{ "nbformat": 4, "nbformat_minor": 0, "metadata": { "colab": { "provenance": [] }, "kernelspec": { "name": "python3", "display_name": "Python 3" }, "language_info": { "name": "python" } }, "cells": [ { "cell_type": "markdown", "source": [ "# 程式特色\n", "1.使用MNIST和EMINIST及SVHN三個DATASET\n", "\n", "2.檢測到已處理好的資料集，直接從 Google Drive 秒速載入！\n", "\n", "3.有googel drive就不再下載data set訓練了\n", "\n", "4使用google drive以加快已訓練好的模型權重載入\n", "\n", "5.使用gradio顯示" ], "metadata": { "id": "UHDi7CALC1bj" } }, { "cell_type": "markdown", "source": [ "# 儲存格 1：環境設定、掛載 Drive 與資料集快取處理\n", "這段程式碼會在第一次執行時花點時間下載並轉換資料，但最後會把處理好的 Numpy 陣列存到你的 Google Drive。下次執行時，它就會直接從 Drive 載入，只要幾秒鐘！\n", "\n", "Python" ], "metadata": { "id": "C7vRLhsYIsiN" } }, { "cell_type": "code", "execution_count": 1, "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "ADOFujykH9rN", "outputId": "4c97169a-65d3-43aa-f06f-a1d8e4cff159" }, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "⏳ 正在掛載 Google Drive...\n", "Mounted at /content/drive\n", "\n", "⏳ 步驟 1: 載入或準備融合資料集...\n", "✅ 檢測到已處理好的資料集，直接從 Google Drive 秒速載入！\n", "\n", "✅ 資料準備完畢！\n", " 融合訓練集數量: 140000 筆\n", " 融合測試集數量: 30000 筆\n" ] } ], "source": [ "import os\n", "import io\n", "import numpy as np\n", "import pandas as pd\n", "import matplotlib.pyplot as plt\n", "from PIL import Image, ImageOps\n", "from sklearn.metrics import classification_report\n", "\n", "import tensorflow as tf\n", "import tensorflow_datasets as tfds\n", "from tensorflow.keras import layers, models\n", "from google.colab import output, files, drive\n", "from IPython.display import HTML, display\n", "\n", "# 1. 掛載 Google Drive\n", "print(\"⏳ 正在掛載 Google Drive...\")\n", "drive.mount('/content/drive')\n", "\n", "# 2. 設定儲存路徑 (在你的 Google Drive 中建立一個專屬資料夾)\n", "DRIVE_DIR = '/content/drive/MyDrive/Digit_Recognition_Colab'\n", "os.makedirs(DRIVE_DIR, exist_ok=True)\n", "\n", "DATA_PATH = os.path.join(DRIVE_DIR, 'processed_dataset.npz')\n", "MODEL_PATH = os.path.join(DRIVE_DIR, 'mnist_emnist_svhn_model.keras') # 使用官方推薦的 .keras 格式\n", "\n", "print(\"\\n⏳ 步驟 1: 載入或準備融合資料集...\")\n", "\n", "# 3. 檢查是否已經有處理好的資料集\n", "if os.path.exists(DATA_PATH):\n", " print(\"✅ 檢測到已處理好的資料集，直接從 Google Drive 秒速載入！\")\n", " data = np.load(DATA_PATH)\n", " x_train, y_train = data['x_train'], data['y_train']\n", " x_test, y_test = data['x_test'], data['y_test']\n", "else:\n", " print(\"⚠️ 未檢測到快取資料，開始下載與轉換資料集 (此步驟較久，但只需執行一次)...\")\n", "\n", " # 載入 MNIST\n", " (x_m_train, y_m_train), (x_m_test, y_m_test) = tf.keras.datasets.mnist.load_data()\n", " x_m_train = x_m_train.reshape(-1, 28, 28, 1) / 255.0\n", " x_m_test = x_m_test.reshape(-1, 28, 28, 1) / 255.0\n", "\n", " # 載入 EMNIST\n", " emnist_train = tfds.load(\"emnist/digits\", split=\"train\", as_supervised=True)\n", " emnist_test = tfds.load(\"emnist/digits\", split=\"test\", as_supervised=True)\n", "\n", " x_e_train, y_e_train = [], []\n", " for img, label in emnist_train.take(40000):\n", " x_e_train.append(tf.image.transpose(img).numpy())\n", " y_e_train.append(label.numpy())\n", " x_e_train, y_e_train = np.array(x_e_train) / 255.0, np.array(y_e_train)\n", "\n", " x_e_test, y_e_test = [], []\n", " for img, label in emnist_test.take(10000):\n", " x_e_test.append(tf.image.transpose(img).numpy())\n", " y_e_test.append(label.numpy())\n", " x_e_test, y_e_test = np.array(x_e_test) / 255.0, np.array(y_e_test)\n", "\n", " # 載入 SVHN\n", " svhn_train = tfds.load(\"svhn_cropped\", split=\"train\", as_supervised=True)\n", " svhn_test = tfds.load(\"svhn_cropped\", split=\"test\", as_supervised=True)\n", "\n", " x_s_train, y_s_train = [], []\n", " for img, label in svhn_train.take(40000):\n", " img_res = tf.image.resize(tf.image.rgb_to_grayscale(img), [28, 28])\n", " x_s_train.append(img_res.numpy())\n", " y_s_train.append(label.numpy())\n", " x_s_train, y_s_train = np.array(x_s_train) / 255.0, np.array(y_s_train)\n", "\n", " x_s_test, y_s_test = [], []\n", " for img, label in svhn_test.take(10000):\n", " img_res = tf.image.resize(tf.image.rgb_to_grayscale(img), [28, 28])\n", " x_s_test.append(img_res.numpy())\n", " y_s_test.append(label.numpy())\n", " x_s_test, y_s_test = np.array(x_s_test) / 255.0, np.array(y_s_test)\n", "\n", " # 合併資料\n", " x_train = np.concatenate([x_m_train, x_e_train, x_s_train], axis=0)\n", " y_train = np.concatenate([y_m_train, y_e_train, y_s_train], axis=0)\n", " x_test = np.concatenate([x_m_test, x_e_test, x_s_test], axis=0)\n", " y_test = np.concatenate([y_m_test, y_e_test, y_s_test], axis=0)\n", "\n", " print(\"\\n💾 正在將處理好的融合資料集壓縮並儲存到 Google Drive...\")\n", " np.savez_compressed(DATA_PATH, x_train=x_train, y_train=y_train, x_test=x_test, y_test=y_test)\n", " print(\"✅ 資料集儲存完畢！下次重開機將直接載入。\")\n", "\n", "print(f\"\\n✅ 資料準備完畢！\")\n", "print(f\" 融合訓練集數量: {x_train.shape[0]} 筆\")\n", "print(f\" 融合測試集數量: {x_test.shape[0]} 筆\")" ] }, { "cell_type": "markdown", "source": [ "# 儲存格 2：模型建構與快取訓練\n", "\n", "這段加入了 Keras 新版的 Input 層以消除警告，並把模型永久保存在你的雲端硬碟中。" ], "metadata": { "id": "8svSxROwIzrk" } }, { "cell_type": "code", "source": [ "print(\"⏳ 步驟 2: 模型建構與訓練...\\n\")\n", "\n", "if os.path.exists(MODEL_PATH):\n", " print(\"✅ 檢測到已存在 Google Drive 的模型，直接載入，跳過訓練！\")\n", " model = tf.keras.models.load_model(MODEL_PATH)\n", "else:\n", " print(\"⚠️ 尚未建立模型，開始訓練...\")\n", " model = models.Sequential([\n", " layers.Input(shape=(28, 28, 1)), # 修正警告：使用獨立的 Input 層\n", " layers.Conv2D(32, (3, 3), activation='relu'),\n", " layers.BatchNormalization(),\n", " layers.MaxPooling2D((2, 2)),\n", " layers.Conv2D(64, (3, 3), activation='relu'),\n", " layers.BatchNormalization(),\n", " layers.MaxPooling2D((2, 2)),\n", " layers.Flatten(),\n", " layers.Dense(128, activation='relu'),\n", " layers.Dense(10, activation='softmax')\n", " ])\n", "\n", " model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])\n", "\n", " # 訓練模型 (若有開啟 GPU，這裡會飛快)\n", " history = model.fit(x_train, y_train, epochs=5, batch_size=128,\n", " validation_data=(x_test, y_test), verbose=1)\n", "\n", " print(\"\\n💾 正在將訓練好的模型儲存到 Google Drive...\")\n", " model.save(MODEL_PATH)\n", " print(f\"✅ 模型已安全儲存至：{MODEL_PATH}\")" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "MX_Jw3nYIKrz", "outputId": "2ac57769-b951-4256-804b-e392f87e8323" }, "execution_count": 2, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "⏳ 步驟 2: 模型建構與訓練...\n", "\n", "✅ 檢測到已存在 Google Drive 的模型，直接載入，跳過訓練！\n" ] } ] }, { "cell_type": "markdown", "source": [ "# 儲存格 3：產生訓練分析報告與驗證數據\n", "\n", "1.正確率\n", "\n", "2.召回率\n", "\n", "3.f1參數" ], "metadata": { "id": "I2qS_xn9I8k0" } }, { "cell_type": "code", "source": [ "print(\"⏳ 步驟 3: 產生訓練分析報告與驗證數據...\\n\")\n", "\n", "# 繪製曲線 (只有剛剛訓練完才會有 history 變數)\n", "if 'history' in locals():\n", " fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 5))\n", "\n", " ax1.plot(history.history['accuracy'], label='Training Accuracy', color='#1a73e8', marker='o')\n", " ax1.plot(history.history['val_accuracy'], label='Validation Accuracy', color='#34a853', marker='x')\n", " ax1.set_title('Model Accuracy Curve', fontsize=14, fontweight='bold')\n", " ax1.set_xlabel('Epochs')\n", " ax1.set_ylabel('Accuracy')\n", " ax1.legend()\n", " ax1.grid(True, linestyle='--')\n", "\n", " ax2.plot(history.history['loss'], label='Training Loss', color='#d93025', marker='o')\n", " ax2.plot(history.history['val_loss'], label='Validation Loss', color='#fbbc04', marker='x')\n", " ax2.set_title('Model Loss Curve', fontsize=14, fontweight='bold')\n", " ax2.set_xlabel('Epochs')\n", " ax2.set_ylabel('Loss')\n", " ax2.legend()\n", " ax2.grid(True, linestyle='--')\n", "\n", " plt.show()\n", "else:\n", " print(\"💡 提示：模型為直接從 Google Drive 載入，本次未重新訓練，無歷史軌跡曲線。\")\n", "\n", "print(\"\\n\" + \"=\"*60 + \"\\n 正在計算測試集 0-9 每個數字的詳細分析指標... \\n\" + \"=\"*60)\n", "\n", "# 預測並計算指標\n", "predictions = model.predict(x_test, batch_size=256, verbose=0)\n", "y_pred = np.argmax(predictions, axis=1)\n", "report_dict = classification_report(y_test, y_pred, output_dict=True)\n", "\n", "analysis_data = []\n", "for i in range(10):\n", " label = str(i)\n", " analysis_data.append({\n", " \"數字\": i,\n", " \"精確率 (Precision)\": f\"{report_dict[label]['precision']*100:.2f}%\",\n", " \"召回率 (Recall)\": f\"{report_dict[label]['recall']*100:.2f}%\",\n", " \"F1-分數 (F1-Score)\": f\"{report_dict[label]['f1-score']*100:.2f}%\",\n", " \"測試樣本總數\": int(report_dict[label]['support'])\n", " })\n", "\n", "df_analysis = pd.DataFrame(analysis_data)\n", "display(HTML(\"

📊 每個數字的分類效能詳細分析表

\"))\n", "display(df_analysis)\n", "\n", "print(f\"\\n✅ 總體測試集平均準確率 (Overall Accuracy): {report_dict['accuracy']*100:.2f}%\")" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 559 }, "id": "oUuMcqlZIPd9", "outputId": "ec65b42e-839b-4381-a1f4-68ce94caeb36" }, "execution_count": 3, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "⏳ 步驟 3: 產生訓練分析報告與驗證數據...\n", "\n", "💡 提示：模型為直接從 Google Drive 載入，本次未重新訓練，無歷史軌跡曲線。\n", "\n", "============================================================\n", " 正在計算測試集 0-9 每個數字的詳細分析指標... \n", "============================================================\n" ] }, { "output_type": "display_data", "data": { "text/plain": [ "" ], "text/html": [ "

📊 每個數字的分類效能詳細分析表

" ] }, "metadata": {} }, { "output_type": "display_data", "data": { "text/plain": [ " 數字 精確率 (Precision) 召回率 (Recall) F1-分數 (F1-Score) 測試樣本總數\n", "0 0 95.40% 95.44% 95.42% 2630\n", "1 1 95.57% 95.73% 95.65% 4171\n", "2 2 94.60% 95.89% 95.24% 3580\n", "3 3 91.12% 93.72% 92.40% 3088\n", "4 4 96.93% 94.77% 95.84% 3000\n", "5 5 96.96% 92.54% 94.70% 2789\n", "6 6 93.35% 95.24% 94.28% 2710\n", "7 7 97.36% 94.55% 95.93% 2842\n", "8 8 95.40% 92.04% 93.69% 2589\n", "9 9 89.88% 95.66% 92.68% 2601" ], "text/html": [ "\n", "
\n", "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
數字精確率 (Precision)召回率 (Recall)F1-分數 (F1-Score)測試樣本總數
0095.40%95.44%95.42%2630
1195.57%95.73%95.65%4171
2294.60%95.89%95.24%3580
3391.12%93.72%92.40%3088
4496.93%94.77%95.84%3000
5596.96%92.54%94.70%2789
6693.35%95.24%94.28%2710
7797.36%94.55%95.93%2842
8895.40%92.04%93.69%2589
9989.88%95.66%92.68%2601
\n", "
\n", "
\n", "\n", "
\n", " \n", "\n", " \n", "\n", " \n", "
\n", "\n", "\n", "
\n", " \n", " \n", " \n", "
\n", "\n", "
\n", "
\n" ], "application/vnd.google.colaboratory.intrinsic+json": { "type": "dataframe", "variable_name": "df_analysis", "summary": "{\n \"name\": \"df_analysis\",\n \"rows\": 10,\n \"fields\": [\n {\n \"column\": \"\\u6578\\u5b57\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 3,\n \"min\": 0,\n \"max\": 9,\n \"num_unique_values\": 10,\n \"samples\": [\n 8,\n 1,\n 5\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"\\u7cbe\\u78ba\\u7387 (Precision)\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 9,\n \"samples\": [\n \"97.36%\",\n \"95.57%\",\n \"96.96%\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"\\u53ec\\u56de\\u7387 (Recall)\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 10,\n \"samples\": [\n \"92.04%\",\n \"95.73%\",\n \"92.54%\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"F1-\\u5206\\u6578 (F1-Score)\",\n \"properties\": {\n \"dtype\": \"string\",\n \"num_unique_values\": 10,\n \"samples\": [\n \"93.69%\",\n \"95.65%\",\n \"94.70%\"\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n },\n {\n \"column\": \"\\u6e2c\\u8a66\\u6a23\\u672c\\u7e3d\\u6578\",\n \"properties\": {\n \"dtype\": \"number\",\n \"std\": 509,\n \"min\": 2589,\n \"max\": 4171,\n \"num_unique_values\": 10,\n \"samples\": [\n 2589,\n 4171,\n 2789\n ],\n \"semantic_type\": \"\",\n \"description\": \"\"\n }\n }\n ]\n}" } }, "metadata": {} }, { "output_type": "stream", "name": "stdout", "text": [ "\n", "✅ 總體測試集平均準確率 (Overall Accuracy): 94.64%\n" ] } ] }, { "cell_type": "markdown", "source": [ "儲存格4 安裝gradio" ], "metadata": { "id": "PYEyZ90Ielag" } }, { "cell_type": "code", "source": [ "!pip install gradio" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "CyXVVo6ae1-f", "outputId": "d169e3c0-22c2-4c81-bc66-180f820095da" }, "execution_count": 4, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Requirement already satisfied: gradio in /usr/local/lib/python3.12/dist-packages (5.50.0)\n", "Requirement already satisfied: aiofiles<25.0,>=22.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (24.1.0)\n", "Requirement already satisfied: anyio<5.0,>=3.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (4.13.0)\n", "Requirement already satisfied: brotli>=1.1.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (1.2.0)\n", "Requirement already satisfied: fastapi<1.0,>=0.115.2 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.136.1)\n", "Requirement already satisfied: ffmpy in /usr/local/lib/python3.12/dist-packages (from gradio) (1.0.0)\n", "Requirement already satisfied: gradio-client==1.14.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (1.14.0)\n", "Requirement already satisfied: groovy~=0.1 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.1.2)\n", "Requirement already satisfied: httpx<1.0,>=0.24.1 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.28.1)\n", "Requirement already satisfied: huggingface-hub<2.0,>=0.33.5 in /usr/local/lib/python3.12/dist-packages (from gradio) (1.16.1)\n", "Requirement already satisfied: jinja2<4.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (3.1.6)\n", "Requirement already satisfied: markupsafe<4.0,>=2.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (3.0.3)\n", "Requirement already satisfied: numpy<3.0,>=1.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.0.2)\n", "Requirement already satisfied: orjson~=3.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (3.11.9)\n", "Requirement already satisfied: packaging in /usr/local/lib/python3.12/dist-packages (from gradio) (26.2)\n", "Requirement already satisfied: pandas<3.0,>=1.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.2.2)\n", "Requirement already satisfied: pillow<12.0,>=8.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (11.3.0)\n", "Requirement already satisfied: pydantic<=2.12.3,>=2.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.12.3)\n", "Requirement already satisfied: pydub in /usr/local/lib/python3.12/dist-packages (from gradio) (0.25.1)\n", "Requirement already satisfied: python-multipart>=0.0.18 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.0.29)\n", "Requirement already satisfied: pyyaml<7.0,>=5.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (6.0.3)\n", "Requirement already satisfied: ruff>=0.9.3 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.15.14)\n", "Requirement already satisfied: safehttpx<0.2.0,>=0.1.6 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.1.7)\n", "Requirement already satisfied: semantic-version~=2.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (2.10.0)\n", "Requirement already satisfied: starlette<1.0,>=0.40.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.52.1)\n", "Requirement already satisfied: tomlkit<0.14.0,>=0.12.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.13.3)\n", "Requirement already satisfied: typer<1.0,>=0.12 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.25.1)\n", "Requirement already satisfied: typing-extensions~=4.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (4.15.0)\n", "Requirement already satisfied: uvicorn>=0.14.0 in /usr/local/lib/python3.12/dist-packages (from gradio) (0.47.0)\n", "Requirement already satisfied: fsspec in /usr/local/lib/python3.12/dist-packages (from gradio-client==1.14.0->gradio) (2025.3.0)\n", "Requirement already satisfied: websockets<16.0,>=13.0 in /usr/local/lib/python3.12/dist-packages (from gradio-client==1.14.0->gradio) (15.0.1)\n", "Requirement already satisfied: idna>=2.8 in /usr/local/lib/python3.12/dist-packages (from anyio<5.0,>=3.0->gradio) (3.15)\n", "Requirement already satisfied: typing-inspection>=0.4.2 in /usr/local/lib/python3.12/dist-packages (from fastapi<1.0,>=0.115.2->gradio) (0.4.2)\n", "Requirement already satisfied: annotated-doc>=0.0.2 in /usr/local/lib/python3.12/dist-packages (from fastapi<1.0,>=0.115.2->gradio) (0.0.4)\n", "Requirement already satisfied: certifi in /usr/local/lib/python3.12/dist-packages (from httpx<1.0,>=0.24.1->gradio) (2026.5.20)\n", "Requirement already satisfied: httpcore==1.* in /usr/local/lib/python3.12/dist-packages (from httpx<1.0,>=0.24.1->gradio) (1.0.9)\n", "Requirement already satisfied: h11>=0.16 in /usr/local/lib/python3.12/dist-packages (from httpcore==1.*->httpx<1.0,>=0.24.1->gradio) (0.16.0)\n", "Requirement already satisfied: filelock>=3.10.0 in /usr/local/lib/python3.12/dist-packages (from huggingface-hub<2.0,>=0.33.5->gradio) (3.29.0)\n", "Requirement already satisfied: hf-xet<2.0.0,>=1.4.3 in /usr/local/lib/python3.12/dist-packages (from huggingface-hub<2.0,>=0.33.5->gradio) (1.5.0)\n", "Requirement already satisfied: tqdm>=4.42.1 in /usr/local/lib/python3.12/dist-packages (from huggingface-hub<2.0,>=0.33.5->gradio) (4.67.3)\n", "Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.12/dist-packages (from pandas<3.0,>=1.0->gradio) (2.9.0.post0)\n", "Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.12/dist-packages (from pandas<3.0,>=1.0->gradio) (2025.2)\n", "Requirement already satisfied: tzdata>=2022.7 in /usr/local/lib/python3.12/dist-packages (from pandas<3.0,>=1.0->gradio) (2026.2)\n", "Requirement already satisfied: annotated-types>=0.6.0 in /usr/local/lib/python3.12/dist-packages (from pydantic<=2.12.3,>=2.0->gradio) (0.7.0)\n", "Requirement already satisfied: pydantic-core==2.41.4 in /usr/local/lib/python3.12/dist-packages (from pydantic<=2.12.3,>=2.0->gradio) (2.41.4)\n", "Requirement already satisfied: click>=8.2.1 in /usr/local/lib/python3.12/dist-packages (from typer<1.0,>=0.12->gradio) (8.4.0)\n", "Requirement already satisfied: shellingham>=1.3.0 in /usr/local/lib/python3.12/dist-packages (from typer<1.0,>=0.12->gradio) (1.5.4)\n", "Requirement already satisfied: rich>=13.8.0 in /usr/local/lib/python3.12/dist-packages (from typer<1.0,>=0.12->gradio) (13.9.4)\n", "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.12/dist-packages (from python-dateutil>=2.8.2->pandas<3.0,>=1.0->gradio) (1.17.0)\n", "Requirement already satisfied: markdown-it-py>=2.2.0 in /usr/local/lib/python3.12/dist-packages (from rich>=13.8.0->typer<1.0,>=0.12->gradio) (4.2.0)\n", "Requirement already satisfied: pygments<3.0.0,>=2.13.0 in /usr/local/lib/python3.12/dist-packages (from rich>=13.8.0->typer<1.0,>=0.12->gradio) (2.20.0)\n", "Requirement already satisfied: mdurl~=0.1 in /usr/local/lib/python3.12/dist-packages (from markdown-it-py>=2.2.0->rich>=13.8.0->typer<1.0,>=0.12->gradio) (0.1.2)\n" ] } ] }, { "cell_type": "markdown", "source": [ "# 儲存格 5：啟動手寫圖片上傳辨識\n" ], "metadata": { "id": "0ks_elSvJFAc" } }, { "cell_type": "code", "source": [ "import gradio as gr\n", "from PIL import Image, ImageOps\n", "import numpy as np\n", "\n", "print(\"⏳ 步驟 4: 啟動手寫圖片上傳辨識介面 (Gradio)...\\\\n\")\n", "\n", "def process_image(img):\n", " \"\"\"處理圖片並回傳模型預測結果給 Gradio\"\"\"\n", " if img is None:\n", " return {\"請上傳圖片\": 0.0}\n", "\n", " # 確保圖片為灰階\n", " img = img.convert('L')\n", "\n", " # 顏色反轉 (原本為白底黑字 -> 黑底白字，符合模型訓練特徵)\n", " img_inv = ImageOps.invert(img)\n", "\n", " # 取得數字邊界框，去除多餘空白\n", " bbox = img_inv.getbbox()\n", "\n", " if bbox:\n", " # 裁切出數字部分\n", " img_final = img_inv.crop(bbox)\n", " # 增加邊框並縮放至 28x28\n", " img_final = ImageOps.expand(img_final, border=4, fill=0).resize((28, 28), Image.Resampling.LANCZOS)\n", " # 轉為模型需要的 Numpy 陣列並標準化\n", " img_array = np.array(img_final).reshape(1, 28, 28, 1) / 255.0\n", "\n", " # 模型預測\n", " res = model.predict(img_array, verbose=0)[0]\n", "\n", " # 回傳字典格式 {數字: 機率}，Gradio 會自動繪製出漂亮的信心度長條圖\n", " return {str(i): float(res[i]) for i in range(10)}\n", " else:\n", " return {\"圖片過於空白或無法辨識\": 0.0}\n", "\n", "# 建立 Gradio 網頁介面\n", "ui = gr.Interface(\n", " fn=process_image,\n", " inputs=gr.Image(type=\"pil\", image_mode=\"L\", label=\"上傳手寫圖片 (建議白底黑字)\"),\n", " outputs=gr.Label(num_top_classes=3, label=\"🤖 AI 預測結果 (Top 3)\"),\n", " title=\"🎨 智光商工電機電子群 數字辨識 AI 網頁版\",\n", " description=\"請點擊下方虛線框上傳圖片，或是直接將圖片拖曳進來。AI 將立刻分析數字並給出信心度評估！\",\n", " theme=gr.themes.Soft() # 使用柔和的現代化主題\n", ")\n", "\n", "# 啟動介面\n", "# share=True 會自動生成一個 public URL (例如 https://xxxx.gradio.live)，72小時內有效\n", "ui.launch(share=True, debug=False)" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 628 }, "id": "A-2eGLRxe9L3", "outputId": "8502b2ff-aaab-4c6d-e854-396da73f5ac9" }, "execution_count": 5, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "⏳ 步驟 4: 啟動手寫圖片上傳辨識介面 (Gradio)...\\n\n", "Colab notebook detected. To show errors in colab notebook, set debug=True in launch()\n", "* Running on public URL: https://1bfa487bb9dee4d6bd.gradio.live\n", "\n", "This share link expires in 1 week. For free permanent hosting and GPU upgrades, run `gradio deploy` from the terminal in the working directory to deploy to Hugging Face Spaces (https://huggingface.co/spaces)\n" ] }, { "output_type": "display_data", "data": { "text/plain": [ "" ], "text/html": [ "
" ] }, "metadata": {} }, { "output_type": "execute_result", "data": { "text/plain": [] }, "metadata": {}, "execution_count": 5 } ] } ] }