shell.execute_reply": "2024-08-02T03:24:49.961685Z"
    },
    "papermill": {
     "duration": 0.071453,
     "end_time": "2024-08-02T03:24:49.966501",
     "exception": false,
     "start_time": "2024-08-02T03:24:49.895048",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "# Check for missing values\n",
    "print(train_df.isnull().sum().sum())\n",
    "print(test_df.isnull().sum().sum())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "341bfb52",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:49.987223Z",
     "iopub.status.busy": "2024-08-02T03:24:49.986747Z",
     "iopub.status.idle": "2024-08-02T03:24:50.008977Z",
     "shell.execute_reply": "2024-08-02T03:24:50.007656Z"
    },
    "papermill": {
     "duration": 0.035821,
     "end_time": "2024-08-02T03:24:50.011727",
     "exception": false,
     "start_time": "2024-08-02T03:24:49.975906",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "label\n",
       "0    4132\n",
       "1    4684\n",
       "2    4177\n",
       "3    4351\n",
       "4    4072\n",
       "5    3795\n",
       "6    4137\n",
       "7    4401\n",
       "8    4063\n",
       "9    4188\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df['label'].value_counts().sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "e6e15fc0",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:50.032908Z",
     "iopub.status.busy": "2024-08-02T03:24:50.032456Z",
     "iopub.status.idle": "2024-08-02T03:24:50.392667Z",
     "shell.execute_reply": "2024-08-02T03:24:50.391231Z"
    },
    "papermill": {
     "duration": 0.374625,
     "end_time": "2024-08-02T03:24:50.395794",
     "exception": false,
     "start_time": "2024-08-02T03:24:50.021169",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualize the distribution of labels\n",
    "sns.countplot(x=train_df['label'])\n",
    "plt.title('Distribution of Labels in Training Set')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "fec0b489",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:50.417567Z",
     "iopub.status.busy": "2024-08-02T03:24:50.417114Z",
     "iopub.status.idle": "2024-08-02T03:24:50.533780Z",
     "shell.execute_reply": "2024-08-02T03:24:50.532382Z"
    },
    "papermill": {
     "duration": 0.130844,
     "end_time": "2024-08-02T03:24:50.536683",
     "exception": false,
     "start_time": "2024-08-02T03:24:50.405839",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Separate features and labels\n",
    "X = train_df.drop(columns=['label']).values\n",
    "y = train_df['label'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "21d194aa",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:50.558425Z",
     "iopub.status.busy": "2024-08-02T03:24:50.558024Z",
     "iopub.status.idle": "2024-08-02T03:24:50.850667Z",
     "shell.execute_reply": "2024-08-02T03:24:50.849045Z"
    },
    "papermill": {
     "duration": 0.306844,
     "end_time": "2024-08-02T03:24:50.853550",
     "exception": false,
     "start_time": "2024-08-02T03:24:50.546706",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7a2f0aefc1f0>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X[0].reshape(28,28))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a4588545",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:50.876176Z",
     "iopub.status.busy": "2024-08-02T03:24:50.875700Z",
     "iopub.status.idle": "2024-08-02T03:24:51.096239Z",
     "shell.execute_reply": "2024-08-02T03:24:51.094410Z"
    },
    "papermill": {
     "duration": 0.235244,
     "end_time": "2024-08-02T03:24:51.099166",
     "exception": false,
     "start_time": "2024-08-02T03:24:50.863922",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Normalize the pixel values\n",
    "X = X / 255.0\n",
    "test_data = test_df.values / 255.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "429c3299",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:51.122518Z",
     "iopub.status.busy": "2024-08-02T03:24:51.122029Z",
     "iopub.status.idle": "2024-08-02T03:24:51.128811Z",
     "shell.execute_reply": "2024-08-02T03:24:51.127453Z"
    },
    "papermill": {
     "duration": 0.021257,
     "end_time": "2024-08-02T03:24:51.131570",
     "exception": false,
     "start_time": "2024-08-02T03:24:51.110313",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Reshape data to fit the model input (28x28x1)\n",
    "X = X.reshape(-1, 28, 28, 1)\n",
    "test_data = test_data.reshape(-1, 28, 28, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "1340c264",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:51.154060Z",
     "iopub.status.busy": "2024-08-02T03:24:51.153623Z",
     "iopub.status.idle": "2024-08-02T03:24:51.162696Z",
     "shell.execute_reply": "2024-08-02T03:24:51.161330Z"
    },
    "papermill": {
     "duration": 0.023575,
     "end_time": "2024-08-02T03:24:51.165532",
     "exception": false,
     "start_time": "2024-08-02T03:24:51.141957",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 0, 1, ..., 7, 6, 9])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "878590af",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:51.188359Z",
     "iopub.status.busy": "2024-08-02T03:24:51.187909Z",
     "iopub.status.idle": "2024-08-02T03:24:51.195538Z",
     "shell.execute_reply": "2024-08-02T03:24:51.194178Z"
    },
    "papermill": {
     "duration": 0.022315,
     "end_time": "2024-08-02T03:24:51.198352",
     "exception": false,
     "start_time": "2024-08-02T03:24:51.176037",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Convert labels to categorical format\n",
    "# 将整数标签转换为二进制类别矩阵\n",
    "y = to_categorical(y, num_classes=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "1c0452c6",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:51.221196Z",
     "iopub.status.busy": "2024-08-02T03:24:51.220755Z",
     "iopub.status.idle": "2024-08-02T03:24:51.230301Z",
     "shell.execute_reply": "2024-08-02T03:24:51.228882Z"
    },
    "papermill": {
     "duration": 0.024287,
     "end_time": "2024-08-02T03:24:51.233134",
     "exception": false,
     "start_time": "2024-08-02T03:24:51.208847",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 1., 0., ..., 0., 0., 0.],\n",
       "       [1., 0., 0., ..., 0., 0., 0.],\n",
       "       [0., 1., 0., ..., 0., 0., 0.],\n",
       "       ...,\n",
       "       [0., 0., 0., ..., 1., 0., 0.],\n",
       "       [0., 0., 0., ..., 0., 0., 0.],\n",
       "       [0., 0., 0., ..., 0., 0., 1.]])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "c490ab49",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:51.257154Z",
     "iopub.status.busy": "2024-08-02T03:24:51.256763Z",
     "iopub.status.idle": "2024-08-02T03:24:51.704616Z",
     "shell.execute_reply": "2024-08-02T03:24:51.703244Z"
    },
    "papermill": {
     "duration": 0.463607,
     "end_time": "2024-08-02T03:24:51.707754",
     "exception": false,
     "start_time": "2024-08-02T03:24:51.244147",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Split the training data into training and validation sets\n",
    "X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ce1dcaf9",
   "metadata": {
    "papermill": {
     "duration": 0.010316,
     "end_time": "2024-08-02T03:24:51.728693",
     "exception": false,
     "start_time": "2024-08-02T03:24:51.718377",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# 2. Quantum Convolutional Neural Network (QCNN)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "65e37768",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:51.752143Z",
     "iopub.status.busy": "2024-08-02T03:24:51.751671Z",
     "iopub.status.idle": "2024-08-02T03:24:51.757193Z",
     "shell.execute_reply": "2024-08-02T03:24:51.755856Z"
    },
    "papermill": {
     "duration": 0.020282,
     "end_time": "2024-08-02T03:24:51.759843",
     "exception": false,
     "start_time": "2024-08-02T03:24:51.739561",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# !pip install pennylane"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "5e767211",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:51.783721Z",
     "iopub.status.busy": "2024-08-02T03:24:51.782427Z",
     "iopub.status.idle": "2024-08-02T03:24:51.789799Z",
     "shell.execute_reply": "2024-08-02T03:24:51.787938Z"
    },
    "papermill": {
     "duration": 0.022446,
     "end_time": "2024-08-02T03:24:51.792989",
     "exception": false,
     "start_time": "2024-08-02T03:24:51.770543",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from tensorflow.keras import layers, models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "06c8a287",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:51.817009Z",
     "iopub.status.busy": "2024-08-02T03:24:51.815721Z",
     "iopub.status.idle": "2024-08-02T03:24:51.954928Z",
     "shell.execute_reply": "2024-08-02T03:24:51.953421Z"
    },
    "papermill": {
     "duration": 0.154299,
     "end_time": "2024-08-02T03:24:51.957974",
     "exception": false,
     "start_time": "2024-08-02T03:24:51.803675",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.10/site-packages/keras/src/layers/convolutional/base_conv.py:107: UserWarning: Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.\n",
      "  super().__init__(activity_regularizer=activity_regularizer, **kwargs)\n"
     ]
    }
   ],
   "source": [
    "# Title: Build and compile the Convolutional Neural Network\n",
    "from tensorflow.keras import layers, models\n",
    "\n",
    "# Build the CNN model\n",
    "# Conv2D 层是二维卷积层，用于在图像上应用卷积操作。卷积操作通过在输入图像上滑动一个小的窗口（称为卷积核或滤波器），并计算窗口内像素值与卷积核权重的点积，来提取图像的特征。\n",
    "# MaxPooling2D 层是二维最大池化层，用于对输入特征图进行下采样。池化操作通过在输入特征图上滑动一个窗口，并取窗口内最大值作为输出，来减少特征图的尺寸。\n",
    "cnn_model = models.Sequential([\n",
    "    # 卷积核：通过权重随机初始化，并通过反向传播修正\n",
    "    layers.Conv2D(32, (3, 3), activation='relu', input_shape=(28, 28, 1)),\n",
    "    layers.MaxPooling2D((2, 2)),\n",
    "    layers.Conv2D(64, (3, 3), activation='relu'),\n",
    "    layers.MaxPooling2D((2, 2)),\n",
    "    layers.Conv2D(64, (3, 3), activation='relu'),\n",
    "    layers.Flatten(),\n",
    "    layers.Dense(64, activation='relu'),\n",
    "    layers.Dense(10, activation='softmax')\n",
    "])\n",
    "\n",
    "# Compile the model\n",
    "cnn_model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ed4ea3cb",
   "metadata": {
    "papermill": {
     "duration": 0.010468,
     "end_time": "2024-08-02T03:24:51.979184",
     "exception": false,
     "start_time": "2024-08-02T03:24:51.968716",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# 3. Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "02663a70",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:24:52.002782Z",
     "iopub.status.busy": "2024-08-02T03:24:52.002317Z",
     "iopub.status.idle": "2024-08-02T03:27:38.768012Z",
     "shell.execute_reply": "2024-08-02T03:27:38.766547Z"
    },
    "papermill": {
     "duration": 166.780808,
     "end_time": "2024-08-02T03:27:38.770883",
     "exception": false,
     "start_time": "2024-08-02T03:24:51.990075",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "\u001b[1m525/525\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m19s\u001b[0m 32ms/step - accuracy: 0.8128 - loss: 0.5869 - val_accuracy: 0.9746 - val_loss: 0.0781\n",
      "Epoch 2/10\n",
      "\u001b[1m525/525\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 30ms/step - accuracy: 0.9794 - loss: 0.0678 - val_accuracy: 0.9840 - val_loss: 0.0505\n",
      "Epoch 3/10\n",
      "\u001b[1m525/525\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 31ms/step - accuracy: 0.9835 - loss: 0.0469 - val_accuracy: 0.9837 - val_loss: 0.0517\n",
      "Epoch 4/10\n",
      "\u001b[1m525/525\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 30ms/step - accuracy: 0.9902 - loss: 0.0330 - val_accuracy: 0.9854 - val_loss: 0.0461\n",
      "Epoch 5/10\n",
      "\u001b[1m525/525\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 31ms/step - accuracy: 0.9913 - loss: 0.0286 - val_accuracy: 0.9879 - val_loss: 0.0379\n",
      "Epoch 6/10\n",
      "\u001b[1m525/525\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 30ms/step - accuracy: 0.9937 - loss: 0.0192 - val_accuracy: 0.9890 - val_loss: 0.0362\n",
      "Epoch 7/10\n",
      "\u001b[1m525/525\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 31ms/step - accuracy: 0.9943 - loss: 0.0177 - val_accuracy: 0.9876 - val_loss: 0.0367\n",
      "Epoch 8/10\n",
      "\u001b[1m525/525\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 30ms/step - accuracy: 0.9945 - loss: 0.0170 - val_accuracy: 0.9893 - val_loss: 0.0377\n",
      "Epoch 9/10\n",
      "\u001b[1m525/525\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m16s\u001b[0m 31ms/step - accuracy: 0.9958 - loss: 0.0144 - val_accuracy: 0.9886 - val_loss: 0.0407\n",
      "Epoch 10/10\n",
      "\u001b[1m525/525\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m20s\u001b[0m 30ms/step - accuracy: 0.9965 - loss: 0.0114 - val_accuracy: 0.9888 - val_loss: 0.0351\n"
     ]
    }
   ],
   "source": [
    "# Title: Train, evaluate, and visualize the CNN model\n",
    "# 准确率提供了模型预测正确率的直观度量，而损失函数则提供了模型性能的连续度量，用于指导模型训练过程中的参数调整\n",
    "history = cnn_model.fit(X_train, y_train, epochs=10, batch_size=64, validation_data=(X_val, y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "4ecfc883",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:27:39.325615Z",
     "iopub.status.busy": "2024-08-02T03:27:39.325154Z",
     "iopub.status.idle": "2024-08-02T03:27:42.019897Z",
     "shell.execute_reply": "2024-08-02T03:27:42.018336Z"
    },
    "papermill": {
     "duration": 2.991366,
     "end_time": "2024-08-02T03:27:42.022815",
     "exception": false,
     "start_time": "2024-08-02T03:27:39.031449",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m263/263\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 7ms/step - accuracy: 0.9873 - loss: 0.0417\n",
      "Validation loss: 0.0351371131837368\n",
      "Validation accuracy: 0.9888095259666443\n"
     ]
    }
   ],
   "source": [
    "# Evaluate the model\n",
    "val_loss, val_accuracy = cnn_model.evaluate(X_val, y_val)\n",
    "print(f'Validation loss: {val_loss}')\n",
    "print(f'Validation accuracy: {val_accuracy}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "ce95f94f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:27:42.563730Z",
     "iopub.status.busy": "2024-08-02T03:27:42.563283Z",
     "iopub.status.idle": "2024-08-02T03:27:43.147386Z",
     "shell.execute_reply": "2024-08-02T03:27:43.146184Z"
    },
    "papermill": {
     "duration": 0.857872,
     "end_time": "2024-08-02T03:27:43.150162",
     "exception": false,
     "start_time": "2024-08-02T03:27:42.292290",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot training & validation accuracy values\n",
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['val_accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Validation'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "# Plot training & validation loss values\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Validation'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "270ffb7a",
   "metadata": {
    "papermill": {
     "duration": 0.269465,
     "end_time": "2024-08-02T03:27:43.688959",
     "exception": false,
     "start_time": "2024-08-02T03:27:43.419494",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# 4. Submission File Generation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "14dfaac7",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-08-02T03:27:44.223948Z",
     "iopub.status.busy": "2024-08-02T03:27:44.223482Z",
     "iopub.status.idle": "2024-08-02T03:27:50.722041Z",
     "shell.execute_reply": "2024-08-02T03:27:50.720410Z"
    },
    "papermill": {
     "duration": 6.77041,
     "end_time": "2024-08-02T03:27:50.724922",
     "exception": false,
     "start_time": "2024-08-02T03:27:43.954512",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m875/875\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 7ms/step\n",
      "   ImageId  Label\n",
      "0        1      2\n",
      "1        2      0\n",
      "2        3      9\n",
      "3        4      0\n",
      "4        5      3\n"
     ]
    }
   ],
   "source": [
    "# Title: Generate predictions and create submission file\n",
    "predictions = cnn_model.predict(test_data)\n",
    "predicted_labels = np.argmax(predictions, axis=1)\n",
    "\n",
    "submission_df = pd.DataFrame({'ImageId': np.arange(1, len(predicted_labels) + 1), 'Label': predicted_labels})\n",
    "submission_df.to_csv('submission.csv', index=False)\n",
    "\n",
    "print(submission_df.head())\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "304149e2",
   "metadata": {
    "papermill": {
     "duration": 0.279117,
     "end_time": "2024-08-02T03:27:51.286395",
     "exception": false,
     "start_time": "2024-08-02T03:27:51.007278",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Summary\n",
    "This process involves loading and preprocessing the MNIST dataset, building a hybrid Quantum Convolutional Neural Network (QCNN) using Pennylane and TensorFlow, training and evaluating the model, visualizing the training history, and generating a submission file. Each step is crucial for achieving high accuracy in classifying handwritten digits and effectively leveraging quantum computing techniques."
   ]
  }
 ],
 "metadata": {
  "kaggle": {
   "accelerator": "none",
   "dataSources": [
    {
     "databundleVersionId": 861823,
     "sourceId": 3004,
     "sourceType": "competition"
    }
   ],
   "dockerImageVersionId": 30746,
   "isGpuEnabled": false,
   "isInternetEnabled": false,
   "language": "python",
   "sourceType": "notebook"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.10.13"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 212.179629,
   "end_time": "2024-08-02T03:27:53.462323",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2024-08-02T03:24:21.282694",
   "version": "2.5.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}