{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.11.11","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"gpu","dataSources":[{"sourceId":21154,"databundleVersionId":1243559,"isSourceIdPinned":false,"sourceType":"competition"}],"dockerImageVersionId":31041,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"!pip install --upgrade pip","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import tensorflow as tf\nfrom tensorflow.keras.applications.densenet import DenseNet169, preprocess_input\nfrom tensorflow.keras.layers import Dense, GlobalAveragePooling2D\nfrom tensorflow.keras.models import Model\nfrom tensorflow.keras.optimizers import Adam\nfrom tensorflow.keras.callbacks import EarlyStopping, Callback\nimport numpy as np\nimport pandas as pd\nimport os\n\n# ========== TPU 設定 ==========\ntry:\n    tpu = tf.distribute.cluster_resolver.TPUClusterResolver()\n    tf.config.experimental_connect_to_cluster(tpu)\n    tf.tpu.experimental.initialize_tpu_system(tpu)\n    strategy = tf.distribute.TPUStrategy(tpu)\n    print(\"✅ TPU initialized\")\nexcept ValueError:\n    strategy = tf.distribute.get_strategy()\n    print(\"✅ Using CPU/GPU\")\n\nIMAGE_SIZE = [384, 384]\nBATCH_SIZE = 16\nEPOCHS = 12\nNUM_CLASSES = 104\n\n# ========== 資料路徑 ==========\nGCS_PATH = '/kaggle/input/tpu-getting-started'\nTRAINING_FILENAMES = tf.io.gfile.glob(GCS_PATH + '/tfrecords-jpeg-512x512/train/*.tfrec')\nVALIDATION_FILENAMES = tf.io.gfile.glob(GCS_PATH + '/tfrecords-jpeg-512x512/val/*.tfrec')\nTEST_FILENAMES = tf.io.gfile.glob(GCS_PATH + '/tfrecords-jpeg-512x512/test/*.tfrec')\n\n# ========== 數值檢查 Callback ==========\nclass NanCheckCallback(Callback):\n    def on_train_batch_end(self, batch, logs=None):\n        for var in self.model.trainable_variables:\n            tf.debugging.check_numerics(var, message=\"❌ 權重出現 NaN/Inf\")\n\n# ========== 資料解析 ==========\ndef decode_image(image_data):\n    image = tf.image.decode_jpeg(image_data, channels=3)\n    image = tf.image.resize(image, IMAGE_SIZE)\n    image = tf.cast(image, tf.float32)\n    image = preprocess_input(image)\n    return image\n\ndef read_labeled_tfrecord(example):\n    LABELED_TFREC_FORMAT = {\n        \"image\": tf.io.FixedLenFeature([], tf.string),\n        \"class\": tf.io.FixedLenFeature([], tf.int64),\n    }\n    example = tf.io.parse_single_example(example, LABELED_TFREC_FORMAT)\n    return decode_image(example['image']), example['class']\n\ndef read_unlabeled_tfrecord(example):\n    UNLABELED_TFREC_FORMAT = {\n        \"image\": tf.io.FixedLenFeature([], tf.string),\n        \"id\": tf.io.FixedLenFeature([], tf.string),\n    }\n    example = tf.io.parse_single_example(example, UNLABELED_TFREC_FORMAT)\n    image = decode_image(example['image'])\n    return image, example['id']\n\ndef load_dataset(filenames, labeled=True, ordered=False):\n    ignore_order = tf.data.Options()\n    if not ordered:\n        ignore_order.experimental_deterministic = False\n    dataset = tf.data.TFRecordDataset(filenames, num_parallel_reads=tf.data.AUTOTUNE)\n    dataset = dataset.with_options(ignore_order)\n    dataset = dataset.map(read_labeled_tfrecord if labeled else read_unlabeled_tfrecord, num_parallel_calls=4)\n    return dataset\n\ndef get_dataset(filenames, labeled=True, batch_size=BATCH_SIZE):\n    dataset = load_dataset(filenames, labeled)\n    if labeled:\n        dataset = dataset.shuffle(2048)\n    dataset = dataset.batch(batch_size)\n    dataset = dataset.prefetch(tf.data.AUTOTUNE)\n    return dataset\n\n# ========== 建立資料集 ==========\ntrain_dataset = get_dataset(TRAINING_FILENAMES, labeled=True)\nval_dataset = get_dataset(VALIDATION_FILENAMES, labeled=True)\ntest_dataset = get_dataset(TEST_FILENAMES, labeled=False)\n\n# ========== 建立 DenseNet169 模型 ==========\nwith strategy.scope():\n    base_model = DenseNet169(\n        input_shape=(*IMAGE_SIZE, 3),\n        include_top=False,\n        weights='imagenet',\n        pooling='avg'\n    )\n\n    output = Dense(NUM_CLASSES, activation='softmax')(base_model.output)\n    model = Model(inputs=base_model.input, outputs=output)\n\n    model.compile(\n        optimizer=Adam(learning_rate=1e-4),\n        loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),\n        metrics=['sparse_categorical_accuracy']\n    )\n\nmodel.summary()\n\n# ========== 模型訓練 ==========\ntry:\n    history = model.fit(\n        train_dataset,\n        validation_data=val_dataset,\n        epochs=EPOCHS,\n        callbacks=[EarlyStopping(patience=3, restore_best_weights=True), NanCheckCallback()]\n    )\nexcept Exception as e:\n    print(\"🚨 模型訓練錯誤：\", e)\n\n# ========== 預測與提交 ==========\nprint(\"Generating predictions...\")\n\nraw_test_dataset = load_dataset(TEST_FILENAMES, labeled=False, ordered=True)\nbatched_test_dataset = raw_test_dataset.batch(BATCH_SIZE, drop_remainder=False)\n\n# 印出測試批次確認\nfor img, idnum in batched_test_dataset.take(1):\n    print(\"✅ 測試圖片形狀:\", img.shape, \"ID 數量:\", idnum.shape)\n\n# 逐批預測\nall_predictions = []\nall_ids = []\n\nfor batch_images, batch_ids in batched_test_dataset:\n    preds = model(batch_images, training=False)\n    all_predictions.append(preds.numpy())\n    all_ids.append(batch_ids.numpy())\n\n# 合併預測與 ID\npredictions = np.concatenate(all_predictions, axis=0)\nimage_ids = np.concatenate(all_ids, axis=0)\npredicted_labels = np.argmax(predictions, axis=-1)\n\nsubmission = pd.DataFrame({\n    \"id\": image_ids.astype(\"U\"),\n    \"label\": predicted_labels\n})\nsubmission.to_csv(\"submission.csv\", index=False)\nprint(\"✅ submission.csv created\")\n\n# ========== 印出驗證損失與準確率 ==========\nif history:\n    val_loss = history.history['val_loss'][-1]\n    val_acc = history.history['val_sparse_categorical_accuracy'][-1]\n    print(f\"✅ Validation Loss: {val_loss:.4f}\")\n    print(f\"✅ Validation Accuracy: {val_acc:.4f}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-06-11T11:12:22.573485Z","iopub.execute_input":"2025-06-11T11:12:22.573996Z","iopub.status.idle":"2025-06-11T12:11:01.878281Z","shell.execute_reply.started":"2025-06-11T11:12:22.573973Z","shell.execute_reply":"2025-06-11T12:11:01.877476Z"},"_kg_hide-input":false},"outputs":[],"execution_count":null}]}