project3_2.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ikn2MLxkgVhs",
        "outputId": "6e4d18c7-9305-43fa-b0ea-43d74e2eb269"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Collecting jamo\n",
            "  Downloading jamo-0.4.1-py3-none-any.whl (9.5 kB)\n",
            "Installing collected packages: jamo\n",
            "Successfully installed jamo-0.4.1\n"
          ]
        }
      ],
      "source": [
        "!pip install jamo"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.layers import Input, Conv1D, LSTM, Dense, Embedding, RepeatVector, Concatenate, Reshape, TimeDistributed\n",
        "from tensorflow.keras.models import Model\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "from tensorflow.keras.losses import MeanSquaredError\n",
        "from tensorflow.keras.callbacks import ModelCheckpoint\n",
        "from scipy.io.wavfile import read, write\n",
        "import librosa\n",
        "from tqdm import tqdm"
      ],
      "metadata": {
        "id": "NxL8PuDqgaDI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SE2XdyY4gafx",
        "outputId": "946e04fc-31cb-4f42-d856-f4f7ddbdc2dd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_data_path = '/content/drive/MyDrive/dataset/voice/ljs_audio_text_train_filelist.txt'\n",
        "val_data_path =  '/content/drive/MyDrive/dataset/voice/ljs_audio_text_val_filelist.txt'\n",
        "wav_path = '/content/drive/MyDrive/dataset/voice/'"
      ],
      "metadata": {
        "id": "AonsUEFCgcVN"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import random\n",
        "\n",
        "def load_data(file_path, wav_path, subset_ratio=None):\n",
        "    data = []\n",
        "    with open(file_path, 'r') as f:\n",
        "        lines = f.readlines()\n",
        "        if subset_ratio:\n",
        "            random.shuffle(lines)\n",
        "            num_samples = int(len(lines) * subset_ratio)\n",
        "            lines = lines[:num_samples]\n",
        "        for line in lines:\n",
        "            wav_file, text = line.strip().split(\"|\")\n",
        "            wav_file_path = os.path.join(wav_path, wav_file[10:])\n",
        "            audio, _ = librosa.load(wav_file_path, sr=None)\n",
        "            data.append((audio, text))\n",
        "    return data"
      ],
      "metadata": {
        "id": "umkTb6UNgeRa"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 데이터셋이 많아 학습 및 로딩시간이 매우 기므로 20%만 랜덤 추출 (시험용)\n",
        "\n",
        "train_data = load_data(train_data_path, wav_path, subset_ratio=0.2)\n",
        "val_data = load_data(val_data_path, wav_path, subset_ratio=0.2)"
      ],
      "metadata": {
        "id": "hSpRs8F4ggna"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 한글을 초성 중성 종성으로 분리한 후 각각 숫자로 임베딩\n",
        "\n",
        "import re\n",
        "import unicodedata\n",
        "\n",
        "CHOSUNGS = \"\".join([chr(_) for _ in range(0x1100, 0x1113)])\n",
        "JOONGSUNGS = \"\".join([chr(_) for _ in range(0x1161, 0x1176)])\n",
        "JONGSUNGS = \"\".join([chr(_) for _ in range(0x11A8, 0x11C3)])\n",
        "ALPHABETS = \"ABCDEFGHIJKLMNOPQRSTUVWXYZ\"\n",
        "NUMBERS = \"0123456789\"\n",
        "SPECIALS = \" ?!\"\n",
        "\n",
        "ALL_VOCABS = \"\".join([\n",
        "    CHOSUNGS,\n",
        "    JOONGSUNGS,\n",
        "    JONGSUNGS,\n",
        "    ALPHABETS,\n",
        "    NUMBERS,\n",
        "    SPECIALS\n",
        "])\n",
        "VOCAB_DICT = {\n",
        "    \"_\": 0,\n",
        "    \"~\": 1,\n",
        "}\n",
        "\n",
        "for idx, v in enumerate(ALL_VOCABS):\n",
        "    VOCAB_DICT[v] = idx + 2\n",
        "\n",
        "\n",
        "def normalize(text):\n",
        "    text = unicodedata.normalize('NFKD', text)\n",
        "    text = text.upper()\n",
        "    text = text.replace('%', unicodedata.normalize('NFKD', '퍼센트'))\n",
        "    regex = unicodedata.normalize('NFKD', r\"[^ \\u11A8-\\u11FF\\u1100-\\u115E\\u1161-\\u11A70-9A-Z?!]\")\n",
        "    text = re.sub(regex, '', text)\n",
        "    text = re.sub(' +', ' ', text)\n",
        "    text = text.strip()\n",
        "    return text\n",
        "\n",
        "\n",
        "def tokenize(text, encoding: bool = True):\n",
        "    tokens = list()\n",
        "\n",
        "    for t in text:\n",
        "        if encoding:\n",
        "            tokens.append(VOCAB_DICT[t])\n",
        "        else:\n",
        "            tokens.append(t)\n",
        "\n",
        "    if encoding:\n",
        "        tokens.append(VOCAB_DICT['~'])\n",
        "    else:\n",
        "        tokens.append('~')\n",
        "\n",
        "    return tokens\n",
        "\n",
        "\n",
        "def text_to_sequence(text):\n",
        "    text = normalize(text)\n",
        "    tokens = tokenize(text, encoding=True)\n",
        "    return tokens\n",
        "\n",
        "# Example usage\n",
        "text = \"안녕하세요\"\n",
        "embedding = text_to_sequence(text)\n",
        "print(embedding)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UQYUNcjAgwxk",
        "outputId": "ec868771-a3c2-4b7f-d2ea-6514309c6279"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[13, 21, 45, 4, 27, 62, 20, 21, 11, 26, 13, 33, 1]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.utils import pad_sequences\n",
        "\n",
        "# Preprocess data(text)\n",
        "train_wavs, train_texts = zip(*train_data)\n",
        "val_wavs, val_texts = zip(*val_data)\n",
        "\n",
        "max_audio_length = 200\n",
        "output_dim = 400\n",
        "\n",
        "# Convert text to sequence\n",
        "train_texts = [text_to_sequence(text) for text in train_texts]\n",
        "val_texts = [text_to_sequence(text) for text in val_texts]\n",
        "\n",
        "# Padding sequences for train and val texts\n",
        "train_texts = pad_sequences(train_texts, dtype='int32', padding='post', maxlen=140)\n",
        "val_texts = pad_sequences(val_texts, dtype='int32', padding='post', maxlen=140)\n",
        "\n",
        "# Expanding the dimension of the text data\n",
        "train_texts = np.expand_dims(train_texts, axis=1)\n",
        "val_texts = np.expand_dims(val_texts, axis=1)"
      ],
      "metadata": {
        "id": "CRHPopD1g1l3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Padding and reshaping train and val wavs\n",
        "train_wavs_mel = []\n",
        "for audio in train_wavs:\n",
        "    if len(audio.shape) == 1:\n",
        "        audio = audio[:80000]  # Cut if size exceeds 80000\n",
        "        audio = np.pad(audio, (0, max(0, 80000 - len(audio))))  # Pad if size is less than 80000\n",
        "        audio = audio.reshape(-1, 400)[:200, :]  # Reshape to 2D and cut to fit into (200, 400)\n",
        "    train_wavs_mel.append(audio)\n",
        "\n",
        "val_wavs_mel = []\n",
        "for audio in val_wavs:\n",
        "    if len(audio.shape) == 1:\n",
        "        audio = audio[:80000]  # Cut if size exceeds 80000\n",
        "        audio = np.pad(audio, (0, max(0, 80000 - len(audio))))  # Pad if size is less than 80000\n",
        "        audio = audio.reshape(-1, 400)[:200, :]  # Reshape to 2D and cut to fit into (200, 400)\n",
        "    val_wavs_mel.append(audio)\n",
        "\n",
        "train_wavs_mel = np.array(train_wavs_mel)\n",
        "val_wavs_mel = np.array(val_wavs_mel)"
      ],
      "metadata": {
        "id": "tz1Gn4eVg4Fg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "max_text_length = 140 # 최대 문장 길이\n",
        "vocab_size = 10000 # 단어 사전 크기\n",
        "embedding_dim = 256 # 임베딩 차원\n",
        "latent_dim = 64 # 잠재 공간 차원\n",
        "output_length = 100 # 출력 길이\n",
        "output_dim = 400  # 출력 차원\n",
        "max_audio_length = 200   # 음성 길이"
      ],
      "metadata": {
        "id": "rdCV0GaVgl_N"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.models import Model\n",
        "from tensorflow.keras.layers import Input, LSTM, Dense, LayerNormalization\n",
        "from tensorflow.keras.layers import MultiHeadAttention\n",
        "\n",
        "def create_tts_model(input_dim_text, input_dim_audio, hidden_dim, num_heads):\n",
        "    # Define inputs\n",
        "    text_input = Input(shape=(1, input_dim_text), name=\"text_input\")\n",
        "    audio_input = Input(shape=(200, input_dim_audio), name=\"audio_input\")\n",
        "\n",
        "    # Use LSTM to map text_input to the same dimension as audio_input\n",
        "    text_mapped = LSTM(input_dim_audio, return_sequences=True)(text_input)\n",
        "\n",
        "    # Transformer encoder\n",
        "    attention = MultiHeadAttention(num_heads=num_heads, key_dim=hidden_dim)(text_mapped, audio_input)\n",
        "    attention = LayerNormalization(epsilon=1e-6)(attention + audio_input)\n",
        "    output = Dense(input_dim_audio)(attention)\n",
        "\n",
        "    model = Model(inputs=[text_input, audio_input], outputs=output)\n",
        "    \n",
        "    return model\n",
        "\n",
        "# Create the TTS model\n",
        "tts_model = create_tts_model(input_dim_text=140, input_dim_audio=400, hidden_dim=64, num_heads=2)\n",
        "tts_model.compile(optimizer='adam', loss='mse')"
      ],
      "metadata": {
        "id": "NSAoNOA7gnhH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 모델 생성\n",
        "model = tts_model\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4I81niA6gsJm",
        "outputId": "b74f91ea-dd41-437d-9524-ff66e47e10e2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_29\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " text_input (InputLayer)        [(None, 1, 140)]     0           []                               \n",
            "                                                                                                  \n",
            " lstm_26 (LSTM)                 (None, 1, 400)       865600      ['text_input[0][0]']             \n",
            "                                                                                                  \n",
            " audio_input (InputLayer)       [(None, 200, 400)]   0           []                               \n",
            "                                                                                                  \n",
            " multi_head_attention_10 (Multi  (None, 1, 400)      205584      ['lstm_26[0][0]',                \n",
            " HeadAttention)                                                   'audio_input[0][0]']            \n",
            "                                                                                                  \n",
            " tf.__operators__.add_14 (TFOpL  (None, 200, 400)    0           ['multi_head_attention_10[0][0]',\n",
            " ambda)                                                           'audio_input[0][0]']            \n",
            "                                                                                                  \n",
            " layer_normalization_10 (LayerN  (None, 200, 400)    800         ['tf.__operators__.add_14[0][0]']\n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " dense_21 (Dense)               (None, 200, 400)     160400      ['layer_normalization_10[0][0]'] \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 1,232,384\n",
            "Trainable params: 1,232,384\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow import keras\n",
        "keras.utils.plot_model(model, show_shapes=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 644
        },
        "id": "uSfV-kcpgtco",
        "outputId": "d92fb52b-790c-4165-a594-9e848d88fa83"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 167
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "text_data = np.array(train_texts, dtype=np.float32).reshape((-1, 1, 140))\n",
        "val_text_data = np.array(val_texts, dtype=np.float32).reshape((-1, 1, 140))\n",
        "audio_data = train_wavs_mel\n",
        "val_audio_data = val_wavs_mel\n",
        "\n",
        "# Training the model\n",
        "tts_model.fit([text_data, audio_data], audio_data, epochs=100, batch_size=32, validation_data=([val_text_data, val_audio_data], val_audio_data))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Jyqb1RMkhRsa",
        "outputId": "edaa8c4e-c5de-4b50-899b-11672d283011"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "19/19 [==============================] - 12s 425ms/step - loss: 0.4280 - val_loss: 0.1457\n",
            "Epoch 2/100\n",
            "19/19 [==============================] - 6s 310ms/step - loss: 0.0796 - val_loss: 0.0132\n",
            "Epoch 3/100\n",
            "19/19 [==============================] - 7s 397ms/step - loss: 0.0095 - val_loss: 0.0060\n",
            "Epoch 4/100\n",
            "19/19 [==============================] - 5s 275ms/step - loss: 0.0059 - val_loss: 0.0050\n",
            "Epoch 5/100\n",
            "19/19 [==============================] - 7s 382ms/step - loss: 0.0051 - val_loss: 0.0045\n",
            "Epoch 6/100\n",
            "19/19 [==============================] - 5s 276ms/step - loss: 0.0046 - val_loss: 0.0041\n",
            "Epoch 7/100\n",
            "19/19 [==============================] - 5s 264ms/step - loss: 0.0042 - val_loss: 0.0038\n",
            "Epoch 8/100\n",
            "19/19 [==============================] - 7s 378ms/step - loss: 0.0039 - val_loss: 0.0035\n",
            "Epoch 9/100\n",
            "19/19 [==============================] - 6s 301ms/step - loss: 0.0036 - val_loss: 0.0032\n",
            "Epoch 10/100\n",
            "19/19 [==============================] - 7s 397ms/step - loss: 0.0034 - val_loss: 0.0030\n",
            "Epoch 11/100\n",
            "19/19 [==============================] - 5s 287ms/step - loss: 0.0032 - val_loss: 0.0028\n",
            "Epoch 12/100\n",
            "19/19 [==============================] - 6s 292ms/step - loss: 0.0030 - val_loss: 0.0027\n",
            "Epoch 13/100\n",
            "19/19 [==============================] - 7s 360ms/step - loss: 0.0029 - val_loss: 0.0025\n",
            "Epoch 14/100\n",
            "19/19 [==============================] - 5s 267ms/step - loss: 0.0028 - val_loss: 0.0024\n",
            "Epoch 15/100\n",
            "19/19 [==============================] - 8s 408ms/step - loss: 0.0026 - val_loss: 0.0023\n",
            "Epoch 16/100\n",
            "19/19 [==============================] - 5s 267ms/step - loss: 0.0025 - val_loss: 0.0022\n",
            "Epoch 17/100\n",
            "19/19 [==============================] - 6s 310ms/step - loss: 0.0024 - val_loss: 0.0021\n",
            "Epoch 18/100\n",
            "19/19 [==============================] - 6s 325ms/step - loss: 0.0024 - val_loss: 0.0020\n",
            "Epoch 19/100\n",
            "19/19 [==============================] - 8s 403ms/step - loss: 0.0023 - val_loss: 0.0020\n",
            "Epoch 20/100\n",
            "19/19 [==============================] - 12s 614ms/step - loss: 0.0022 - val_loss: 0.0019\n",
            "Epoch 21/100\n",
            "19/19 [==============================] - 7s 367ms/step - loss: 0.0021 - val_loss: 0.0018\n",
            "Epoch 22/100\n",
            "19/19 [==============================] - 8s 444ms/step - loss: 0.0021 - val_loss: 0.0018\n",
            "Epoch 23/100\n",
            "19/19 [==============================] - 7s 388ms/step - loss: 0.0020 - val_loss: 0.0017\n",
            "Epoch 24/100\n",
            "19/19 [==============================] - 5s 273ms/step - loss: 0.0020 - val_loss: 0.0017\n",
            "Epoch 25/100\n",
            "19/19 [==============================] - 6s 344ms/step - loss: 0.0019 - val_loss: 0.0016\n",
            "Epoch 26/100\n",
            "19/19 [==============================] - 6s 310ms/step - loss: 0.0019 - val_loss: 0.0016\n",
            "Epoch 27/100\n",
            "19/19 [==============================] - 6s 293ms/step - loss: 0.0018 - val_loss: 0.0016\n",
            "Epoch 28/100\n",
            "19/19 [==============================] - 7s 367ms/step - loss: 0.0018 - val_loss: 0.0015\n",
            "Epoch 29/100\n",
            "19/19 [==============================] - 5s 273ms/step - loss: 0.0017 - val_loss: 0.0015\n",
            "Epoch 30/100\n",
            "19/19 [==============================] - 7s 384ms/step - loss: 0.0017 - val_loss: 0.0015\n",
            "Epoch 31/100\n",
            "19/19 [==============================] - 5s 285ms/step - loss: 0.0017 - val_loss: 0.0014\n",
            "Epoch 32/100\n",
            "19/19 [==============================] - 6s 307ms/step - loss: 0.0016 - val_loss: 0.0014\n",
            "Epoch 33/100\n",
            "19/19 [==============================] - 7s 340ms/step - loss: 0.0016 - val_loss: 0.0014\n",
            "Epoch 34/100\n",
            "19/19 [==============================] - 5s 269ms/step - loss: 0.0016 - val_loss: 0.0013\n",
            "Epoch 35/100\n",
            "19/19 [==============================] - 7s 368ms/step - loss: 0.0015 - val_loss: 0.0013\n",
            "Epoch 36/100\n",
            "19/19 [==============================] - 6s 297ms/step - loss: 0.0015 - val_loss: 0.0013\n",
            "Epoch 37/100\n",
            "19/19 [==============================] - 7s 356ms/step - loss: 0.0015 - val_loss: 0.0013\n",
            "Epoch 38/100\n",
            "19/19 [==============================] - 6s 294ms/step - loss: 0.0015 - val_loss: 0.0012\n",
            "Epoch 39/100\n",
            "19/19 [==============================] - 5s 287ms/step - loss: 0.0014 - val_loss: 0.0012\n",
            "Epoch 40/100\n",
            "19/19 [==============================] - 7s 392ms/step - loss: 0.0014 - val_loss: 0.0012\n",
            "Epoch 41/100\n",
            "19/19 [==============================] - 5s 278ms/step - loss: 0.0014 - val_loss: 0.0012\n",
            "Epoch 42/100\n",
            "19/19 [==============================] - 7s 373ms/step - loss: 0.0014 - val_loss: 0.0012\n",
            "Epoch 43/100\n",
            "19/19 [==============================] - 5s 271ms/step - loss: 0.0013 - val_loss: 0.0011\n",
            "Epoch 44/100\n",
            "19/19 [==============================] - 5s 280ms/step - loss: 0.0013 - val_loss: 0.0011\n",
            "Epoch 45/100\n",
            "19/19 [==============================] - 7s 381ms/step - loss: 0.0013 - val_loss: 0.0011\n",
            "Epoch 46/100\n",
            "19/19 [==============================] - 5s 277ms/step - loss: 0.0013 - val_loss: 0.0011\n",
            "Epoch 47/100\n",
            "19/19 [==============================] - 7s 392ms/step - loss: 0.0013 - val_loss: 0.0011\n",
            "Epoch 48/100\n",
            "19/19 [==============================] - 5s 271ms/step - loss: 0.0013 - val_loss: 0.0011\n",
            "Epoch 49/100\n",
            "19/19 [==============================] - 6s 330ms/step - loss: 0.0012 - val_loss: 0.0010\n",
            "Epoch 50/100\n",
            "19/19 [==============================] - 7s 343ms/step - loss: 0.0012 - val_loss: 0.0010\n",
            "Epoch 51/100\n",
            "19/19 [==============================] - 5s 282ms/step - loss: 0.0012 - val_loss: 0.0010\n",
            "Epoch 52/100\n",
            "19/19 [==============================] - 7s 400ms/step - loss: 0.0012 - val_loss: 0.0010\n",
            "Epoch 53/100\n",
            "19/19 [==============================] - 5s 280ms/step - loss: 0.0012 - val_loss: 9.9170e-04\n",
            "Epoch 54/100\n",
            "19/19 [==============================] - 8s 410ms/step - loss: 0.0012 - val_loss: 9.7571e-04\n",
            "Epoch 55/100\n",
            "19/19 [==============================] - 5s 282ms/step - loss: 0.0012 - val_loss: 9.6795e-04\n",
            "Epoch 56/100\n",
            "19/19 [==============================] - 6s 321ms/step - loss: 0.0011 - val_loss: 9.5417e-04\n",
            "Epoch 57/100\n",
            "19/19 [==============================] - 7s 353ms/step - loss: 0.0011 - val_loss: 9.5314e-04\n",
            "Epoch 58/100\n",
            "19/19 [==============================] - 6s 321ms/step - loss: 0.0011 - val_loss: 9.4275e-04\n",
            "Epoch 59/100\n",
            "19/19 [==============================] - 8s 401ms/step - loss: 0.0011 - val_loss: 9.3005e-04\n",
            "Epoch 60/100\n",
            "19/19 [==============================] - 5s 288ms/step - loss: 0.0011 - val_loss: 9.5843e-04\n",
            "Epoch 61/100\n",
            "19/19 [==============================] - 7s 389ms/step - loss: 0.0011 - val_loss: 9.3716e-04\n",
            "Epoch 62/100\n",
            "19/19 [==============================] - 5s 283ms/step - loss: 0.0011 - val_loss: 8.9959e-04\n",
            "Epoch 63/100\n",
            "19/19 [==============================] - 5s 286ms/step - loss: 0.0011 - val_loss: 8.8687e-04\n",
            "Epoch 64/100\n",
            "19/19 [==============================] - 7s 357ms/step - loss: 0.0010 - val_loss: 8.8514e-04\n",
            "Epoch 65/100\n",
            "19/19 [==============================] - 5s 269ms/step - loss: 0.0010 - val_loss: 8.5015e-04\n",
            "Epoch 66/100\n",
            "19/19 [==============================] - 7s 397ms/step - loss: 0.0010 - val_loss: 8.5973e-04\n",
            "Epoch 67/100\n",
            "19/19 [==============================] - 5s 274ms/step - loss: 0.0010 - val_loss: 8.6785e-04\n",
            "Epoch 68/100\n",
            "19/19 [==============================] - 6s 341ms/step - loss: 0.0010 - val_loss: 8.5376e-04\n",
            "Epoch 69/100\n",
            "19/19 [==============================] - 6s 308ms/step - loss: 9.9458e-04 - val_loss: 8.5037e-04\n",
            "Epoch 70/100\n",
            "19/19 [==============================] - 5s 287ms/step - loss: 9.9091e-04 - val_loss: 8.4568e-04\n",
            "Epoch 71/100\n",
            "19/19 [==============================] - 7s 381ms/step - loss: 9.8165e-04 - val_loss: 8.4441e-04\n",
            "Epoch 72/100\n",
            "19/19 [==============================] - 5s 276ms/step - loss: 9.8958e-04 - val_loss: 8.1849e-04\n",
            "Epoch 73/100\n",
            "19/19 [==============================] - 7s 383ms/step - loss: 9.5705e-04 - val_loss: 8.1268e-04\n",
            "Epoch 74/100\n",
            "19/19 [==============================] - 5s 252ms/step - loss: 9.5260e-04 - val_loss: 8.2156e-04\n",
            "Epoch 75/100\n",
            "19/19 [==============================] - 5s 267ms/step - loss: 9.4032e-04 - val_loss: 7.8233e-04\n",
            "Epoch 76/100\n",
            "19/19 [==============================] - 7s 391ms/step - loss: 9.4364e-04 - val_loss: 7.8626e-04\n",
            "Epoch 77/100\n",
            "19/19 [==============================] - 5s 272ms/step - loss: 9.3300e-04 - val_loss: 7.9545e-04\n",
            "Epoch 78/100\n",
            "19/19 [==============================] - 7s 363ms/step - loss: 9.2931e-04 - val_loss: 7.7171e-04\n",
            "Epoch 79/100\n",
            "19/19 [==============================] - 5s 260ms/step - loss: 9.0094e-04 - val_loss: 7.7616e-04\n",
            "Epoch 80/100\n",
            "19/19 [==============================] - 5s 279ms/step - loss: 9.2317e-04 - val_loss: 7.8265e-04\n",
            "Epoch 81/100\n",
            "19/19 [==============================] - 7s 367ms/step - loss: 9.0056e-04 - val_loss: 7.5875e-04\n",
            "Epoch 82/100\n",
            "19/19 [==============================] - 5s 266ms/step - loss: 8.7908e-04 - val_loss: 7.2325e-04\n",
            "Epoch 83/100\n",
            "19/19 [==============================] - 7s 365ms/step - loss: 8.7592e-04 - val_loss: 7.2820e-04\n",
            "Epoch 84/100\n",
            "19/19 [==============================] - 5s 268ms/step - loss: 8.6631e-04 - val_loss: 7.5699e-04\n",
            "Epoch 85/100\n",
            "19/19 [==============================] - 5s 276ms/step - loss: 8.5724e-04 - val_loss: 7.1741e-04\n",
            "Epoch 86/100\n",
            "19/19 [==============================] - 7s 391ms/step - loss: 8.3983e-04 - val_loss: 7.2186e-04\n",
            "Epoch 87/100\n",
            "19/19 [==============================] - 5s 265ms/step - loss: 8.3676e-04 - val_loss: 7.1778e-04\n",
            "Epoch 88/100\n",
            "19/19 [==============================] - 6s 341ms/step - loss: 8.4505e-04 - val_loss: 7.1051e-04\n",
            "Epoch 89/100\n",
            "19/19 [==============================] - 5s 261ms/step - loss: 8.5183e-04 - val_loss: 7.0100e-04\n",
            "Epoch 90/100\n",
            "19/19 [==============================] - 5s 272ms/step - loss: 8.1199e-04 - val_loss: 6.8858e-04\n",
            "Epoch 91/100\n",
            "19/19 [==============================] - 7s 386ms/step - loss: 7.9528e-04 - val_loss: 6.6210e-04\n",
            "Epoch 92/100\n",
            "19/19 [==============================] - 5s 270ms/step - loss: 7.9401e-04 - val_loss: 6.9941e-04\n",
            "Epoch 93/100\n",
            "19/19 [==============================] - 7s 385ms/step - loss: 8.0289e-04 - val_loss: 6.6954e-04\n",
            "Epoch 94/100\n",
            "19/19 [==============================] - 6s 285ms/step - loss: 7.9016e-04 - val_loss: 6.6941e-04\n",
            "Epoch 95/100\n",
            "19/19 [==============================] - 6s 299ms/step - loss: 7.9176e-04 - val_loss: 6.8145e-04\n",
            "Epoch 96/100\n",
            "19/19 [==============================] - 7s 359ms/step - loss: 7.6194e-04 - val_loss: 6.5211e-04\n",
            "Epoch 97/100\n",
            "19/19 [==============================] - 5s 274ms/step - loss: 7.4466e-04 - val_loss: 6.3916e-04\n",
            "Epoch 98/100\n",
            "19/19 [==============================] - 7s 381ms/step - loss: 7.4703e-04 - val_loss: 6.1912e-04\n",
            "Epoch 99/100\n",
            "19/19 [==============================] - 5s 263ms/step - loss: 7.3794e-04 - val_loss: 6.7602e-04\n",
            "Epoch 100/100\n",
            "19/19 [==============================] - 6s 340ms/step - loss: 7.9396e-04 - val_loss: 7.1785e-04\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f21f1ed25f0>"
            ]
          },
          "metadata": {},
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "tts_model.save('my_model.h5')  # creates a HDF5 file 'my_model.h5'\n",
        "# import matplotlib.pyplot as plt\n",
        "\n",
        "# # Save loss and validation loss data\n",
        "# np.savez(\"loss_history.npz\", loss=history.history['loss'], val_loss=history.history['val_loss'])\n",
        "\n",
        "# # Plot the loss and validation loss curves\n",
        "# epochs = range(1, len(history.history['loss']) + 1)\n",
        "\n",
        "# plt.plot(epochs, history.history['loss'], 'b', label='Training Loss')\n",
        "# plt.plot(epochs, history.history['val_loss'], 'r', label='Validation Loss')\n",
        "# plt.title('Training and Validation Loss')\n",
        "# plt.xlabel('Epochs')\n",
        "# plt.ylabel('Loss')\n",
        "# plt.legend()\n",
        "# plt.savefig(\"loss_curve.png\")\n",
        "# plt.show();"
      ],
      "metadata": {
        "id": "SSjry3mnhTuB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.layers import Input, Lambda, Flatten\n",
        "from tensorflow.keras import backend as K\n",
        "\n",
        "# def create_text_to_voice_model(input_dim_text, hidden_dim, num_heads):\n",
        "#     # Define the input for text\n",
        "#     text_input = Input(shape=(1, input_dim_text), name=\"text_input\")\n",
        "\n",
        "#     # Use the \"text-to-voice\" model layers up to the attention layer\n",
        "#     text_mapped = tts_model.get_layer(\"lstm_8\")(text_input)\n",
        "#     attention = tts_model.get_layer(\"multi_head_attention_9\")(text_mapped, tts_model.get_layer(\"audio_input\")(text_mapped))\n",
        "#     attention = tts_model.get_layer(\"layer_normalization_9\")(attention)\n",
        "\n",
        "#     # Adjust the time dimension to match the desired shape\n",
        "#     output = Lambda(lambda x: K.repeat_elements(x, 200, axis=1))(attention)\n",
        "\n",
        "#     # Create the model\n",
        "#     model = Model(inputs=text_input, outputs=output)\n",
        "\n",
        "#     # Set the pre-trained model's layers as non-trainable and transfer the weights\n",
        "#     for layer in model.layers:\n",
        "#         if layer.name in tts_model.layers:\n",
        "#             layer.set_weights(tts_model.get_layer(layer.name).get_weights())\n",
        "#             layer.trainable = False\n",
        "\n",
        "#     return model\n",
        "\n",
        "def create_text_to_voice_model(input_dim_text, hidden_dim, output_sequence_length, output_feature_dim, pretrained_model_path):\n",
        "    # Load the pretrained model\n",
        "    pretrained_model = keras.models.load_model(pretrained_model_path)\n",
        "\n",
        "    # Create a new input for the text data\n",
        "    text_input = Input(shape=(1, input_dim_text), name=\"text_input\")\n",
        "\n",
        "    # Pass the text input through the pretrained model's LSTM and attention layers\n",
        "    lstm_output = pretrained_model.get_layer(\"lstm_26\")(text_input)\n",
        "    attention_output = pretrained_model.get_layer(\"multi_head_attention_10\")(lstm_output, lstm_output)\n",
        "    normalized_output = pretrained_model.get_layer(\"layer_normalization_10\")(attention_output)\n",
        "\n",
        "    # Add a new LSTM layer with the specified hidden dimension\n",
        "    lstm_layer = LSTM(hidden_dim, return_sequences=True)(normalized_output)\n",
        "\n",
        "    # Apply Dense layer to each time step\n",
        "    dense_output = TimeDistributed(Dense(output_feature_dim))(lstm_layer)\n",
        "\n",
        "    # Repeat the output sequence length using a Lambda layer\n",
        "    repeated_output = Lambda(lambda x: K.repeat_elements(x, output_sequence_length, axis=1))(dense_output)\n",
        "\n",
        "    # Create a model with the text input and the repeated output\n",
        "    model = Model(text_input, repeated_output)\n",
        "\n",
        "    return model"
      ],
      "metadata": {
        "id": "_M5sRj3DzxAK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "text_to_voice_model = create_text_to_voice_model(input_dim_text=140, hidden_dim=64, output_sequence_length=200, output_feature_dim=400, pretrained_model_path='my_model.h5')\n",
        "text_to_voice_model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_EziHDzv1-wX",
        "outputId": "feeec277-3ff1-42e6-f838-a54b95e575f2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_31\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " text_input (InputLayer)        [(None, 1, 140)]     0           []                               \n",
            "                                                                                                  \n",
            " lstm_26 (LSTM)                 (None, 1, 400)       865600      ['text_input[0][0]']             \n",
            "                                                                                                  \n",
            " multi_head_attention_10 (Multi  (None, 1, 400)      205584      ['lstm_26[1][0]',                \n",
            " HeadAttention)                                                   'lstm_26[1][0]']                \n",
            "                                                                                                  \n",
            " layer_normalization_10 (LayerN  multiple            800         ['multi_head_attention_10[1][0]']\n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " lstm_34 (LSTM)                 (None, 1, 64)        119040      ['layer_normalization_10[1][0]'] \n",
            "                                                                                                  \n",
            " time_distributed_4 (TimeDistri  (None, 1, 400)      26000       ['lstm_34[0][0]']                \n",
            " buted)                                                                                           \n",
            "                                                                                                  \n",
            " lambda_4 (Lambda)              (None, 200, 400)     0           ['time_distributed_4[0][0]']     \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 1,217,024\n",
            "Trainable params: 1,217,024\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "keras.utils.plot_model(text_to_voice_model, show_shapes=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 754
        },
        "id": "BmbwpaAh2aOl",
        "outputId": "58411cb6-ea92-473b-8f80-af8e2d7c3ba0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 188
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.callbacks import History\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# Create an instance of History callback\n",
        "history = History()\n",
        "\n",
        "text_to_voice_model.compile(optimizer='adam', loss='mse')\n",
        "text_to_voice_model.fit(text_data, audio_data, epochs=100, batch_size=32, validation_data=(val_text_data, val_audio_data), callbacks=[history])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ejv74n-16L8t",
        "outputId": "4ec1b7e0-7751-4a6f-fab0-8058cc358db7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "19/19 [==============================] - 10s 161ms/step - loss: 0.0071 - val_loss: 0.0064\n",
            "Epoch 2/100\n",
            "19/19 [==============================] - 1s 61ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 3/100\n",
            "19/19 [==============================] - 1s 61ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 4/100\n",
            "19/19 [==============================] - 1s 60ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 5/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 6/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 7/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 8/100\n",
            "19/19 [==============================] - 1s 52ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 9/100\n",
            "19/19 [==============================] - 2s 87ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 10/100\n",
            "19/19 [==============================] - 2s 90ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 11/100\n",
            "19/19 [==============================] - 1s 63ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 12/100\n",
            "19/19 [==============================] - 1s 48ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 13/100\n",
            "19/19 [==============================] - 1s 61ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 14/100\n",
            "19/19 [==============================] - 1s 47ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 15/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 16/100\n",
            "19/19 [==============================] - 1s 45ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 17/100\n",
            "19/19 [==============================] - 1s 45ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 18/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 19/100\n",
            "19/19 [==============================] - 1s 54ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 20/100\n",
            "19/19 [==============================] - 1s 52ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 21/100\n",
            "19/19 [==============================] - 1s 52ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 22/100\n",
            "19/19 [==============================] - 2s 83ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 23/100\n",
            "19/19 [==============================] - 2s 79ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 24/100\n",
            "19/19 [==============================] - 1s 68ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 25/100\n",
            "19/19 [==============================] - 1s 54ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 26/100\n",
            "19/19 [==============================] - 1s 45ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 27/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 28/100\n",
            "19/19 [==============================] - 1s 54ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 29/100\n",
            "19/19 [==============================] - 1s 54ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 30/100\n",
            "19/19 [==============================] - 1s 54ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 31/100\n",
            "19/19 [==============================] - 1s 61ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 32/100\n",
            "19/19 [==============================] - 1s 61ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 33/100\n",
            "19/19 [==============================] - 1s 66ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 34/100\n",
            "19/19 [==============================] - 1s 70ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 35/100\n",
            "19/19 [==============================] - 1s 78ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 36/100\n",
            "19/19 [==============================] - 1s 71ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 37/100\n",
            "19/19 [==============================] - 1s 71ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 38/100\n",
            "19/19 [==============================] - 1s 53ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 39/100\n",
            "19/19 [==============================] - 1s 57ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 40/100\n",
            "19/19 [==============================] - 1s 48ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 41/100\n",
            "19/19 [==============================] - 1s 47ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 42/100\n",
            "19/19 [==============================] - 1s 47ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 43/100\n",
            "19/19 [==============================] - 1s 47ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 44/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 45/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 46/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 47/100\n",
            "19/19 [==============================] - 1s 55ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 48/100\n",
            "19/19 [==============================] - 1s 58ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 49/100\n",
            "19/19 [==============================] - 2s 86ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 50/100\n",
            "19/19 [==============================] - 2s 88ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 51/100\n",
            "19/19 [==============================] - 1s 76ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 52/100\n",
            "19/19 [==============================] - 1s 62ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 53/100\n",
            "19/19 [==============================] - 1s 61ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 54/100\n",
            "19/19 [==============================] - 1s 62ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 55/100\n",
            "19/19 [==============================] - 1s 51ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 56/100\n",
            "19/19 [==============================] - 1s 45ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 57/100\n",
            "19/19 [==============================] - 1s 54ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 58/100\n",
            "19/19 [==============================] - 1s 62ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 59/100\n",
            "19/19 [==============================] - 1s 51ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 60/100\n",
            "19/19 [==============================] - 2s 85ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 61/100\n",
            "19/19 [==============================] - 1s 75ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 62/100\n",
            "19/19 [==============================] - 2s 87ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 63/100\n",
            "19/19 [==============================] - 1s 57ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 64/100\n",
            "19/19 [==============================] - 1s 48ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 65/100\n",
            "19/19 [==============================] - 1s 64ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 66/100\n",
            "19/19 [==============================] - 1s 60ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 67/100\n",
            "19/19 [==============================] - 1s 60ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 68/100\n",
            "19/19 [==============================] - 1s 60ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 69/100\n",
            "19/19 [==============================] - 1s 50ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 70/100\n",
            "19/19 [==============================] - 1s 45ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 71/100\n",
            "19/19 [==============================] - 1s 54ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 72/100\n",
            "19/19 [==============================] - 1s 68ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 73/100\n",
            "19/19 [==============================] - 1s 66ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 74/100\n",
            "19/19 [==============================] - 1s 65ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 75/100\n",
            "19/19 [==============================] - 1s 65ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 76/100\n",
            "19/19 [==============================] - 1s 54ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 77/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 78/100\n",
            "19/19 [==============================] - 1s 45ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 79/100\n",
            "19/19 [==============================] - 1s 45ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 80/100\n",
            "19/19 [==============================] - 1s 47ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 81/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 82/100\n",
            "19/19 [==============================] - 1s 58ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 83/100\n",
            "19/19 [==============================] - 1s 47ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 84/100\n",
            "19/19 [==============================] - 1s 44ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 85/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 86/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 87/100\n",
            "19/19 [==============================] - 1s 58ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 88/100\n",
            "19/19 [==============================] - 1s 68ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 89/100\n",
            "19/19 [==============================] - 1s 67ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 90/100\n",
            "19/19 [==============================] - 1s 68ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 91/100\n",
            "19/19 [==============================] - 1s 49ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 92/100\n",
            "19/19 [==============================] - 1s 46ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 93/100\n",
            "19/19 [==============================] - 1s 47ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 94/100\n",
            "19/19 [==============================] - 1s 45ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 95/100\n",
            "19/19 [==============================] - 1s 47ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 96/100\n",
            "19/19 [==============================] - 1s 47ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 97/100\n",
            "19/19 [==============================] - 1s 52ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 98/100\n",
            "19/19 [==============================] - 1s 56ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 99/100\n",
            "19/19 [==============================] - 1s 47ms/step - loss: 0.0064 - val_loss: 0.0064\n",
            "Epoch 100/100\n",
            "19/19 [==============================] - 1s 50ms/step - loss: 0.0064 - val_loss: 0.0064\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f21db7556f0>"
            ]
          },
          "metadata": {},
          "execution_count": 189
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "# Plot the loss and val_loss\n",
        "epochs = range(1, len(loss) + 1)\n",
        "plt.plot(epochs, loss, 'b', label='Training loss')\n",
        "plt.plot(epochs, val_loss, 'r', label='Validation loss')\n",
        "plt.title('Training and Validation Loss')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "n1V5OE-f51ny",
        "outputId": "7c4e9d4c-5a41-4089-b486-4c88407b98e0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def generate_speech(text, model):\n",
        "    # Preprocess the input text\n",
        "    input_text = pad_sequences([text_to_sequence(text)], dtype='int32', padding='post', maxlen=140)\n",
        "    input_text = np.expand_dims(input_text, axis=1)\n",
        "\n",
        "    generated_speech = model.predict([input_text])\n",
        "\n",
        "    return generated_speech\n",
        "\n",
        "# Use the function to generate speech\n",
        "text = \"안녕하세요 반갑습니다. 저는 발표자입니다.\"\n",
        "generated_speech = generate_speech(text, text_to_voice_model)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jOUO0SpohZ7t",
        "outputId": "5adeb051-0dbe-4e34-88fd-5b52d5785de7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 1s 1s/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "generated_speech.flatten().shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xnXXrJachdBV",
        "outputId": "e9761244-780f-440f-8edb-5912461e195f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(80000,)"
            ]
          },
          "metadata": {},
          "execution_count": 192
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "generated_speech.T"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q_80vmYFC06O",
        "outputId": "91cc121c-fe4e-41a3-8f4e-255e163697b9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[[ 0.00013245],\n",
              "        [ 0.00013245],\n",
              "        [ 0.00013245],\n",
              "        ...,\n",
              "        [ 0.00013245],\n",
              "        [ 0.00013245],\n",
              "        [ 0.00013245]],\n",
              "\n",
              "       [[ 0.00216484],\n",
              "        [ 0.00216484],\n",
              "        [ 0.00216484],\n",
              "        ...,\n",
              "        [ 0.00216484],\n",
              "        [ 0.00216484],\n",
              "        [ 0.00216484]],\n",
              "\n",
              "       [[-0.00031283],\n",
              "        [-0.00031283],\n",
              "        [-0.00031283],\n",
              "        ...,\n",
              "        [-0.00031283],\n",
              "        [-0.00031283],\n",
              "        [-0.00031283]],\n",
              "\n",
              "       ...,\n",
              "\n",
              "       [[ 0.00104678],\n",
              "        [ 0.00104678],\n",
              "        [ 0.00104678],\n",
              "        ...,\n",
              "        [ 0.00104678],\n",
              "        [ 0.00104678],\n",
              "        [ 0.00104678]],\n",
              "\n",
              "       [[ 0.00123615],\n",
              "        [ 0.00123615],\n",
              "        [ 0.00123615],\n",
              "        ...,\n",
              "        [ 0.00123615],\n",
              "        [ 0.00123615],\n",
              "        [ 0.00123615]],\n",
              "\n",
              "       [[ 0.00087974],\n",
              "        [ 0.00087974],\n",
              "        [ 0.00087974],\n",
              "        ...,\n",
              "        [ 0.00087974],\n",
              "        [ 0.00087974],\n",
              "        [ 0.00087974]]], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 193
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import soundfile as sf\n",
        "reshaped_array = generated_speech.flatten()\n",
        "mp3_file_path = \"/content/drive/MyDrive/generated_voice_5.mp3\"  # Replace with the desired output file path\n",
        "sf.write(mp3_file_path, reshaped_array, 22050, format='mp3')"
      ],
      "metadata": {
        "id": "u-sXCcIQhbzk"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}