{
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "id": "hVaj2LWCWxZA",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "309371cc-c56a-46de-df46-b004ed7de597"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Download and untar the yesno dataset"
      ],
      "metadata": {
        "id": "t6wXW_Ifaobx"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!wget https://us.openslr.org/resources/1/waves_yesno.tar.gz\n",
        "!mkdir working\n",
        "!tar -xzvf waves_yesno.tar.gz -C working"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TIM0W34CMr4i",
        "outputId": "50e1b1ad-6226-4a55-e65c-b84889c7d27f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2024-05-10 19:41:43--  https://us.openslr.org/resources/1/waves_yesno.tar.gz\n",
            "Resolving us.openslr.org (us.openslr.org)... 46.101.158.64\n",
            "Connecting to us.openslr.org (us.openslr.org)|46.101.158.64|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 4703754 (4.5M) [application/x-gzip]\n",
            "Saving to: ‘waves_yesno.tar.gz’\n",
            "\n",
            "waves_yesno.tar.gz  100%[===================>]   4.49M  5.68MB/s    in 0.8s    \n",
            "\n",
            "2024-05-10 19:41:44 (5.68 MB/s) - ‘waves_yesno.tar.gz’ saved [4703754/4703754]\n",
            "\n",
            "waves_yesno/\n",
            "waves_yesno/1_0_0_0_0_0_1_1.wav\n",
            "waves_yesno/1_1_0_0_1_0_1_0.wav\n",
            "waves_yesno/1_0_1_1_1_1_0_1.wav\n",
            "waves_yesno/1_1_1_1_0_1_0_0.wav\n",
            "waves_yesno/0_0_1_1_1_0_0_0.wav\n",
            "waves_yesno/0_1_1_1_1_1_1_1.wav\n",
            "waves_yesno/0_1_0_1_1_1_0_0.wav\n",
            "waves_yesno/1_0_1_1_1_0_1_0.wav\n",
            "waves_yesno/1_0_0_1_0_1_1_1.wav\n",
            "waves_yesno/0_0_1_0_1_0_0_0.wav\n",
            "waves_yesno/0_1_0_1_1_0_1_0.wav\n",
            "waves_yesno/0_0_1_1_0_1_1_0.wav\n",
            "waves_yesno/1_0_0_0_1_0_0_1.wav\n",
            "waves_yesno/1_1_0_1_1_1_1_0.wav\n",
            "waves_yesno/0_0_1_1_1_1_0_0.wav\n",
            "waves_yesno/1_1_0_0_1_1_1_0.wav\n",
            "waves_yesno/0_0_1_1_0_1_1_1.wav\n",
            "waves_yesno/1_1_0_1_0_1_1_0.wav\n",
            "waves_yesno/0_1_0_0_0_1_1_0.wav\n",
            "waves_yesno/0_0_0_1_0_0_0_1.wav\n",
            "waves_yesno/0_0_1_0_1_0_1_1.wav\n",
            "waves_yesno/0_0_1_0_0_0_1_0.wav\n",
            "waves_yesno/1_1_0_1_1_0_0_1.wav\n",
            "waves_yesno/0_1_1_1_0_1_0_1.wav\n",
            "waves_yesno/0_1_1_1_0_0_0_0.wav\n",
            "waves_yesno/README~\n",
            "waves_yesno/0_1_0_0_0_1_0_0.wav\n",
            "waves_yesno/1_0_0_0_0_0_0_1.wav\n",
            "waves_yesno/1_1_0_1_1_0_1_1.wav\n",
            "waves_yesno/1_1_0_0_0_0_0_1.wav\n",
            "waves_yesno/1_0_0_0_0_0_0_0.wav\n",
            "waves_yesno/0_1_1_1_1_0_1_0.wav\n",
            "waves_yesno/0_0_1_1_0_1_0_0.wav\n",
            "waves_yesno/1_1_1_0_0_0_0_1.wav\n",
            "waves_yesno/1_0_1_0_1_0_0_1.wav\n",
            "waves_yesno/0_1_0_0_1_0_1_1.wav\n",
            "waves_yesno/0_0_1_1_1_1_1_0.wav\n",
            "waves_yesno/1_1_0_0_0_1_1_1.wav\n",
            "waves_yesno/0_1_1_1_0_0_1_0.wav\n",
            "waves_yesno/1_1_0_1_0_1_0_0.wav\n",
            "waves_yesno/1_1_1_1_1_1_1_1.wav\n",
            "waves_yesno/0_0_1_0_1_0_0_1.wav\n",
            "waves_yesno/1_1_1_1_0_0_1_0.wav\n",
            "waves_yesno/0_0_1_1_1_0_0_1.wav\n",
            "waves_yesno/0_1_0_1_0_0_0_0.wav\n",
            "waves_yesno/1_1_1_1_1_0_0_0.wav\n",
            "waves_yesno/README\n",
            "waves_yesno/0_1_1_0_0_1_1_1.wav\n",
            "waves_yesno/0_0_1_0_0_1_1_0.wav\n",
            "waves_yesno/1_1_0_0_1_0_1_1.wav\n",
            "waves_yesno/1_1_1_0_0_1_0_1.wav\n",
            "waves_yesno/0_0_1_0_0_1_1_1.wav\n",
            "waves_yesno/0_0_1_1_0_0_0_1.wav\n",
            "waves_yesno/1_0_1_1_0_1_1_1.wav\n",
            "waves_yesno/1_1_1_0_1_0_1_0.wav\n",
            "waves_yesno/1_1_1_0_1_0_1_1.wav\n",
            "waves_yesno/0_1_0_0_1_0_1_0.wav\n",
            "waves_yesno/1_1_1_0_0_1_1_1.wav\n",
            "waves_yesno/0_1_1_0_0_1_1_0.wav\n",
            "waves_yesno/0_0_0_1_0_1_1_0.wav\n",
            "waves_yesno/1_1_1_1_1_1_0_0.wav\n",
            "waves_yesno/0_0_0_0_1_1_1_1.wav\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "import re\n",
        "\n",
        "# Directory containing the audio files\n",
        "audio_dir = '/content/working/waves_yesno/'\n",
        "\n",
        "# Create a dictionary to store audio paths and text transcriptions\n",
        "audio_transcription_dict = {}\n",
        "\n",
        "# Iterate through the audio files in the directory\n",
        "for filename in os.listdir(audio_dir):\n",
        "    if filename.endswith('.wav'):\n",
        "        # Extract the numeric labels from the filename\n",
        "        labels = re.findall(r'\\d+', filename)\n",
        "\n",
        "        labels = [int(label) for label in labels if label.isdigit()]\n",
        "\n",
        "        # Convert the numeric labels to \"yes\" and \"no\"\n",
        "        transcription = ' '.join(['yes' if label == 1 else 'no' for label in labels])\n",
        "\n",
        "\n",
        "        # Store the audio path and transcription in the dictionary\n",
        "        audio_path = os.path.join(audio_dir, filename)\n",
        "        audio_transcription_dict[audio_path] = transcription\n",
        "\n",
        "# Sort the dictionary by transcription (which ensures both lists are in the same order)\n",
        "sorted_audio_transcription = sorted(audio_transcription_dict.items(), key=lambda x: x[1])\n",
        "\n",
        "# Separate the sorted data into audio paths and transcriptions\n",
        "sorted_audio_paths, sorted_transcriptions = zip(*sorted_audio_transcription)\n",
        "\n",
        "# Create a text file to store the audio paths\n",
        "paths_output_file = 'paths.txt'\n",
        "with open(paths_output_file, 'w') as f_paths:\n",
        "    # Write each audio path on a separate line\n",
        "    f_paths.write('\\n'.join(sorted_audio_paths))\n",
        "\n",
        "print(f\"Audio paths have been saved to {paths_output_file}\")\n",
        "\n",
        "# Create a text file to store the transcriptions\n",
        "transcriptions_output_file = 'transcriptions.txt'\n",
        "with open(transcriptions_output_file, 'w') as f_transcriptions:\n",
        "    # Write each transcription on a separate line\n",
        "    f_transcriptions.write('\\n'.join(sorted_transcriptions))\n",
        "\n",
        "print(f\"Transcriptions have been saved to {transcriptions_output_file}\")\n"
      ],
      "metadata": {
        "id": "Y6xhhB_ATQzW",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "160f138d-ad36-41d2-baca-77c9f290992d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Audio paths have been saved to paths.txt\n",
            "Transcriptions have been saved to transcriptions.txt\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "WHKy_KEgRkYH",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "72c4369e-815e-49e6-88f1-7c49f848253f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting jiwer\n",
            "  Downloading jiwer-3.0.4-py3-none-any.whl (21 kB)\n",
            "Requirement already satisfied: click<9.0.0,>=8.1.3 in /usr/local/lib/python3.10/dist-packages (from jiwer) (8.1.7)\n",
            "Collecting rapidfuzz<4,>=3 (from jiwer)\n",
            "  Downloading rapidfuzz-3.9.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (3.4 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m3.4/3.4 MB\u001b[0m \u001b[31m23.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hInstalling collected packages: rapidfuzz, jiwer\n",
            "Successfully installed jiwer-3.0.4 rapidfuzz-3.9.0\n"
          ]
        }
      ],
      "source": [
        "# @title install required libs and import them\n",
        "!pip install jiwer\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "import matplotlib.pyplot as plt\n",
        "from IPython import display\n",
        "from jiwer import wer\n",
        "\n",
        "from tensorflow.keras.preprocessing.text import Tokenizer"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "audio_files_path = \"/content/paths.txt\"\n",
        "texttranscribe_path = \"/content/transcriptions.txt\"\n",
        "\n",
        "# Load the audio file paths and corresponding text labels\n",
        "with open(audio_files_path, 'r') as f:\n",
        "    audio_files = f.read().splitlines()\n",
        "with open(texttranscribe_path, 'r') as f:\n",
        "    audio_labels_text = f.read().splitlines()\n",
        "# Create a dictionary with the data\n",
        "data = {'Audio_Path': audio_files, 'Transcription': audio_labels_text}\n",
        "\n",
        "# Create a Pandas DataFrame\n",
        "df = pd.DataFrame(data)\n",
        "\n",
        "metadata_df  = df\n",
        "\n",
        "# Display the first few rows of the DataFrame\n",
        "print(df.head())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Rya-e-KYTQju",
        "outputId": "67c8db63-6353-40f5-aa56-bd3285cf030e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "                                         Audio_Path  \\\n",
            "0  /content/working/waves_yesno/0_0_0_0_1_1_1_1.wav   \n",
            "1  /content/working/waves_yesno/0_0_0_1_0_0_0_1.wav   \n",
            "2  /content/working/waves_yesno/0_0_0_1_0_1_1_0.wav   \n",
            "3  /content/working/waves_yesno/0_0_1_0_0_0_1_0.wav   \n",
            "4  /content/working/waves_yesno/0_0_1_0_0_1_1_0.wav   \n",
            "\n",
            "                 Transcription  \n",
            "0  no no no no yes yes yes yes  \n",
            "1    no no no yes no no no yes  \n",
            "2   no no no yes no yes yes no  \n",
            "3    no no yes no no no yes no  \n",
            "4   no no yes no no yes yes no  \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "text = audio_labels_text\n",
        "\n",
        "\n",
        "charlevel=True\n",
        "\n",
        "# Tokenize the labels and convert to binary matrix\n",
        "tokenizer = Tokenizer(char_level=charlevel, oov_token='<UNK>')\n",
        "tokenizer.fit_on_texts(text)\n",
        "\n",
        "audio_labels = tokenizer.texts_to_sequences(text)\n",
        "\n",
        "word_index = tokenizer.word_index\n",
        "print(word_index)\n",
        "\n",
        "# Extract only the characters (keys) from the dictionary\n",
        "characters = [char for char in word_index.keys()]\n",
        "\n",
        "# Join the list elements into a single string\n",
        "vocabresult = ''.join(characters)\n",
        "\n",
        "# Print the result\n",
        "print(vocabresult)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0YXztnyueesT",
        "outputId": "ebc0d02c-0701-44fd-bac3-9dd3a261420c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "{'<UNK>': 1, ' ': 2, 'y': 3, 'e': 4, 's': 5, 'n': 6, 'o': 7}\n",
            "<UNK> yesno\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7aQkJNTTRkYM",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "11a1bf98-abc0-45cc-9e4f-4651b0cc7241"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Size of the training set: 54\n",
            "Size of the val set: 6\n"
          ]
        }
      ],
      "source": [
        "split = int(len(metadata_df) * 0.90)\n",
        "df_train = metadata_df[:split]\n",
        "df_val = metadata_df[split:]\n",
        "\n",
        "print(f\"Size of the training set: {len(df_train)}\")\n",
        "print(f\"Size of the val set: {len(df_val)}\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6ts1U5VfRkYN"
      },
      "source": [
        "## Preprocessing\n",
        "\n",
        "We first prepare the vocabulary to be used."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "huysjs_qRkYO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e7e2d236-b145-4364-8c08-394455f8965e"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The vocabulary is: ['', '<', 'U', 'N', 'K', '>', ' ', 'y', 'e', 's', 'n', 'o'] (size =12)\n"
          ]
        }
      ],
      "source": [
        "# The set of characters accepted in the transcription.\n",
        "#characters = [x for x in \"abcdefghijklmnopqrstuvwxyz'?! \"]\n",
        "characters = [x for x in vocabresult]\n",
        "# Mapping characters to integers\n",
        "char_to_num = keras.layers.StringLookup(vocabulary=characters, oov_token=\"\")\n",
        "# Mapping integers back to original characters\n",
        "num_to_char = keras.layers.StringLookup(\n",
        "    vocabulary=char_to_num.get_vocabulary(), oov_token=\"\", invert=True\n",
        ")\n",
        "\n",
        "print(\n",
        "    f\"The vocabulary is: {char_to_num.get_vocabulary()} \"\n",
        "    f\"(size ={char_to_num.vocabulary_size()})\"\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8DyXAkBRRkYP"
      },
      "source": [
        "Next, we create the function that describes the transformation that we apply to each\n",
        "element of our dataset."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ODBBRjudRkYP"
      },
      "outputs": [],
      "source": [
        "# An integer scalar Tensor. The window length in samples.\n",
        "frame_length = 256\n",
        "# An integer scalar Tensor. The number of samples to step.\n",
        "frame_step = 160\n",
        "# An integer scalar Tensor. The size of the FFT to apply.\n",
        "# If not provided, uses the smallest power of 2 enclosing frame_length.\n",
        "fft_length = 384\n",
        "\n",
        "\n",
        "def encode_single_sample(wav_file, label):\n",
        "    ###########################################\n",
        "    ##  Process the Audio\n",
        "    ##########################################\n",
        "    # 1. Read wav file\n",
        "    #file = tf.io.read_file(wavs_path + wav_file + \".wav\")\n",
        "    file = tf.io.read_file(wav_file)\n",
        "    # 2. Decode the wav file\n",
        "    audio, _ = tf.audio.decode_wav(file)\n",
        "    audio = tf.squeeze(audio, axis=-1)\n",
        "    # 3. Change type to float\n",
        "    audio = tf.cast(audio, tf.float32)\n",
        "    # 4. Get the spectrogram\n",
        "    spectrogram = tf.signal.stft(\n",
        "        audio, frame_length=frame_length, frame_step=frame_step, fft_length=fft_length\n",
        "    )\n",
        "    # 5. We only need the magnitude, which can be derived by applying tf.abs\n",
        "    spectrogram = tf.abs(spectrogram)\n",
        "    spectrogram = tf.math.pow(spectrogram, 0.5)\n",
        "    # 6. normalisation\n",
        "    means = tf.math.reduce_mean(spectrogram, 1, keepdims=True)\n",
        "    stddevs = tf.math.reduce_std(spectrogram, 1, keepdims=True)\n",
        "    spectrogram = (spectrogram - means) / (stddevs + 1e-10)\n",
        "    ###########################################\n",
        "    ##  Process the label\n",
        "    ##########################################\n",
        "    # 7. Convert label to Lower case\n",
        "    label = tf.strings.lower(label)\n",
        "    # 8. Split the label\n",
        "    label = tf.strings.unicode_split(label, input_encoding=\"UTF-8\")\n",
        "    # 9. Map the characters in label to numbers\n",
        "    label = char_to_num(label)\n",
        "    # 10. Return a dict as our model is expecting two inputs\n",
        "    return spectrogram, label\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ReqsqF65RkYQ"
      },
      "source": [
        "## Creating `Dataset` objects\n",
        "\n",
        "We create a `tf.data.Dataset` object that yields\n",
        "the transformed elements, in the same order as they\n",
        "appeared in the input."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ZmCwCErURkYR"
      },
      "outputs": [],
      "source": [
        "batch_size = 32\n",
        "# Define the training dataset\n",
        "train_dataset = tf.data.Dataset.from_tensor_slices(\n",
        "    #old(list(df_train[\"file_name\"]), list(df_train[\"normalized_transcription\"]))\n",
        "    (list(df_train[\"Audio_Path\"]), list(df_train[\"Transcription\"]))\n",
        "\n",
        ")\n",
        "train_dataset = (\n",
        "    train_dataset.map(encode_single_sample, num_parallel_calls=tf.data.AUTOTUNE)\n",
        "    .padded_batch(batch_size)\n",
        "    .prefetch(buffer_size=tf.data.AUTOTUNE)\n",
        ")\n",
        "\n",
        "# Define the validation dataset\n",
        "validation_dataset = tf.data.Dataset.from_tensor_slices(\n",
        "    (list(df_val[\"Audio_Path\"]), list(df_val[\"Transcription\"]))\n",
        ")\n",
        "validation_dataset = (\n",
        "    validation_dataset.map(encode_single_sample, num_parallel_calls=tf.data.AUTOTUNE)\n",
        "    .padded_batch(batch_size)\n",
        "    .prefetch(buffer_size=tf.data.AUTOTUNE)\n",
        ")\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bvwBXcb_RkYR"
      },
      "source": [
        "## Visualize the data\n",
        "\n",
        "Let's visualize an example in our dataset, including the\n",
        "audio clip, the spectrogram and the corresponding label."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "WsKwBIiaRkYS",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 526
        },
        "outputId": "856f0452-bd3b-4695-d56d-1d438651622c"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.lib.display.Audio object>"
            ],
            "text/html": [
              "\n",
              "                <audio  controls=\"controls\" >\n",
              "                    <source src=\"data:audio/wav;base64,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\" type=\"audio/wav\" />\n",
              "                    Your browser does not support the audio element.\n",
              "                </audio>\n",
              "              "
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x500 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "fig = plt.figure(figsize=(8, 5))\n",
        "for batch in train_dataset.take(1):\n",
        "    spectrogram = batch[0][0].numpy()\n",
        "    spectrogram = np.array([np.trim_zeros(x) for x in np.transpose(spectrogram)])\n",
        "    label = batch[1][0]\n",
        "    # Spectrogram\n",
        "    label = tf.strings.reduce_join(num_to_char(label)).numpy().decode(\"utf-8\")\n",
        "    ax = plt.subplot(2, 1, 1)\n",
        "    ax.imshow(spectrogram, vmax=1)\n",
        "    ax.set_title(label)\n",
        "    ax.axis(\"off\")\n",
        "    # Wav\n",
        "    #file = tf.io.read_file(wavs_path + list(df_train[\"file_name\"])[0] + \".wav\")\n",
        "    file = tf.io.read_file(list(df_train[\"Audio_Path\"])[0])\n",
        "    audio, _ = tf.audio.decode_wav(file)\n",
        "    audio = audio.numpy()\n",
        "    ax = plt.subplot(2, 1, 2)\n",
        "    plt.plot(audio)\n",
        "    ax.set_title(\"Signal Wave\")\n",
        "    ax.set_xlim(0, len(audio))\n",
        "    display.display(display.Audio(np.transpose(audio), rate=16000))\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Pn9ingJaRkYS"
      },
      "source": [
        "## Model\n",
        "\n",
        "We first define the CTC Loss function."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7lXw7JNNRkYT"
      },
      "outputs": [],
      "source": [
        "\n",
        "def CTCLoss(y_true, y_pred):\n",
        "    # Compute the training-time loss value\n",
        "    batch_len = tf.cast(tf.shape(y_true)[0], dtype=\"int64\")\n",
        "    input_length = tf.cast(tf.shape(y_pred)[1], dtype=\"int64\")\n",
        "    label_length = tf.cast(tf.shape(y_true)[1], dtype=\"int64\")\n",
        "\n",
        "    input_length = input_length * tf.ones(shape=(batch_len, 1), dtype=\"int64\")\n",
        "    label_length = label_length * tf.ones(shape=(batch_len, 1), dtype=\"int64\")\n",
        "\n",
        "    loss = keras.backend.ctc_batch_cost(y_true, y_pred, input_length, label_length)\n",
        "    return loss\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JSYnFIIkRkYT"
      },
      "source": [
        "We now define our model. We will define a model similar to\n",
        "[DeepSpeech2](https://nvidia.github.io/OpenSeq2Seq/html/speech-recognition/deepspeech2.html)."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-D9m8LV2RkYT",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "40e00241-9de9-4daa-f6ef-4115e192b2cb"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"zmasr\"\n",
            "______________________________________________________________________________________________________________\n",
            " Layer (type)                                    Output Shape                                Param #          \n",
            "==============================================================================================================\n",
            " input (InputLayer)                              [(None, None, 193)]                         0                \n",
            "                                                                                                              \n",
            " expand_dim (Reshape)                            (None, None, 193, 1)                        0                \n",
            "                                                                                                              \n",
            " conv_1 (Conv2D)                                 (None, None, 97, 32)                        14432            \n",
            "                                                                                                              \n",
            " conv_1_bn (BatchNormalization)                  (None, None, 97, 32)                        128              \n",
            "                                                                                                              \n",
            " conv_1_relu (ReLU)                              (None, None, 97, 32)                        0                \n",
            "                                                                                                              \n",
            " conv_2 (Conv2D)                                 (None, None, 49, 32)                        236544           \n",
            "                                                                                                              \n",
            " conv_2_bn (BatchNormalization)                  (None, None, 49, 32)                        128              \n",
            "                                                                                                              \n",
            " conv_2_relu (ReLU)                              (None, None, 49, 32)                        0                \n",
            "                                                                                                              \n",
            " reshape (Reshape)                               (None, None, 1568)                          0                \n",
            "                                                                                                              \n",
            " bidirectional_1 (Bidirectional)                 (None, None, 1024)                          6395904          \n",
            "                                                                                                              \n",
            " dropout (Dropout)                               (None, None, 1024)                          0                \n",
            "                                                                                                              \n",
            " bidirectional_2 (Bidirectional)                 (None, None, 1024)                          4724736          \n",
            "                                                                                                              \n",
            " dropout_1 (Dropout)                             (None, None, 1024)                          0                \n",
            "                                                                                                              \n",
            " bidirectional_3 (Bidirectional)                 (None, None, 1024)                          4724736          \n",
            "                                                                                                              \n",
            " dropout_2 (Dropout)                             (None, None, 1024)                          0                \n",
            "                                                                                                              \n",
            " bidirectional_4 (Bidirectional)                 (None, None, 1024)                          4724736          \n",
            "                                                                                                              \n",
            " dropout_3 (Dropout)                             (None, None, 1024)                          0                \n",
            "                                                                                                              \n",
            " bidirectional_5 (Bidirectional)                 (None, None, 1024)                          4724736          \n",
            "                                                                                                              \n",
            " dense_1 (Dense)                                 (None, None, 1024)                          1049600          \n",
            "                                                                                                              \n",
            " dense_1_relu (ReLU)                             (None, None, 1024)                          0                \n",
            "                                                                                                              \n",
            " dropout_4 (Dropout)                             (None, None, 1024)                          0                \n",
            "                                                                                                              \n",
            " dense (Dense)                                   (None, None, 13)                            13325            \n",
            "                                                                                                              \n",
            "==============================================================================================================\n",
            "Total params: 26609005 (101.51 MB)\n",
            "Trainable params: 26608877 (101.50 MB)\n",
            "Non-trainable params: 128 (512.00 Byte)\n",
            "______________________________________________________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "\n",
        "def build_model(input_dim, output_dim, rnn_layers=5, rnn_units=128):\n",
        "    \"\"\"Model.\"\"\"\n",
        "    # Model's input\n",
        "    input_spectrogram = layers.Input((None, input_dim), name=\"input\")\n",
        "    # Expand the dimension to use 2D CNN.\n",
        "    x = layers.Reshape((-1, input_dim, 1), name=\"expand_dim\")(input_spectrogram)\n",
        "    # Convolution layer 1\n",
        "    x = layers.Conv2D(\n",
        "        filters=32,\n",
        "        kernel_size=[11, 41],\n",
        "        strides=[2, 2],\n",
        "        padding=\"same\",\n",
        "        use_bias=False,\n",
        "        name=\"conv_1\",\n",
        "    )(x)\n",
        "    x = layers.BatchNormalization(name=\"conv_1_bn\")(x)\n",
        "    x = layers.ReLU(name=\"conv_1_relu\")(x)\n",
        "    # Convolution layer 2\n",
        "    x = layers.Conv2D(\n",
        "        filters=32,\n",
        "        kernel_size=[11, 21],\n",
        "        strides=[1, 2],\n",
        "        padding=\"same\",\n",
        "        use_bias=False,\n",
        "        name=\"conv_2\",\n",
        "    )(x)\n",
        "    x = layers.BatchNormalization(name=\"conv_2_bn\")(x)\n",
        "    x = layers.ReLU(name=\"conv_2_relu\")(x)\n",
        "    # Reshape the resulted volume to feed the RNNs layers\n",
        "    x = layers.Reshape((-1, x.shape[-2] * x.shape[-1]))(x)\n",
        "    # RNN layers\n",
        "    for i in range(1, rnn_layers + 1):\n",
        "        recurrent = layers.GRU(\n",
        "            units=rnn_units,\n",
        "            activation=\"tanh\",\n",
        "            recurrent_activation=\"sigmoid\",\n",
        "            use_bias=True,\n",
        "            return_sequences=True,\n",
        "            reset_after=True,\n",
        "            name=f\"gru_{i}\",\n",
        "        )\n",
        "        x = layers.Bidirectional(\n",
        "            recurrent, name=f\"bidirectional_{i}\", merge_mode=\"concat\"\n",
        "        )(x)\n",
        "        if i < rnn_layers:\n",
        "            x = layers.Dropout(rate=0.5)(x)\n",
        "    # Dense layer\n",
        "    x = layers.Dense(units=rnn_units * 2, name=\"dense_1\")(x)\n",
        "    x = layers.ReLU(name=\"dense_1_relu\")(x)\n",
        "    x = layers.Dropout(rate=0.5)(x)\n",
        "    # Classification layer\n",
        "    output = layers.Dense(units=output_dim + 1, activation=\"softmax\")(x)\n",
        "    # Model\n",
        "    model = keras.Model(input_spectrogram, output, name=\"zmasr\")\n",
        "    # Optimizer\n",
        "    opt = keras.optimizers.Adam(learning_rate=1e-4)\n",
        "    # Compile the model and return\n",
        "    model.compile(optimizer=opt, loss=CTCLoss)\n",
        "    return model\n",
        "\n",
        "\n",
        "# Get the model\n",
        "model = build_model(\n",
        "    input_dim=fft_length // 2 + 1,\n",
        "    output_dim=char_to_num.vocabulary_size(),\n",
        "    rnn_units=512,\n",
        ")\n",
        "model.summary(line_length=110)"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "XgSVEpffqySU"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "t_oyQ58ERkYU"
      },
      "source": [
        "## Training and Evaluating"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "R0SD621WRkYU"
      },
      "outputs": [],
      "source": [
        "# A utility function to decode the output of the network\n",
        "def decode_batch_predictions(pred):\n",
        "    input_len = np.ones(pred.shape[0]) * pred.shape[1]\n",
        "    # Use greedy search. For complex tasks, you can use beam search\n",
        "    results = keras.backend.ctc_decode(pred, input_length=input_len, greedy=True)[0][0]\n",
        "    # Iterate over the results and get back the text\n",
        "    output_text = []\n",
        "    for result in results:\n",
        "        result = tf.strings.reduce_join(num_to_char(result)).numpy().decode(\"utf-8\")\n",
        "        output_text.append(result)\n",
        "    return output_text\n",
        "\n",
        "\n",
        "# A callback class to output a few transcriptions during training\n",
        "class CallbackEval(keras.callbacks.Callback):\n",
        "    \"\"\"Displays a batch of outputs after every epoch.\"\"\"\n",
        "\n",
        "    def __init__(self, dataset):\n",
        "        super().__init__()\n",
        "        self.dataset = dataset\n",
        "\n",
        "    def on_epoch_end(self, epoch: int, logs=None):\n",
        "        predictions = []\n",
        "        targets = []\n",
        "        for batch in self.dataset:\n",
        "            X, y = batch\n",
        "            batch_predictions = model.predict(X)\n",
        "            batch_predictions = decode_batch_predictions(batch_predictions)\n",
        "            predictions.extend(batch_predictions)\n",
        "            for label in y:\n",
        "                label = (\n",
        "                    tf.strings.reduce_join(num_to_char(label)).numpy().decode(\"utf-8\")\n",
        "                )\n",
        "                targets.append(label)\n",
        "        wer_score = wer(targets, predictions)\n",
        "        print(\"-\" * 100)\n",
        "        print(f\"Word Error Rate: {wer_score:.4f}\")\n",
        "        print(\"-\" * 100)\n",
        "        for i in np.random.randint(0, len(predictions), 2):\n",
        "            print(f\"Target    : {targets[i]}\")\n",
        "            print(f\"Prediction: {predictions[i]}\")\n",
        "            print(\"-\" * 100)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NavxqQegRkYV"
      },
      "source": [
        "Let's start the training process."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vugxVG2-RkYW",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ffbb0fc8-d6a4-4901-e542-450e2de51d1a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/200\n",
            "1/1 [==============================] - 4s 4s/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 54s 11s/step - loss: 280.3683 - val_loss: 190.5100\n",
            "Epoch 2/200\n",
            "1/1 [==============================] - 0s 75ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 839ms/step - loss: 86.6413 - val_loss: 74.3344\n",
            "Epoch 3/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 667ms/step - loss: 83.9221 - val_loss: 70.4005\n",
            "Epoch 4/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 680ms/step - loss: 74.4365 - val_loss: 71.7059\n",
            "Epoch 5/200\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 669ms/step - loss: 65.9256 - val_loss: 74.2275\n",
            "Epoch 6/200\n",
            "1/1 [==============================] - 0s 65ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 665ms/step - loss: 63.7472 - val_loss: 64.9777\n",
            "Epoch 7/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 666ms/step - loss: 59.0318 - val_loss: 72.0029\n",
            "Epoch 8/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 668ms/step - loss: 61.5732 - val_loss: 68.8926\n",
            "Epoch 9/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 688ms/step - loss: 57.8475 - val_loss: 61.7637\n",
            "Epoch 10/200\n",
            "1/1 [==============================] - 0s 81ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 56.3472 - val_loss: 59.8899\n",
            "Epoch 11/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 811ms/step - loss: 56.9523 - val_loss: 60.8580\n",
            "Epoch 12/200\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 687ms/step - loss: 55.3890 - val_loss: 65.2349\n",
            "Epoch 13/200\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 672ms/step - loss: 55.4309 - val_loss: 66.3740\n",
            "Epoch 14/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 685ms/step - loss: 54.9646 - val_loss: 62.1246\n",
            "Epoch 15/200\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 670ms/step - loss: 53.7092 - val_loss: 59.1405\n",
            "Epoch 16/200\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 672ms/step - loss: 54.0880 - val_loss: 58.7334\n",
            "Epoch 17/200\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 690ms/step - loss: 53.4504 - val_loss: 60.5394\n",
            "Epoch 18/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 695ms/step - loss: 52.8594 - val_loss: 62.2759\n",
            "Epoch 19/200\n",
            "1/1 [==============================] - 0s 94ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 946ms/step - loss: 53.1221 - val_loss: 60.4830\n",
            "Epoch 20/200\n",
            "1/1 [==============================] - 0s 92ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 52.5530 - val_loss: 57.9976\n",
            "Epoch 21/200\n",
            "1/1 [==============================] - 0s 111ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 52.3053 - val_loss: 57.3616\n",
            "Epoch 22/200\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1000ms/step - loss: 51.9913 - val_loss: 58.6424\n",
            "Epoch 23/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 841ms/step - loss: 51.9597 - val_loss: 59.6710\n",
            "Epoch 24/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 681ms/step - loss: 51.7355 - val_loss: 58.4488\n",
            "Epoch 25/200\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 852ms/step - loss: 51.3883 - val_loss: 57.2756\n",
            "Epoch 26/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 708ms/step - loss: 51.4069 - val_loss: 57.7176\n",
            "Epoch 27/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 707ms/step - loss: 51.1882 - val_loss: 58.7833\n",
            "Epoch 28/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 705ms/step - loss: 50.9066 - val_loss: 57.7158\n",
            "Epoch 29/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 703ms/step - loss: 50.5527 - val_loss: 56.7773\n",
            "Epoch 30/200\n",
            "1/1 [==============================] - 0s 75ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 50.3671 - val_loss: 58.6454\n",
            "Epoch 31/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 895ms/step - loss: 50.2194 - val_loss: 59.6487\n",
            "Epoch 32/200\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 686ms/step - loss: 49.7212 - val_loss: 57.4765\n",
            "Epoch 33/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 691ms/step - loss: 49.3390 - val_loss: 58.0322\n",
            "Epoch 34/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 696ms/step - loss: 48.7770 - val_loss: 60.4925\n",
            "Epoch 35/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 710ms/step - loss: 47.9643 - val_loss: 59.0265\n",
            "Epoch 36/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 705ms/step - loss: 47.3073 - val_loss: 59.1847\n",
            "Epoch 37/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 717ms/step - loss: 46.2350 - val_loss: 62.4832\n",
            "Epoch 38/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 705ms/step - loss: 44.9732 - val_loss: 60.8203\n",
            "Epoch 39/200\n",
            "1/1 [==============================] - 0s 76ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 975ms/step - loss: 43.6342 - val_loss: 62.6938\n",
            "Epoch 40/200\n",
            "1/1 [==============================] - 0s 82ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 42.5279 - val_loss: 67.4025\n",
            "Epoch 41/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 692ms/step - loss: 41.5513 - val_loss: 67.6657\n",
            "Epoch 42/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 697ms/step - loss: 40.8318 - val_loss: 73.1254\n",
            "Epoch 43/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 703ms/step - loss: 40.6209 - val_loss: 78.6465\n",
            "Epoch 44/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 713ms/step - loss: 39.9639 - val_loss: 73.6232\n",
            "Epoch 45/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 718ms/step - loss: 38.8675 - val_loss: 82.8925\n",
            "Epoch 46/200\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 702ms/step - loss: 38.5207 - val_loss: 76.4856\n",
            "Epoch 47/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 697ms/step - loss: 37.6787 - val_loss: 69.6204\n",
            "Epoch 48/200\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 851ms/step - loss: 37.1372 - val_loss: 72.7817\n",
            "Epoch 49/200\n",
            "1/1 [==============================] - 0s 76ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 973ms/step - loss: 36.5018 - val_loss: 70.9699\n",
            "Epoch 50/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 874ms/step - loss: 35.9560 - val_loss: 69.9423\n",
            "Epoch 51/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 715ms/step - loss: 35.1836 - val_loss: 72.0382\n",
            "Epoch 52/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 850ms/step - loss: 34.3618 - val_loss: 79.1895\n",
            "Epoch 53/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 716ms/step - loss: 33.1099 - val_loss: 74.7553\n",
            "Epoch 54/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 709ms/step - loss: 31.4292 - val_loss: 71.6617\n",
            "Epoch 55/200\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 723ms/step - loss: 29.0624 - val_loss: 79.6929\n",
            "Epoch 56/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: y\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 785ms/step - loss: 27.4510 - val_loss: 81.9462\n",
            "Epoch 57/200\n",
            "1/1 [==============================] - 0s 83ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 905ms/step - loss: 25.4366 - val_loss: 76.9334\n",
            "Epoch 58/200\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: eeyeye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 907ms/step - loss: 23.9502 - val_loss: 64.2882\n",
            "Epoch 59/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: \n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yeyeye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 706ms/step - loss: 22.3700 - val_loss: 47.6921\n",
            "Epoch 60/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: eyeyesyesyeyeyese\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: eyeyesyesyeyeyese\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 20.3380 - val_loss: 46.8438\n",
            "Epoch 61/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yesyesyesyeyeyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: esyesyesyesyeses\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 836ms/step - loss: 18.8931 - val_loss: 43.5477\n",
            "Epoch 62/200\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eeees\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: esyesyesyesyesyesyese\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 737ms/step - loss: 17.7420 - val_loss: 50.6988\n",
            "Epoch 63/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: eesesyes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyesyesyesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 712ms/step - loss: 16.7118 - val_loss: 54.4376\n",
            "Epoch 64/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eeeees\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyesyesyesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 713ms/step - loss: 16.0422 - val_loss: 55.7261\n",
            "Epoch 65/200\n",
            "1/1 [==============================] - 0s 82ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyesyesyesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyesyesyesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 14.8190 - val_loss: 63.4300\n",
            "Epoch 66/200\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: eesesyesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: eesesyesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 13.8171 - val_loss: 69.2596\n",
            "Epoch 67/200\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esesesesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yeseseseseses\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 728ms/step - loss: 12.8497 - val_loss: 79.3719\n",
            "Epoch 68/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: esyesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: eeesesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 718ms/step - loss: 11.8487 - val_loss: 84.4914\n",
            "Epoch 69/200\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eeeees\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eeeees\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 707ms/step - loss: 10.8563 - val_loss: 87.1034\n",
            "Epoch 70/200\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: eseseso yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yesyesyeseseses\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 719ms/step - loss: 9.6162 - val_loss: 74.7305\n",
            "Epoch 71/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: eseseso yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eeeees\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 729ms/step - loss: 8.9566 - val_loss: 67.3843\n",
            "Epoch 72/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: esesoyesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: eesesoyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 710ms/step - loss: 7.9415 - val_loss: 77.9798\n",
            "Epoch 73/200\n",
            "1/1 [==============================] - 0s 88ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: eesyesesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: eeseso yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 981ms/step - loss: 7.0656 - val_loss: 76.4265\n",
            "Epoch 74/200\n",
            "1/1 [==============================] - 0s 81ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9583\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyesyesyesonoyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: esyes oyeso yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 6.3355 - val_loss: 50.5775\n",
            "Epoch 75/200\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9583\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yesyesyesyesyesyeso\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: eyesyesyesyesyesyese\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 875ms/step - loss: 5.7279 - val_loss: 48.1116\n",
            "Epoch 76/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: esyesyeseees\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: eyesyesyesyesyesyee\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 870ms/step - loss: 5.3120 - val_loss: 75.6176\n",
            "Epoch 77/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: eyesyesyesonyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: eyesyesyesyesyesyese\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 868ms/step - loss: 4.8044 - val_loss: 66.4928\n",
            "Epoch 78/200\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: eyesyesyesyesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yesyesyesyesyesyeso n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 746ms/step - loss: 4.2239 - val_loss: 42.4932\n",
            "Epoch 79/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: eyesyesyesonoyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: esyesyesno yesoyeses\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 720ms/step - loss: 4.1175 - val_loss: 52.5719\n",
            "Epoch 80/200\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: eyesyesyesonoyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: eyesesnoyesoyeses\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 747ms/step - loss: 4.1333 - val_loss: 67.0228\n",
            "Epoch 81/200\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yesyesyesyesyeyeson\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: eyesyesyesyesyesyese\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 875ms/step - loss: 3.6794 - val_loss: 59.1241\n",
            "Epoch 82/200\n",
            "1/1 [==============================] - 0s 87ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9583\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yesyesyesyesyesyesno n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yesyesyesyesyesyesno n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 973ms/step - loss: 3.3159 - val_loss: 42.5774\n",
            "Epoch 83/200\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: esyesyes no yesno yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyeseyesesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 743ms/step - loss: 3.2797 - val_loss: 37.5323\n",
            "Epoch 84/200\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: eyesyesyesyesyesyese\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: eyesyesyesyesyesyese\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 731ms/step - loss: 3.2228 - val_loss: 53.0730\n",
            "Epoch 85/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eeeeso\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eeeeso\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 724ms/step - loss: 3.1060 - val_loss: 62.3882\n",
            "Epoch 86/200\n",
            "1/1 [==============================] - 0s 74ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eeeesooy\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: eyesyesyesonoyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 744ms/step - loss: 3.0630 - val_loss: 59.2805\n",
            "Epoch 87/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 1.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: esyesyesno yesno yeses\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: esyesyesno yesno yeses\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 864ms/step - loss: 2.9543 - val_loss: 51.2364\n",
            "Epoch 88/200\n",
            "1/1 [==============================] - 0s 74ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yesyesyesyesyesyesno n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yesyesyesyesyesyesno n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 733ms/step - loss: 2.6354 - val_loss: 46.2535\n",
            "Epoch 89/200\n",
            "1/1 [==============================] - 0s 94ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yesyesyesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yesyesyesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 2.8762 - val_loss: 47.3027\n",
            "Epoch 90/200\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: eyesyesyesyesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eeeeesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 755ms/step - loss: 2.6828 - val_loss: 51.0762\n",
            "Epoch 91/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: eeyesyesnoyesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: eeyesyesnoyesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 726ms/step - loss: 2.4903 - val_loss: 47.2116\n",
            "Epoch 92/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8958\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yesyesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yesno yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 719ms/step - loss: 2.7518 - val_loss: 35.7837\n",
            "Epoch 93/200\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8125\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yesyesno no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: esesyes yesno yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 729ms/step - loss: 2.4681 - val_loss: 27.8956\n",
            "Epoch 94/200\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8958\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: esesyes yesno yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: esesyes yesno yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 728ms/step - loss: 2.3752 - val_loss: 35.2397\n",
            "Epoch 95/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyeyesesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: esyesyes no yesno yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 714ms/step - loss: 2.6356 - val_loss: 42.0822\n",
            "Epoch 96/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: eesyesyesno yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: eesyesyesno yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 728ms/step - loss: 2.3680 - val_loss: 46.1826\n",
            "Epoch 97/200\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: esyesyes no yesno yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: eyesyes yesyesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 996ms/step - loss: 2.2399 - val_loss: 45.9417\n",
            "Epoch 98/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8958\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yesyesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yesno yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 783ms/step - loss: 2.3434 - val_loss: 38.5010\n",
            "Epoch 99/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8958\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: esyesyes yesyesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: esesyes yesno yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 729ms/step - loss: 2.3613 - val_loss: 34.7916\n",
            "Epoch 100/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8958\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: esyesyes yesyesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yesyesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 727ms/step - loss: 2.2511 - val_loss: 35.8907\n",
            "Epoch 101/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9583\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: es yesyesyesyesyesno n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyeyesesno no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 868ms/step - loss: 2.2026 - val_loss: 43.3531\n",
            "Epoch 102/200\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: esyesyes yesyesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eeeyeesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 879ms/step - loss: 2.1807 - val_loss: 49.7006\n",
            "Epoch 103/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9583\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: esyesyes yesyesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: esyesyesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 883ms/step - loss: 2.1129 - val_loss: 48.1502\n",
            "Epoch 104/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9375\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: es yesyesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: esyesyes yesyesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 872ms/step - loss: 2.0269 - val_loss: 44.7339\n",
            "Epoch 105/200\n",
            "1/1 [==============================] - 0s 76ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8750\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: es yesyesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: es yesyesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 910ms/step - loss: 2.1771 - val_loss: 37.7323\n",
            "Epoch 106/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8958\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyesno no yen\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yesyesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 950ms/step - loss: 2.0686 - val_loss: 35.8053\n",
            "Epoch 107/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.9167\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: esyesyes yesyesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yesno yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 865ms/step - loss: 2.2343 - val_loss: 30.6255\n",
            "Epoch 108/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8542\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: esesyes yesno yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: esesyes yesno yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 719ms/step - loss: 1.9240 - val_loss: 31.4071\n",
            "Epoch 109/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8542\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yesyesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 873ms/step - loss: 2.1481 - val_loss: 28.8162\n",
            "Epoch 110/200\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8542\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yesyesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 864ms/step - loss: 1.9513 - val_loss: 34.7927\n",
            "Epoch 111/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8542\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yesyes yes yesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 869ms/step - loss: 2.2712 - val_loss: 32.4843\n",
            "Epoch 112/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8750\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyesno no yen\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yesyesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 861ms/step - loss: 2.0218 - val_loss: 32.4852\n",
            "Epoch 113/200\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7500\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yesyesyesyesyesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 858ms/step - loss: 2.3283 - val_loss: 23.5414\n",
            "Epoch 114/200\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8125\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyesno no n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: esyesyes yesno yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 970ms/step - loss: 1.9481 - val_loss: 22.6650\n",
            "Epoch 115/200\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7292\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyesno no n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 895ms/step - loss: 1.9019 - val_loss: 20.6532\n",
            "Epoch 116/200\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7083\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yes yesno no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yesyes yes yesyes yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 871ms/step - loss: 1.7562 - val_loss: 20.2097\n",
            "Epoch 117/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7500\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yesyesyesyesyesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yesyes yes yesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 717ms/step - loss: 1.8052 - val_loss: 22.2932\n",
            "Epoch 118/200\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.6875\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yesyes yes yesyes yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 734ms/step - loss: 1.7678 - val_loss: 21.1680\n",
            "Epoch 119/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.6875\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyes no no yen\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yes yes no no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 724ms/step - loss: 1.7065 - val_loss: 21.0673\n",
            "Epoch 120/200\n",
            "1/1 [==============================] - 0s 66ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.5417\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yesyes yes yesno yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyes no no ye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 929ms/step - loss: 1.8452 - val_loss: 18.9185\n",
            "Epoch 121/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.6458\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yesyesyesyesyesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yesyesyesyesyesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 869ms/step - loss: 1.7903 - val_loss: 20.1284\n",
            "Epoch 122/200\n",
            "1/1 [==============================] - 0s 81ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7917\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yes yesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 981ms/step - loss: 1.6992 - val_loss: 23.4997\n",
            "Epoch 123/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7500\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyesno no yen\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yesyes yes yesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 786ms/step - loss: 1.8053 - val_loss: 24.7002\n",
            "Epoch 124/200\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8750\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yesyesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 875ms/step - loss: 1.7967 - val_loss: 29.6860\n",
            "Epoch 125/200\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8750\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yesyesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 866ms/step - loss: 2.0938 - val_loss: 32.9771\n",
            "Epoch 126/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8750\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: esyesyes yesno yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 875ms/step - loss: 1.8254 - val_loss: 33.4772\n",
            "Epoch 127/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7083\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yesyes no no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 742ms/step - loss: 2.3324 - val_loss: 23.4185\n",
            "Epoch 128/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4583\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yesyes yes yes no yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 858ms/step - loss: 1.7535 - val_loss: 15.8728\n",
            "Epoch 129/200\n",
            "1/1 [==============================] - 0s 76ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4375\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yesyes yes yes no yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 873ms/step - loss: 1.8281 - val_loss: 11.8264\n",
            "Epoch 130/200\n",
            "1/1 [==============================] - 0s 82ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.5208\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyes no no ye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 1.7891 - val_loss: 18.2115\n",
            "Epoch 131/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7083\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yes yesno no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyesno no ye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 866ms/step - loss: 1.6895 - val_loss: 23.2124\n",
            "Epoch 132/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7708\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yesyes yes yesno yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 877ms/step - loss: 1.5542 - val_loss: 23.8865\n",
            "Epoch 133/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7083\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyesno no yen\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 728ms/step - loss: 1.5610 - val_loss: 24.2938\n",
            "Epoch 134/200\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.5833\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yes yes yes yes yes yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yes yes yes yes yes yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 870ms/step - loss: 1.5056 - val_loss: 19.9511\n",
            "Epoch 135/200\n",
            "1/1 [==============================] - 0s 74ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yesyesyes yesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 728ms/step - loss: 1.4277 - val_loss: 13.9049\n",
            "Epoch 136/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4167\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 874ms/step - loss: 1.3726 - val_loss: 12.5963\n",
            "Epoch 137/200\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4167\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yesyes yes yes no yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 1.4157 - val_loss: 12.9085\n",
            "Epoch 138/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4375\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyes no no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 946ms/step - loss: 1.4176 - val_loss: 13.4406\n",
            "Epoch 139/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4583\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yesyesyesyesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: es yes yes yesno no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 724ms/step - loss: 1.4159 - val_loss: 17.4657\n",
            "Epoch 140/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.5208\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yesyes yes yes no yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 737ms/step - loss: 1.4305 - val_loss: 19.3620\n",
            "Epoch 141/200\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4583\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: esyesyesyesyes no no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yesyesyes yesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 730ms/step - loss: 1.4325 - val_loss: 16.0638\n",
            "Epoch 142/200\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4375\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: esyesyesyesyes no no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 874ms/step - loss: 1.3951 - val_loss: 13.6185\n",
            "Epoch 143/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: es yes yes yesno no yesn\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 867ms/step - loss: 1.3674 - val_loss: 15.0745\n",
            "Epoch 144/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4583\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: es yes yes yesno no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yesyes yes yes no yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 867ms/step - loss: 1.3724 - val_loss: 18.8238\n",
            "Epoch 145/200\n",
            "1/1 [==============================] - 0s 76ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.6042\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yesyesyesyesyesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 931ms/step - loss: 1.3297 - val_loss: 22.1614\n",
            "Epoch 146/200\n",
            "1/1 [==============================] - 0s 78ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.6458\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yes yes yes yesyes yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: es yes yes yesno noyesn\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 1.2756 - val_loss: 20.8377\n",
            "Epoch 147/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4583\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: es yes yes yes no no yesn\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yesyesyes yesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 726ms/step - loss: 1.3017 - val_loss: 16.1885\n",
            "Epoch 148/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.6875\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yes yesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 882ms/step - loss: 1.3151 - val_loss: 23.6632\n",
            "Epoch 149/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7708\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yesyesnonoyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: es yesyesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 879ms/step - loss: 1.2588 - val_loss: 30.1208\n",
            "Epoch 150/200\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.6458\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yesyes yes yesyesyes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 1.2596 - val_loss: 22.3811\n",
            "Epoch 151/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4375\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyes no no yen\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 955ms/step - loss: 1.2185 - val_loss: 15.1495\n",
            "Epoch 152/200\n",
            "1/1 [==============================] - 0s 81ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.3750\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 1.1831 - val_loss: 13.7388\n",
            "Epoch 153/200\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 1.1752 - val_loss: 17.9248\n",
            "Epoch 154/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7292\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: es yesyesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 868ms/step - loss: 1.2295 - val_loss: 26.1474\n",
            "Epoch 155/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8125\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: esyes yesyesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: es yesyesyesyesyesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 869ms/step - loss: 1.1787 - val_loss: 35.4152\n",
            "Epoch 156/200\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.8958\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: eyesyesyesyesno noyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yesyes yesyesyesyesye\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 726ms/step - loss: 1.1374 - val_loss: 40.8815\n",
            "Epoch 157/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.7708\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: esyes yes yes no yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yesyes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 870ms/step - loss: 1.1870 - val_loss: 29.8804\n",
            "Epoch 158/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4792\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: es yes yes yesno noyesn\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 872ms/step - loss: 1.1428 - val_loss: 16.6342\n",
            "Epoch 159/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.3542\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 865ms/step - loss: 1.1403 - val_loss: 11.5896\n",
            "Epoch 160/200\n",
            "1/1 [==============================] - 0s 84ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.3750\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 970ms/step - loss: 1.1483 - val_loss: 11.3536\n",
            "Epoch 161/200\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4167\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yesyes yes yes no yesnon\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 928ms/step - loss: 1.1321 - val_loss: 13.8546\n",
            "Epoch 162/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4167\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yesno no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: es yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 865ms/step - loss: 1.1853 - val_loss: 14.6645\n",
            "Epoch 163/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.4167\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: es yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: esyesyesyesyes no no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 936ms/step - loss: 1.2289 - val_loss: 12.8536\n",
            "Epoch 164/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.2292\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yes yes yes yes no yes no n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 876ms/step - loss: 1.1714 - val_loss: 7.8979\n",
            "Epoch 165/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1250\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 858ms/step - loss: 1.0773 - val_loss: 5.3538\n",
            "Epoch 166/200\n",
            "1/1 [==============================] - 0s 67ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1042\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: yes yes yes yes no no yesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: yes yes yes yes no no yesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 712ms/step - loss: 1.1715 - val_loss: 4.2681\n",
            "Epoch 167/200\n",
            "1/1 [==============================] - 0s 82ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1042\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: es yes yes yes yes no no yn\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 878ms/step - loss: 1.1033 - val_loss: 4.2497\n",
            "Epoch 168/200\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1042\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: es yes yes yes yes no no n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 1.1601 - val_loss: 4.1096\n",
            "Epoch 169/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1458\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: yes yes yes yes no no yesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: yes yes yes yes no no yesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 730ms/step - loss: 1.0787 - val_loss: 4.6171\n",
            "Epoch 170/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1042\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 878ms/step - loss: 1.1329 - val_loss: 4.1856\n",
            "Epoch 171/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1042\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 870ms/step - loss: 1.0870 - val_loss: 3.9529\n",
            "Epoch 172/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0833\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 893ms/step - loss: 1.0476 - val_loss: 3.6631\n",
            "Epoch 173/200\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1042\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 734ms/step - loss: 1.0729 - val_loss: 3.8514\n",
            "Epoch 174/200\n",
            "1/1 [==============================] - 0s 74ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1042\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: es yes yes yes no no yesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 744ms/step - loss: 1.0712 - val_loss: 4.0790\n",
            "Epoch 175/200\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1250\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 982ms/step - loss: 1.0623 - val_loss: 4.1853\n",
            "Epoch 176/200\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1667\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: es yes yes yes no no yesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yes yes yes yes no yes no n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 956ms/step - loss: 1.0977 - val_loss: 5.6953\n",
            "Epoch 177/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1250\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yes yes yes yes no yes no n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: es yes yes yes no no yesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 725ms/step - loss: 1.1747 - val_loss: 5.8280\n",
            "Epoch 178/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1042\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yes yes yes yes no yes no n\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 734ms/step - loss: 1.0427 - val_loss: 4.5344\n",
            "Epoch 179/200\n",
            "1/1 [==============================] - 0s 76ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.1042\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: es yes yes yes yes no no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: es yes yes yes yes no no yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 879ms/step - loss: 0.9838 - val_loss: 4.1291\n",
            "Epoch 180/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0625\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: yes yes yes yes no no yesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: yes yes yes yes no no yesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 880ms/step - loss: 0.9841 - val_loss: 3.8380\n",
            "Epoch 181/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0625\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yes yes yes yes no yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 794ms/step - loss: 1.0445 - val_loss: 3.2274\n",
            "Epoch 182/200\n",
            "1/1 [==============================] - 0s 70ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0625\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 797ms/step - loss: 0.9741 - val_loss: 2.9388\n",
            "Epoch 183/200\n",
            "1/1 [==============================] - 0s 79ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0417\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 761ms/step - loss: 0.9929 - val_loss: 3.0111\n",
            "Epoch 184/200\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0417\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: yes yes yes yes no no yesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yes yes yes yes no yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 1.0417 - val_loss: 3.1504\n",
            "Epoch 185/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0417\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yes yes yes yes no yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 886ms/step - loss: 1.0880 - val_loss: 2.9475\n",
            "Epoch 186/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0417\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yes yes yes yes no yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 876ms/step - loss: 1.0694 - val_loss: 2.7470\n",
            "Epoch 187/200\n",
            "1/1 [==============================] - 0s 72ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: yes yes yes yes yes no no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 735ms/step - loss: 1.0205 - val_loss: 2.6253\n",
            "Epoch 188/200\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0208\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: yes yes yes yes no no yes no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 723ms/step - loss: 0.9886 - val_loss: 2.6864\n",
            "Epoch 189/200\n",
            "1/1 [==============================] - 0s 68ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: yes yes yes yes yes no no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 881ms/step - loss: 0.9939 - val_loss: 2.2641\n",
            "Epoch 190/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0417\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: yes yes yes yes no no yesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 874ms/step - loss: 1.0252 - val_loss: 2.3229\n",
            "Epoch 191/200\n",
            "1/1 [==============================] - 0s 82ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0417\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: yes yes yes yes no no yesno\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 996ms/step - loss: 0.9696 - val_loss: 2.4037\n",
            "Epoch 192/200\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0417\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yes yes yes yes no yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 895ms/step - loss: 1.0141 - val_loss: 2.5212\n",
            "Epoch 193/200\n",
            "1/1 [==============================] - 0s 80ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0417\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yes yes yes yes no yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 1s/step - loss: 0.9695 - val_loss: 2.6592\n",
            "Epoch 194/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0208\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no yes no no\n",
            "Prediction: yes yes yes yes no yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 873ms/step - loss: 0.9821 - val_loss: 2.9898\n",
            "Epoch 195/200\n",
            "1/1 [==============================] - 0s 77ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0208\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 750ms/step - loss: 0.9678 - val_loss: 2.6875\n",
            "Epoch 196/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 727ms/step - loss: 0.9613 - val_loss: 2.4534\n",
            "Epoch 197/200\n",
            "1/1 [==============================] - 0s 71ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: yes yes yes yes no no yes no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 872ms/step - loss: 0.9611 - val_loss: 2.4519\n",
            "Epoch 198/200\n",
            "1/1 [==============================] - 0s 73ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: yes yes yes yes yes no no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 865ms/step - loss: 0.9940 - val_loss: 2.2739\n",
            "Epoch 199/200\n",
            "1/1 [==============================] - 0s 69ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: yes yes yes yes yes no no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 1s 875ms/step - loss: 0.9569 - val_loss: 2.6029\n",
            "Epoch 200/200\n",
            "1/1 [==============================] - 0s 76ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes no no yes no\n",
            "Prediction: yes yes yes yes no no yes no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "2/2 [==============================] - 2s 941ms/step - loss: 0.9044 - val_loss: 2.4798\n"
          ]
        }
      ],
      "source": [
        "# Define the number of epochs.\n",
        "epochs = 200\n",
        "# Callback function to check transcription on the val set.\n",
        "validation_callback = CallbackEval(validation_dataset)\n",
        "# Train the model\n",
        "history = model.fit(\n",
        "    train_dataset,\n",
        "    validation_data=validation_dataset,\n",
        "    epochs=epochs,\n",
        "    callbacks=[validation_callback],\n",
        ")\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rXbUKPvJRkYW"
      },
      "source": [
        "## Inference"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Q77qMiFFRkYX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a4b88303-9b57-4188-c79e-6decf308214a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 78ms/step\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Word Error Rate: 0.0000\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: yes yes yes yes yes no no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes no no\n",
            "Prediction: yes yes yes yes yes yes no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes no no no\n",
            "Prediction: yes yes yes yes yes no no no\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes no yes no yes yes\n",
            "Prediction: yes yes yes no yes no yes yes\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Target    : yes yes yes yes yes yes yes yes\n",
            "Prediction: yes yes yes yes yes yes yes yes\n",
            "----------------------------------------------------------------------------------------------------\n"
          ]
        }
      ],
      "source": [
        "# Let's check results on more validation samples\n",
        "predictions = []\n",
        "targets = []\n",
        "for batch in validation_dataset:\n",
        "    X, y = batch\n",
        "    batch_predictions = model.predict(X)\n",
        "    batch_predictions = decode_batch_predictions(batch_predictions)\n",
        "    predictions.extend(batch_predictions)\n",
        "    for label in y:\n",
        "        label = tf.strings.reduce_join(num_to_char(label)).numpy().decode(\"utf-8\")\n",
        "        targets.append(label)\n",
        "wer_score = wer(targets, predictions)\n",
        "print(\"-\" * 100)\n",
        "print(f\"Word Error Rate: {wer_score:.4f}\")\n",
        "print(\"-\" * 100)\n",
        "for i in np.random.randint(0, len(predictions), 5):\n",
        "    print(f\"Target    : {targets[i]}\")\n",
        "    print(f\"Prediction: {predictions[i]}\")\n",
        "    print(\"-\" * 100)\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "#############################################################\n",
        "################### Give Audio Path And Run #################\n",
        "#############################################################\n",
        "\n",
        "aud = \"/content/working/waves_yesno/0_0_1_1_1_1_0_0.wav\"\n",
        "\n",
        "#############################################################\n",
        "\n",
        "\n",
        "def encode_aud(wav_file):\n",
        "    \"\"\"\n",
        "    audio file encoder\n",
        "    params : wav audio file path\n",
        "    return : spectogram\n",
        "     \"\"\"\n",
        "    file = tf.io.read_file(wav_file)\n",
        "    audio, _ = tf.audio.decode_wav(file)\n",
        "    audio = tf.squeeze(audio, axis=-1)\n",
        "    audio = tf.cast(audio, tf.float32)\n",
        "    spectrogram = tf.signal.stft(\n",
        "        audio, frame_length=frame_length, frame_step=frame_step, fft_length=fft_length\n",
        "    )\n",
        "    spectrogram = tf.abs(spectrogram)\n",
        "    spectrogram = tf.math.pow(spectrogram, 0.5)\n",
        "    means = tf.math.reduce_mean(spectrogram, 1, keepdims=True)\n",
        "    stddevs = tf.math.reduce_std(spectrogram, 1, keepdims=True)\n",
        "    spectrogram = (spectrogram - means) / (stddevs + 1e-10)\n",
        "\n",
        "    return spectrogram\n",
        "\n",
        "\n",
        "aud = encode_aud(aud)\n",
        "aud = tf.expand_dims(aud, axis=0)\n",
        "#Make preds by ai\n",
        "pred = model.predict(aud)\n",
        "#decode and detokenize it\n",
        "pred = decode_batch_predictions(pred)\n",
        "print(pred)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0Inb51ZrRZpA",
        "outputId": "58378512-fbb3-44ab-8025-4d6a9339ae44"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 3s 3s/step\n",
            "['no no yes yes yes yes no no']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# @title Save models\n",
        "!mkdir model\n",
        "import json\n",
        "\n",
        "# Save the model weights and architecture\n",
        "model.save(\"model/zoomasr\")\n",
        "\n",
        "\n",
        "#with open(\"model/transformer_model_architecture.json\", \"w\") as json_file:\n",
        " #   json_file.write(model.to_json())\n",
        "\n",
        "# Save the tokenizer vocabulary as a JSON file\n",
        "with open(\"model/tokenizer_vocab.txt\", \"w\", encoding=\"utf-8\") as file:\n",
        "    file.write(vocabresult)"
      ],
      "metadata": {
        "id": "RvVrphYysRSE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# @title load model and tokenizer\n",
        "import json\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "\n",
        "# Load the tokenizer vocabulary from the JSON file\n",
        "with open(\"model/tokenizer_vocab.txt\", \"r\") as file:\n",
        "    idx_to_char = file.read()\n",
        "\n",
        "characters = [x for x in idx_to_char]\n",
        "# Mapping characters to integers\n",
        "char_to_num = keras.layers.StringLookup(vocabulary=characters, oov_token=\"\")\n",
        "# Mapping integers back to original characters\n",
        "num_to_char = keras.layers.StringLookup(\n",
        "    vocabulary=char_to_num.get_vocabulary(), oov_token=\"\", invert=True\n",
        ")\n",
        "\n",
        "print(\n",
        "    f\"The vocabulary is: {char_to_num.get_vocabulary()} \"\n",
        "    f\"(size ={char_to_num.vocabulary_size()})\"\n",
        ")\n",
        "\n",
        "# A utility function to decode the output of the network\n",
        "def decode_batch_predictions(pred):\n",
        "    input_len = np.ones(pred.shape[0]) * pred.shape[1]\n",
        "    # Use greedy search. For complex tasks, you can use beam search\n",
        "    results = keras.backend.ctc_decode(pred, input_length=input_len, greedy=True)[0][0]\n",
        "    # Iterate over the results and get back the text\n",
        "    output_text = []\n",
        "    for result in results:\n",
        "        result = tf.strings.reduce_join(num_to_char(result)).numpy().decode(\"utf-8\")\n",
        "        output_text.append(result)\n",
        "    return output_text\n",
        "\n",
        "\n",
        "def CTCLoss(y_true, y_pred):\n",
        "    # Compute the training-time loss value\n",
        "    batch_len = tf.cast(tf.shape(y_true)[0], dtype=\"int64\")\n",
        "    input_length = tf.cast(tf.shape(y_pred)[1], dtype=\"int64\")\n",
        "    label_length = tf.cast(tf.shape(y_true)[1], dtype=\"int64\")\n",
        "\n",
        "    input_length = input_length * tf.ones(shape=(batch_len, 1), dtype=\"int64\")\n",
        "    label_length = label_length * tf.ones(shape=(batch_len, 1), dtype=\"int64\")\n",
        "\n",
        "    loss = keras.backend.ctc_batch_cost(y_true, y_pred, input_length, label_length)\n",
        "    return loss\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "with keras.utils.custom_object_scope({'CTCLoss': CTCLoss}):\n",
        "    loaded_model = tf.keras.models.load_model(\"model/zoomasr\")\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ooFyqRYCuL7T",
        "outputId": "6670dcae-4025-47ab-8719-b6b98b4a1a44"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The vocabulary is: ['', '<', 'U', 'N', 'K', '>', ' ', 'y', 'e', 's', 'n', 'o'] (size =12)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "#############################################################\n",
        "################### Give Audio Path And Run #################\n",
        "#############################################################\n",
        "\n",
        "aud = \"/content/working/waves_yesno/0_0_1_1_1_1_0_0.wav\"\n",
        "\n",
        "#############################################################\n",
        "import numpy as np\n",
        "# An integer scalar Tensor. The window length in samples.\n",
        "frame_length = 256\n",
        "# An integer scalar Tensor. The number of samples to step.\n",
        "frame_step = 160\n",
        "# An integer scalar Tensor. The size of the FFT to apply.\n",
        "# If not provided, uses the smallest power of 2 enclosing frame_length.\n",
        "fft_length = 384\n",
        "\n",
        "\n",
        "def encode_aud(wav_file):\n",
        "    \"\"\"\n",
        "    audio file encoder\n",
        "    params : wav audio file path\n",
        "    return : spectogram\n",
        "     \"\"\"\n",
        "    file = tf.io.read_file(wav_file)\n",
        "    audio, _ = tf.audio.decode_wav(file)\n",
        "    audio = tf.squeeze(audio, axis=-1)\n",
        "    audio = tf.cast(audio, tf.float32)\n",
        "    spectrogram = tf.signal.stft(\n",
        "        audio, frame_length=frame_length, frame_step=frame_step, fft_length=fft_length\n",
        "    )\n",
        "    spectrogram = tf.abs(spectrogram)\n",
        "    spectrogram = tf.math.pow(spectrogram, 0.5)\n",
        "    means = tf.math.reduce_mean(spectrogram, 1, keepdims=True)\n",
        "    stddevs = tf.math.reduce_std(spectrogram, 1, keepdims=True)\n",
        "    spectrogram = (spectrogram - means) / (stddevs + 1e-10)\n",
        "\n",
        "    return spectrogram\n",
        "\n",
        "\n",
        "aud = encode_aud(aud)\n",
        "aud = tf.expand_dims(aud, axis=0)\n",
        "#Make preds by ai\n",
        "pred = loaded_model.predict(aud)\n",
        "#decode and detokenize it\n",
        "pred = decode_batch_predictions(pred)\n",
        "print(pred)\n"
      ],
      "metadata": {
        "id": "P0TYnqxVyYTs",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8370d96e-0c48-4629-f606-d6510fc8e811"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 3s 3s/step\n",
            "['no no yes yes yes yes no no']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# @title io ops\n",
        "!zip -r yesno_zoomasr_model.zip /content/model"
      ],
      "metadata": {
        "id": "-IspmV3gxLBL",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "edb2ebaa-fbbb-4688-a61a-476a965d2e3e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "  adding: content/model/ (stored 0%)\n",
            "  adding: content/model/tokenizer_vocab.txt (stored 0%)\n",
            "  adding: content/model/zoomasr/ (stored 0%)\n",
            "  adding: content/model/zoomasr/fingerprint.pb (stored 0%)\n",
            "  adding: content/model/zoomasr/saved_model.pb (deflated 91%)\n",
            "  adding: content/model/zoomasr/assets/ (stored 0%)\n",
            "  adding: content/model/zoomasr/keras_metadata.pb (deflated 95%)\n",
            "  adding: content/model/zoomasr/variables/ (stored 0%)\n",
            "  adding: content/model/zoomasr/variables/variables.data-00000-of-00001 (deflated 7%)\n",
            "  adding: content/model/zoomasr/variables/variables.index (deflated 72%)\n"
          ]
        }
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": []
    },
    "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": "3.7.0"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}