material/deep_learning/template_text_sequence_data_classification.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "template_text_sequence_data_classification.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mZAFnuGfwqy-"
      },
      "source": [
        "# 텍스트 순차열 데이터 분류 Template"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Hb0EOx6OyNIw"
      },
      "source": [
        "# 데이터 준비"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VBfCpbwh0h9R"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "B5AtaEi6yPhX"
      },
      "source": [
        "## 기반 문자열 생성"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V6wKtOwFwqXM"
      },
      "source": [
        "BASE_STRING = \"abcde12345\"\n",
        "text_data = BASE_STRING*10\n",
        "CATEGORY_NUM = 2"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6RAkHiJbxNfE"
      },
      "source": [
        "SEQUENCE_LENGTH = 5\n",
        "OFFSET = 1"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f-z-YNfNDBn5"
      },
      "source": [
        "## 입출력 데이터 생성\n",
        "\n",
        "'abcde12345abcde12345'의 문자열로 다음의 입출력 데이터를 생성한다.\n",
        "```\n",
        "'abcde' -> T\n",
        "'bcde1' -> F\n",
        "'cde12' -> T\n",
        "'de123' -> F\n",
        "...\n",
        "'e1234' -> T\n",
        "```\n",
        "\n",
        "다음 문자열이 알파벳이면 1, 숫자이면 0을 출력으로 한다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yYxpmUQtwqTY"
      },
      "source": [
        "raw_x = []\n",
        "raw_y = []\n",
        "\n",
        "for i in range(len(text_data)):\n",
        "  if i+SEQUENCE_LENGTH+OFFSET >= len(text_data): break\n",
        "  raw_x.append(text_data[i:i+SEQUENCE_LENGTH])\n",
        "  next_ch = text_data[i+SEQUENCE_LENGTH+OFFSET-1]\n",
        "  raw_y.append([ int(next_ch in 'abcde') ])\n",
        "#            0         1         2      \n",
        "#            012345678901234567890123456789\n",
        "# textdata = abcde12345abcde12345abcde12345\n",
        "#               i = 3\n",
        "#               <-x->   = [3:8] = 'de123'\n",
        "#                    y  = [8]   = '4' is alpahbe = False = 0\n"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "15RBiNi9wqQU",
        "outputId": "e7503ed2-0bdc-4504-8fc0-d0acdcee4c02"
      },
      "source": [
        "print(len(raw_x))\n",
        "# len(raw_x) = 94 = len(text_data) - SEQUENCE_LENGTH - OFFSET = 100 - 5 - 1"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "94\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S7eFX6jIyrnx"
      },
      "source": [
        "## 데이터를 2차원 벡터화 하기\n",
        "\n",
        "RNN은 2차원 데이터를 입력으로 받는다. 1차원 문자열의 각 문자를 one-hot 인코딩하여 2차원 데이터로 만드렁 준다.<br>\n",
        "입력 데이터와 출력 데이터를 다음의 형태로 준비한다.<br>\n",
        "입력 데이터\n",
        "```\n",
        "문자열 'abc12'을 인코딩한\n",
        "[[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
        " [0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
        " [0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
        " [1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
        " [0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]]\n",
        "```\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4jO8skDb7p3f"
      },
      "source": [
        "인코딩 디코딩을 위한 유틸 클래스 CharacterTable"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-atmjECTwqNi"
      },
      "source": [
        "class CharacterTable:\n",
        "\n",
        "    # '1', '2', '3', '4', '5', 'a', 'b', 'c', 'd', 'e'\n",
        "    #  0    1    2    3    4    5    6    7    8    9 \n",
        "\n",
        "    def __init__(self, chars):\n",
        "        self.chars = sorted(set(chars))\n",
        "        self.char_indices = dict((c, i) for i, c in enumerate(self.chars)) # '+' ->  1\n",
        "        self.indices_char = dict((i, c) for i, c in enumerate(self.chars)) #  1  -> '+'\n",
        "\n",
        "    def encode(self, C, num_rows=SEQUENCE_LENGTH):\n",
        "        x = np.zeros((num_rows, len(self.chars)))\n",
        "        for i, c in enumerate(C): # 각 자리 문자마다\n",
        "            x[i, self.char_indices[c]] = 1 # 각 줄의 index만 1로 만든다. one-hot encoding한다.\n",
        "        return x\n",
        "\n",
        "    def decode(self, x):\n",
        "        x = x.argmax(axis=-1)\n",
        "        return \"\".join(self.indices_char[x] for x in x)\n"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MW8bFIwAwqKa"
      },
      "source": [
        "ctable = CharacterTable(BASE_STRING)"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "I_fJZA7b7Y-e"
      },
      "source": [
        "다음은 인코딩, 디코딩 예이다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vwm40AA-r4tt",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9f8987a3-f8a6-40d3-d9ef-01a51d434246"
      },
      "source": [
        "# 원 문자열\n",
        "org_str = 'abc12'\n",
        "print(f\"org_str='{org_str}'\")\n",
        "\n",
        "# 인코딩된 값\n",
        "encoded = ctable.encode(org_str)\n",
        "# [[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]    <==== 'a'\n",
        "#  [0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]    <==== 'b'\n",
        "#  [0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]    <==== 'c'\n",
        "#  [1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]    <==== '1'\n",
        "#  [0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]]   <==== '2'\n",
        "print(f\"encoded=\\n{encoded}\")\n",
        "print(f\"encoded.shape={encoded.shape}\")\n",
        "\n",
        "# 다시 디코딩해서 원복한 값\n",
        "decoded = ctable.decode(encoded)\n",
        "print(f\"decoded='{decoded}'\")"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "org_str='abc12'\n",
            "encoded=\n",
            "[[0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
            " [0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
            " [0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
            " [1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            " [0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]]\n",
            "encoded.shape=(5, 10)\n",
            "decoded='abc12'\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4mg54AkZ3boa",
        "outputId": "23366423-3fc9-4066-cc40-5a5f4d87f89d"
      },
      "source": [
        "print(\"Vectorization...\")\n",
        "\n",
        "x = np.zeros((len(raw_x), len(raw_x[0]), len(BASE_STRING)), dtype=np.bool)\n",
        "for i, sentence in enumerate(raw_x):\n",
        "    x[i] = ctable.encode(sentence)\n",
        "y = np.array(raw_y).astype(np.int16)\n",
        "\n",
        "print(x.shape)\n",
        "print(y.shape)\n",
        "\n",
        "split_at = int(len(x)*0.9)\n",
        "(train_x, test_x) = x[:split_at], x[split_at:]\n",
        "(train_y, test_y) = y[:split_at], y[split_at:]\n",
        "\n",
        "print(\"Training Data:\")\n",
        "print(train_x.shape)\n",
        "print(train_y.shape)\n",
        "\n",
        "print(\"Test Data:\")\n",
        "print(test_x.shape)\n",
        "print(test_y.shape)\n"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Vectorization...\n",
            "(94, 5, 10)\n",
            "(94, 1)\n",
            "Training Data:\n",
            "(84, 5, 10)\n",
            "(84, 1)\n",
            "Test Data:\n",
            "(10, 5, 10)\n",
            "(10, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lMq2Om3w8qNy"
      },
      "source": [
        "준비된 데이터는 다음과 같습니다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "41UcKyn-q5Ez",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "98ed6a38-03d6-4016-b18f-bb76ea02789f"
      },
      "source": [
        "print(\"train_x[0]\\n\", train_x[0].astype(np.int16))\n",
        "print()\n",
        "print(\"train_y[0]\\n\", train_y[0].astype(np.int16))"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "train_x[0]\n",
            " [[0 0 0 0 0 1 0 0 0 0]\n",
            " [0 0 0 0 0 0 1 0 0 0]\n",
            " [0 0 0 0 0 0 0 1 0 0]\n",
            " [0 0 0 0 0 0 0 0 1 0]\n",
            " [0 0 0 0 0 0 0 0 0 1]]\n",
            "\n",
            "train_y[0]\n",
            " [0]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wa-gM3Nq84XK"
      },
      "source": [
        "준비된 데이터는 인코딩된 값들입니다. 디코딩에서 본 원 문자열은 다음과 같습니다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zw_RFfJvrGFM",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ccc018cd-1168-4957-b649-a3f5f99ad984"
      },
      "source": [
        "for i in range(10):\n",
        "  print(\"'\"+ctable.decode(train_x[i])+\"' -> \"+str(train_y[i][0]))"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "'abcde' -> 0\n",
            "'bcde1' -> 0\n",
            "'cde12' -> 0\n",
            "'de123' -> 0\n",
            "'e1234' -> 0\n",
            "'12345' -> 1\n",
            "'2345a' -> 1\n",
            "'345ab' -> 1\n",
            "'45abc' -> 1\n",
            "'5abcd' -> 1\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0-mHtIAk_qU9"
      },
      "source": [
        "# Template"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "k3CG02lDAI73"
      },
      "source": [
        "## LSTM"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 819
        },
        "id": "GEffW4RR3bgJ",
        "outputId": "30b1d04a-e8e3-4694-c173-1fc0c6819e05"
      },
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras.layers import Dense, RepeatVector, Flatten\n",
        "from tensorflow.keras.layers import Bidirectional, LSTM, GRU\n",
        "\n",
        "input_shape = train_x.shape[1:]\n",
        "# input_shape = (5,10)\n",
        "\n",
        "model = keras.Sequential()\n",
        "model.add(LSTM(128, input_shape=input_shape))\n",
        "model.add(Flatten())\n",
        "model.add(Dense(CATEGORY_NUM, activation=\"softmax\"))\n",
        "model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
        "model.summary()\n",
        "\n",
        "history = model.fit(train_x, train_y, epochs=100, verbose=1, validation_split=0.1, batch_size=128)\n",
        "\n",
        "\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.show()\n",
        "\n",
        "\n",
        "\n",
        "loss, acc = model.evaluate(test_x, test_y)\n",
        "print(\"loss=\", loss)\n",
        "print(\"acc=\", acc)\n",
        "\n",
        "\n",
        "\n",
        "y_ = model.predict(test_x)\n",
        "predicted = np.argmax(y_, axis=-1)\n",
        "\n",
        "plt.plot(test_y, \"o\")\n",
        "plt.plot(predicted, \".\")\n",
        "plt.show()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "lstm (LSTM)                  (None, 128)               71168     \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 2)                 258       \n",
            "=================================================================\n",
            "Total params: 71,426\n",
            "Trainable params: 71,426\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "1/1 [==============================] - 0s 16ms/step - loss: 8.1507e-05 - accuracy: 1.0000\n",
            "loss= 8.150665235007182e-05\n",
            "acc= 1.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M1sV7tBtAWEz"
      },
      "source": [
        "## GRU"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 819
        },
        "id": "usltj8hI3bdL",
        "outputId": "d19d47fc-a4e4-42a0-ca1f-4e8a68040894"
      },
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras.layers import Dense, RepeatVector, Flatten\n",
        "from tensorflow.keras.layers import Bidirectional, LSTM, GRU\n",
        "\n",
        "input_shape = train_x.shape[1:]\n",
        "# input_shape = (5,10)\n",
        "\n",
        "model = keras.Sequential()\n",
        "model.add(GRU(128, input_shape=input_shape))\n",
        "model.add(Flatten())\n",
        "model.add(Dense(CATEGORY_NUM, activation=\"softmax\"))\n",
        "model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
        "model.summary()\n",
        "\n",
        "history = model.fit(train_x, train_y, epochs=100, verbose=1, validation_split=0.1, batch_size=128)\n",
        "\n",
        "\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.show()\n",
        "\n",
        "\n",
        "\n",
        "loss, acc = model.evaluate(test_x, test_y)\n",
        "print(\"loss=\", loss)\n",
        "print(\"acc=\", acc)\n",
        "\n",
        "\n",
        "\n",
        "y_ = model.predict(test_x)\n",
        "predicted = np.argmax(y_, axis=-1)\n",
        "\n",
        "plt.plot(test_y, \"o\")\n",
        "plt.plot(predicted, \".\")\n",
        "plt.show()"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "gru (GRU)                    (None, 128)               53760     \n",
            "_________________________________________________________________\n",
            "flatten_1 (Flatten)          (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 2)                 258       \n",
            "=================================================================\n",
            "Total params: 54,018\n",
            "Trainable params: 54,018\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "1/1 [==============================] - 0s 14ms/step - loss: 5.2615e-05 - accuracy: 1.0000\n",
            "loss= 5.26148833159823e-05\n",
            "acc= 1.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jCBiKCg9AeHT"
      },
      "source": [
        "## Bidirectional + LSTM"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "azGH_FDbwqHT",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 819
        },
        "outputId": "f3237175-e8dd-4c6e-f260-2ea1696ed48e"
      },
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras.layers import Dense, RepeatVector, Flatten\n",
        "from tensorflow.keras.layers import Bidirectional, LSTM, GRU\n",
        "\n",
        "input_shape = train_x.shape[1:]\n",
        "# input_shape = (5,10)\n",
        "\n",
        "model = keras.Sequential()\n",
        "model.add(Bidirectional(LSTM(128), input_shape=input_shape))\n",
        "model.add(Flatten())\n",
        "model.add(Dense(CATEGORY_NUM, activation=\"softmax\"))\n",
        "model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
        "model.summary()\n",
        "\n",
        "history = model.fit(train_x, train_y, epochs=100, verbose=1, validation_split=0.1, batch_size=128)\n",
        "\n",
        "\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.show()\n",
        "\n",
        "\n",
        "\n",
        "loss, acc = model.evaluate(test_x, test_y)\n",
        "print(\"loss=\", loss)\n",
        "print(\"acc=\", acc)\n",
        "\n",
        "\n",
        "\n",
        "y_ = model.predict(test_x)\n",
        "predicted = np.argmax(y_, axis=-1)\n",
        "\n",
        "plt.plot(test_y, \"o\")\n",
        "plt.plot(predicted, \".\")\n",
        "plt.show()"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "bidirectional (Bidirectional (None, 256)               142336    \n",
            "_________________________________________________________________\n",
            "flatten_2 (Flatten)          (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 2)                 514       \n",
            "=================================================================\n",
            "Total params: 142,850\n",
            "Trainable params: 142,850\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "1/1 [==============================] - 0s 14ms/step - loss: 4.6560e-05 - accuracy: 1.0000\n",
            "loss= 4.6560388000216335e-05\n",
            "acc= 1.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TfY_p_SJA5Rf"
      },
      "source": [
        "## Bidirectional + GRU"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OcjycCe8wqEH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 819
        },
        "outputId": "2dcbbd85-26c8-4b49-c52d-7db43dc14ca1"
      },
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras.layers import Dense, RepeatVector, Flatten\n",
        "from tensorflow.keras.layers import Bidirectional, LSTM, GRU\n",
        "\n",
        "input_shape = train_x.shape[1:]\n",
        "# input_shape = (5,10)\n",
        "\n",
        "model = keras.Sequential()\n",
        "model.add(Bidirectional(GRU(128), input_shape=input_shape))\n",
        "model.add(Flatten())\n",
        "model.add(Dense(CATEGORY_NUM, activation=\"softmax\"))\n",
        "model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
        "model.summary()\n",
        "\n",
        "history = model.fit(train_x, train_y, epochs=100, verbose=1, validation_split=0.1, batch_size=128)\n",
        "\n",
        "\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.show()\n",
        "\n",
        "\n",
        "\n",
        "loss, acc = model.evaluate(test_x, test_y)\n",
        "print(\"loss=\", loss)\n",
        "print(\"acc=\", acc)\n",
        "\n",
        "\n",
        "\n",
        "y_ = model.predict(test_x)\n",
        "predicted = np.argmax(y_, axis=-1)\n",
        "\n",
        "plt.plot(test_y, \"o\")\n",
        "plt.plot(predicted, \".\")\n",
        "plt.show()"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_3\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "bidirectional_1 (Bidirection (None, 256)               107520    \n",
            "_________________________________________________________________\n",
            "flatten_3 (Flatten)          (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 2)                 514       \n",
            "=================================================================\n",
            "Total params: 108,034\n",
            "Trainable params: 108,034\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "1/1 [==============================] - 0s 15ms/step - loss: 7.0089e-05 - accuracy: 1.0000\n",
            "loss= 7.008880493231118e-05\n",
            "acc= 1.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "P3aO3WEwA-lR"
      },
      "source": [
        "## stacked"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xIYi9GUDwqA6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 893
        },
        "outputId": "cde4289d-0e68-4310-efcc-6aafc95896a1"
      },
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras.layers import Dense, RepeatVector, Flatten\n",
        "from tensorflow.keras.layers import Bidirectional, LSTM, GRU\n",
        "\n",
        "input_shape = train_x.shape[1:]\n",
        "# input_shape = (5,10)\n",
        "\n",
        "model = keras.Sequential()\n",
        "model.add(LSTM(128, input_shape=input_shape, return_sequences=True))\n",
        "model.add(LSTM(128))\n",
        "model.add(Flatten())\n",
        "model.add(Dense(CATEGORY_NUM, activation=\"softmax\"))\n",
        "model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
        "model.summary()\n",
        "\n",
        "history = model.fit(train_x, train_y, epochs=100, verbose=1, validation_split=0.1, batch_size=128)\n",
        "\n",
        "\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.show()\n",
        "\n",
        "\n",
        "\n",
        "loss, acc = model.evaluate(test_x, test_y)\n",
        "print(\"loss=\", loss)\n",
        "print(\"acc=\", acc)\n",
        "\n",
        "\n",
        "\n",
        "y_ = model.predict(test_x)\n",
        "predicted = np.argmax(y_, axis=-1)\n",
        "\n",
        "plt.plot(test_y, \"o\")\n",
        "plt.plot(predicted, \".\")\n",
        "plt.show()"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_4\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "lstm_2 (LSTM)                (None, 5, 128)            71168     \n",
            "_________________________________________________________________\n",
            "lstm_3 (LSTM)                (None, 128)               131584    \n",
            "_________________________________________________________________\n",
            "flatten_4 (Flatten)          (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 2)                 258       \n",
            "=================================================================\n",
            "Total params: 203,010\n",
            "Trainable params: 203,010\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD4CAYAAAD8Zh1EAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de3xU9Z3/8dcnCUlIgkAugEAkCQQwKIpGQEGwor/iZaVaa8HatauWdlu32vbXXd325691t7u9WLf9/WrdUrfdtlulLrUWFSuKqNwlKIrhmoRAglzC/Z7rZ/+YwUYMMMBMTmbm/Xw88pg553yZ+RwOvDl85sz3mLsjIiLxLyXoAkREJDoU6CIiCUKBLiKSIBToIiIJQoEuIpIg0oJ64/z8fC8qKgrq7UVE4tKKFSt2untBR9sCC/SioiIqKiqCensRkbhkZptOtE0tFxGRBKFAFxFJEAp0EZEEoUAXEUkQEQW6mU02s3VmVmVmD3Sw/d/MbGX4Z72Z7Y1+qSIicjKnvMrFzFKBx4BrgXpguZnNdvfVx8a4+1fbjf87YFQMahURkZOI5Ax9NFDl7jXu3gTMBKacZPw04KloFCciIpGLJNAHAHXtluvD6z7CzAYBxcCrJ9g+3cwqzKyioaHhdGsFoHbnIb7/57W0tWnaXxGR9qL9oehUYJa7t3a00d1nuHu5u5cXFHT4RadTmrt6G4+/Vs13nqtEc7mLiPxFJN8U3QIUtlseGF7XkanAl8+2qJP5/LhBtO1Yx/eWbCI7I42/nzw8lm8nIhI3IjlDXw6UmlmxmaUTCu3Zxw8ys+FAb2BJdEs87n0WPMIX1t7Fw+fX87PXqvnpqxti+XYiInHjlIHu7i3AvcBLwBrgaXevNLOHzeymdkOnAjM91n2Q8ruxgmF8tvZB/rX4XR6Zu54nFtTE9C1FROJBRJNzufscYM5x6x46bvnb0SvrJHIK4HPPY7//LNNqvkf3AZ/n/hecbqkp3HlFUaeUICLSFQU22+JZyegBtz8Nf/oSn1j1C3L67Obzs0OhfvuY84KuTkQkEPEZ6ABp6XDzDMjK55plj/O7vH389R//mvS0FG69dGDQ1YmIdLr4DXSAlBSY/K+QnccVr/4zs3rt5/ZZf0tmtxRuHNk/6OpERDpVfAc6gBlM+AZk5XHR81/j6R6N3D7zfjLSxnFtWd+gqxMR6TSJM9ti+V3YzT9nRPMqZmb9kL//3UIWVe0MuioRkU6TOIEOcNGnsVt/yflt63kq83t8/Tevs6p+X9BViYh0isQKdIARN2O3/ZZhvpEZqT/gi798g407DwVdlYhIzCVeoAMMvx679T+4kPU82vZ97n5iATv2Hw26KhGRmErMQAcom4JNeYwxrOJbR37IF3+9lKPNHc4ZJiKSEBI30AEuvh2uf4SrrYJbd/yErz+9UtPuikjCSuxABxj9eRj/VW5PfZW+q3/Fj+dpMi8RSUyJH+gAVz+ED7+Rb3X7He/Of5rn330/6IpERKIuOQI9JQW7ZQb0u4CfZfyUGbNeoLrhYNBViYhEVXIEOkB6NinTZpLRPYdHU37C1/5rMUea9CGpiCSO5Al0gJ4DSL3l5wyhjtt2/zsP/em9oCsSEYma5Ap0gCGTYNx9fCZ1HgfefoY/rKgPuiIRkahIvkAH+Ni38P6X8EjGE/x89mu8v/dI0BWJiJy15Az0tHTs1l+SleZ8i1/wD7PeIdZ3zhMRibXkDHSA3GJSJj3EBFtJj5oXePLNzUFXJCJyViIKdDObbGbrzKzKzB44wZjbzGy1mVWa2ZPRLTNGLrsHP/civpv5X/z4hRXU7T4cdEUiImfslIFuZqnAY8B1QBkwzczKjhtTCjwIjHP3EcD9Mag1+lLTsBv/jV5te7jPnuabz76n1ouIxK1IztBHA1XuXuPuTcBMYMpxYz4PPObuewDcfUd0y4yhAZdil93D7fYSuzcs4+XV24OuSETkjEQS6AOAunbL9eF17Q0FhprZIjNbamaTO3ohM5tuZhVmVtHQ0HBmFcfCpP+DZefzL1lP8fBzlZqVUUTiUrQ+FE0DSoGrgGnAL8ys1/GD3H2Gu5e7e3lBQUGU3joKMntiE77ByNZKive/yb+/Xh10RSIipy2SQN8CFLZbHhhe1149MNvdm919I7CeUMDHj0vvhJ6F/HOPP/L4a1X6gFRE4k4kgb4cKDWzYjNLB6YCs48b8yyhs3PMLJ9QC6YminXGXloGXPUAgxrXca2t4Hsvrg26IhGR03LKQHf3FuBe4CVgDfC0u1ea2cNmdlN42EvALjNbDcwHvuHuu2JVdMyMnAp5pXw75xleXLWFyvd1g2kRiR8W1GV65eXlXlFREch7n9R7f4BZd/EAf8fO4ik8cedlQVckIvIBM1vh7uUdbUveb4qeSNnN0GcE38iawytrtvPW5j1BVyQiEhEF+vFSUuDyL5N3uJrrs9by6Nz1QVckIhIRBXpHLrwVsvvwQO/5LKzayZLq+Ps4QESSjwK9I2kZcNk9nLdrIaN7NPCTeTpLF5GuT4F+IuV3QWoG3+6zkKU1u1lVryteRKRrU6CfSE4BjPwU5+94gf4ZR/nFgvi6rF5Eko8C/WTGfglrPsw/D1zOC6u2skV3NhKRLkyBfjJ9R0DxBCYceAGjjf9ctDHoikRETkiBfiqjPkva/s383eAGZr5Zx4GjzUFXJCLSIQX6qQy/ETLO4bOZCznQ2MLvl9ed+teIiARAgX4q6Vkw4mZya1/kykHd+fWSWtradFcjEel6FOiRuPgz0HyIr/WvpG73ERbri0Yi0gUp0CNROBpyB3PRrjn0yurGU8s3B12RiMhHKNAjYQYX307K5sXcVebMrdzG7kNNQVclIvIhCvRIXTQNMG7PWERzq/PMW/VBVyQi8iEK9Ej1HAAlV5Ff/SyjCnsyc3kdQc0lLyLSEQX66bjgFti7ib8ddpCqHQdZsUlzpYtI16FAPx3DbgBL5arWJWSnp/LUm7omXUS6DgX66cjOg+IrSV/3HH818lzmrNrK4aaWoKsSEQEiDHQzm2xm68ysyswe6GD758yswcxWhn/uiX6pXUTZFNhdzdSigxxpbmXemh1BVyQiAkQQ6GaWCjwGXAeUAdPMrKyDob9394vDP09Euc6uY/iNYCmM3P8afXpk8Nw77wddkYgIENkZ+migyt1r3L0JmAlMiW1ZXVhOHxg0jpQ1s7lh5Lm8tq6BfUc0YZeIBC+SQB8AtP/0rz687nifNLN3zWyWmRVGpbquqmwKNKzltvMO0dTaxtzKbUFXJCIStQ9FnwOK3H0k8DLw644Gmdl0M6sws4qGhoYovXUAht8Yetgzn8Lc7sxW20VEuoBIAn0L0P6Me2B43QfcfZe7N4YXnwAu7eiF3H2Gu5e7e3lBQcGZ1Ns1nHMuFI7F1szmr0b2Z3H1LnYebDz1rxMRiaFIAn05UGpmxWaWDkwFZrcfYGbntlu8CVgTvRK7qOHXw/b3uGUItLY5L67aGnRFIpLkThno7t4C3Au8RCion3b3SjN72MxuCg/7iplVmtk7wFeAz8Wq4C6j9OMADN6zmNI+OWq7iEjg0iIZ5O5zgDnHrXuo3fMHgQejW1oXVzAMep2HbZjLjSP/iR/PW8+OA0fp0yMz6MpEJEnpm6Jnyix0lr7xdT4+rCfu6EtGIhIoBfrZGPpxaD7MsKPvMrB3d15evT3oikQkiSnQz0bReEjrjm2Yy7VlfVlYtZNDjZrbRUSCoUA/G926Q/EE2PAS157fh6aWNt5YH8fX14tIXFOgn62h/wv21DK6x256du+mtouIBEaBfrbCly+mVb/MpOF9eHXdDlpa2wIuSkSSkQL9bPUqhD5lobZLWV/2Hm5mea3uZCQinU+BHg2l18KmJUwo6k56WgpzV2uyLhHpfAr0aCj5GLQ1k71tOeOH5PPy6u26gbSIdDoFejScNxZSM6B6PpPO70P9niNUNxwMuioRSTIK9Gjo1h0GXQ41rzFxaGgWydfW6fJFEelcCvRoKbkKdlQyMO0AQ/rk8LquRxeRTqZAj5aSj4UeN77OxKEFLKvZzeEmfWtURDqPAj1a+o2E7rkftF2aWttYVrM76KpEJIko0KMlJQVKJkL1fEYX9SazWwqvrdPsiyLSeRTo0VRyFRx4n8x9NVxekqc+uoh0KgV6NB3ro9fMZ+LQAmp3HaZ256FgaxKRpKFAj6beg6B3caiPPqwPAG9s0Fm6iHQOBXq0lVwFGxdQ3DuDQXlZvK7r0UWkkyjQo61kIjQdgK0rmTi0gMXVuzja3Bp0VSKSBCIKdDObbGbrzKzKzB44ybhPmpmbWXn0Sowzg8aHHje+wZWlBRxpbuWtzZp9UURi75SBbmapwGPAdUAZMM3MyjoY1wO4D1gW7SLjSk4BFJwPtQsZU5JLaoqxuGpX0FWJSBKI5Ax9NFDl7jXu3gTMBKZ0MO6fgO8DR6NYX3wqvhI2L+WcbjByYE8WVe8MuiIRSQKRBPoAoK7dcn143QfM7BKg0N1fONkLmdl0M6sws4qGhgT+sLBoPDQfgi1vMW5wPu/W72P/0eagqxKRBHfWH4qaWQrwKPD1U4119xnuXu7u5QUFBWf71l3XsT567QLGDcmntc01DYCIxFwkgb4FKGy3PDC87pgewAXAa2ZWC4wFZif1B6PZedBnBNQu4JJBvcjslsKiKrVdRCS2Ign05UCpmRWbWTowFZh9bKO773P3fHcvcvciYClwk7tXxKTieFF8JWxeRgYtXFaUy2L10UUkxk4Z6O7eAtwLvASsAZ5290oze9jMbop1gXGr6EpoORLqow/JZ/32g+zYr8+LRSR20iIZ5O5zgDnHrXvoBGOvOvuyEsCgKwAL9dEHfwGAxdW7+MSoASf/dSIiZ0jfFI2VrFzodwFsfIOy/ufQs3s39dFFJKYU6LFUdCXUvUlqayNXDM5jUdVO3D3oqkQkQSnQY6loPLQ2wpYVXDEkn/f3HaV21+GgqxKRBKVAj6XzLgcMNi3iisF5ACyt0TQAIhIbCvRYysqFviOgdiEl+dkU9MhQoItIzCjQY23QOKh7E2ttZmxJHktrdqmPLiIxoUCPtaJxoevR33+bsSW5bN/fyEbdlk5EYkCBHmuDxoUeNy1kbMmxPrrmdRGR6FOgx1p2fnh+9EWU5GfTR310EYkRBXpnKBoHdcuwtlb10UUkZhTonWHQOGg6CFvfYWxJHjsOqI8uItGnQO8MH+qj5wKwRG0XEYkyBXpn6NEX8kqhdhHFH/TR9cGoiESXAr2zFI2DzUswb+Pyweqji0j0KdA7y6Dx0Lgftr3L2JI8Gg40Ut2gPrqIRI8CvbMUhfvotYsYUxzqoy/bqD66iESPAr2znNMfckugduEHfXTdOFpEokmB3pmKxsPmxZi3MUbXo4tIlCnQO1PRlXB0H2x/jzHFuew40Kj50UUkaiIKdDObbGbrzKzKzB7oYPsXzWyVma00s4VmVhb9UhPAsevRa/8yr8syXY8uIlFyykA3s1TgMeA6oAyY1kFgP+nuF7r7xcAPgEejXmki6DkAehdD7SIGF2STn6N5XUQkeiI5Qx8NVLl7jbs3ATOBKe0HuPv+dovZgBrDJ1I0HjYtwtwZU5zLso271UcXkaiIJNAHAHXtluvD6z7EzL5sZtWEztC/Ep3yElDReDi6F3ZUMqYkl637jlK3+0jQVYlIAojah6Lu/pi7Dwb+AfhWR2PMbLqZVZhZRUNDQ7TeOr500EdfquvRRSQKIgn0LUBhu+WB4XUnMhP4REcb3H2Gu5e7e3lBQUHkVSaSXoXQuwhqF1LaJ4fc7HT10UUkKiIJ9OVAqZkVm1k6MBWY3X6AmZW2W7wB2BC9EhPQoL/00UcX5eoLRiISFacMdHdvAe4FXgLWAE+7e6WZPWxmN4WH3WtmlWa2EvgacGfMKk4ERePhyB7YUcnYkly27D1C3W5djy4iZyctkkHuPgeYc9y6h9o9vy/KdSW2ovGhx40LGFNyBwDLNu6mMDcrwKJEJN7pm6JB6FUYmtdl4xsM69uD3lndWFKtPrqInB0FelCKJ8KmRaR4K2OKNa+LiJw9BXpQiieE5kffupLLB+exZe8R6vfoenQROXMK9KAUTwg9bnydyweHrkdX20VEzoYCPSjZ+dD3Aqh5ndI+OeRlp+vG0SJyVhToQSqeAHXLsJZGxpbksaRafXQROXMK9CAVT4CWo1C/nLGD89i2/yibND+6iJwhBXqQBo0DSw310cPzuqjtIiJnSoEepMxzoP8o2PgGgwuyKeiRoQ9GReSMKdCDVjwBtqzAmg6G+ui6Hl1EzpACPWglE6GtBTYt5vKSPBoONFLdcCjoqkQkDinQg1Y4FtK6Q/Wr7a5H3xlwUSISjxToQeuWCUXjoOoVivKyGNCrOwurFOgicvoU6F3BkGtgVxW2dxPjh+SzuHoXLa1tQVclInFGgd4VDJ4Ueqyax/jSfA4cbWHVln3B1iQicUeB3hXkl0LPQqh+lSvCffSFG9R2EZHTo0DvCsxgyCSoeZ287imM6H8OC9RHF5HTpEDvKgZPgqYDUPcm40vzeXvzHg41tgRdlYjEEQV6V1EyMTQNQPU8rhxSQHOrs2yjvjUqIpFToHcVmT2hcDRUvUJ5UW8y0lJYuEGBLiKRiyjQzWyyma0zsyoze6CD7V8zs9Vm9q6ZzTOzQdEvNQkMngRb3yGzcTeXFeWysKoh6IpEJI6cMtDNLBV4DLgOKAOmmVnZccPeBsrdfSQwC/hBtAtNCkPCly9Why5fXL/9INv3Hw22JhGJG5GcoY8Gqty9xt2bgJnAlPYD3H2+ux+byHspMDC6ZSaJcy+GnL6w7kXGD8kHdPmiiEQukkAfANS1W64PrzuRu4EXO9pgZtPNrMLMKhoa1E74iJQUGDoZquZRVpBBfk4G89ftCLoqEYkTUf1Q1MzuAMqBH3a03d1nuHu5u5cXFBRE860Tx7DroekAKZsX8rFhBbyxvoFmTQMgIhGIJNC3AIXtlgeG132ImV0DfBO4yd0bo1NeEiqZCN2yYN2LTDq/D/uPtrBi056gqxKROBBJoC8HSs2s2MzSganA7PYDzGwU8HNCYa4ewdno1h0GX/1BH71bqjF/rX5LReTUThno7t4C3Au8BKwBnnb3SjN72MxuCg/7IZAD/LeZrTSz2Sd4OYnEsOtg/xZydlcypjiPeQp0EYlAWiSD3H0OMOe4dQ+1e35NlOtKbkMnAwbrXuTq4VN5+PnVbN51mPPysoKuTES6MH1TtCvKzofCMbBuDpPO7wPAq2u3B1yUiHR1CvSuath1sO1dBqXuZnBBttouInJKCvSuavgNocd1c7h6eB+W1ezW7IsiclIK9K4qvxQKhkPls1w9vC9NrW2616iInJQCvSsbcQtsXkJ57hHOyUxjbqX66CJyYgr0rmzEzYDTbd1zXFvWj5dXb6OpRd8aFZGOKdC7soKh0PcCqPwjN4zsx/6jLSxS20VETkCB3tWN+ATULWN8QSM9MtN4YdXWoCsSkS5Kgd7VjbgFgPR1s7m2rC9zK9V2EZGOKdC7urzB0G8kVD7DDReeG2q7VKvtIiIfpUCPBxfcAltWML7gMD0y0pjzrtouIvJRCvR4MOJmADLWPhtqu6zerraLiHyEAj0e9C6CgaPhnZlcf0E/9h1pVttFRD5CgR4vLr4dGtYyoUc9PTLSeP4dtV1E5MMU6PFixM2Qlkn6qpncMPJcXnxvq+Z2EZEPUaDHi+69YPiNsOq/uW1UAYebWnVNuoh8iAI9nlw8DY7uZdSRZZTkZzOroj7oikSkC1Ggx5OSj0GPc7F3nuKTlw7kzdrd1O48FHRVItJFKNDjSUoqXDQVNrzMp4alk2Lwh7d0li4iIREFuplNNrN1ZlZlZg90sH2Cmb1lZi1mdmv0y5QPXHQ7eCt9Nj7LlaUF/GFFPa1tHnRVItIFnDLQzSwVeAy4DigDpplZ2XHDNgOfA56MdoFynIKhofuNrvhPbr2kP+/vO8qS6l1BVyUiXUAkZ+ijgSp3r3H3JmAmMKX9AHevdfd3AX19sTNc9nnYXc3HM1dzTmYaM5dvDroiEekCIgn0AUBdu+X68LrTZmbTzazCzCoaGhrO5CUEoGwKZPch/a1f8qnyQv783ja27TsadFUiErBO/VDU3We4e7m7lxcUFHTmWyeWtHS49E5Y/2fuKkuh1Z3fLq0NuioRCVgkgb4FKGy3PDC8ToJ06d+ApTCg6ndcc35fnly2maPNrUFXJSIBiiTQlwOlZlZsZunAVGB2bMuSU+o5AIbfAG//lrvH9GPP4Wb+tFL/zooks1MGuru3APcCLwFrgKfdvdLMHjazmwDM7DIzqwc+BfzczCpjWbSEjZ4OR/Yw5tB8hvfrwa8W1eKuSxhFklVEPXR3n+PuQ919sLt/N7zuIXefHX6+3N0Hunu2u+e5+4hYFi1hReOhTxm29GfcdcUg1m47oEsYRZKYvikaz8xg/Fdhx2o+kfUOudnpPLFwY9BViUhAFOjxbsQt0LuY9EU/4m8uH8Sra3ewsm5v0FWJSAAU6PEuNQ2u/DpsXck9/WvondWNR19eH3RVIhIABXoiGPlp6FlI98U/4gsTSnhjfQMVtbuDrkpEOpkCPRGkpcP4+6H+TT7XfzP5ORn8aK7O0kWSjQI9UVx8B+T0I3PB9/nSxBKW1OxicZVuJC2STBToiaJbJnzsH6FuKXfkLKffOZn84KV1tGlqXZGkoUBPJKM+C/0vIX3eQzw4aSAr6/by5JuaiVEkWSjQE0lKClz/Qzi4jZv2/RfjhuTx/RfXsn2/ZmIUSQYK9EQzsBxG3YEtfZwfTMykqbWN7zynmRhEkoECPRFN+jZ0y2LAom9x39UlzFm1jVdWbw+6KhGJMQV6IsopgI9/F2oX8AWfxbC+Pfjms6vYodaLSEJToCeqUXfAxZ8hdcEP+cXle9h/pIXpv12hOdNFEpgCPVGZwfWPQN8RnPf6fTx+Yz4r6/byj39cpSl2RRKUAj2RpWfBbb+Btlauevt+HpzYh2fe2sIvFtQEXZmIxIACPdHlDYZbfwUN65lecy+3n5/Gv8xZy09f3aAzdZEEo0BPBqXXwB1/wPbV893d/5t7ytp4ZO56HnxmFc2tbUFXJyJRokBPFsVXwp3PYU0H+ebWr/CzCzcwc/lm/uZXy6nfczjo6kQkChToyWTAJXD3XCy3hOs3/F+WDvz/7Ni0mkk/ep0fv7JeV8CIxDkLqo9aXl7uFRUVgbx30mtrhRW/gle+gzcf5p2sy3l09xVU5VzGbaMH8clLBlKYmxV0lSLSATNb4e7lHW6LJNDNbDLwEyAVeMLdv3fc9gzgN8ClwC7g0+5ee7LXVKB3Afu3wtLHYOWTcHgXO1MLmNc0gqWt59PYfwylQ8u4rDiPi8/rRU5GWtDVighnGehmlgqsB64F6oHlwDR3X91uzJeAke7+RTObCtzs7p8+2esq0LuQlkZY8xy89wxttYtIaQzdk/SQZ7DRz2Wj9+NQegFtWX1I69mX9JxcMnN6kXVOb7pn9SAzKzv0mNmd9IwMMjIyyeiWRlqKkZpimFnAOyiSOE4W6JGcdo0Gqty9JvxiM4EpwOp2Y6YA3w4/nwX81MzMdV1cfEjLgAtvhQtvJaWtDXZUQt2bpG1fR/8taxi4ZyPdG1eSceAoHIjsJVs8hVZSaSSFtnY/bkYbhmM4KYT+gBxbDj+30PIx7Z9zgvUf/oNmHT798OucieT+h0l/maNn16X3c+kN90T9dSMJ9AFAXbvlemDMica4e4uZ7QPygA/dMsfMpgPTAc4777wzLFliKiUF+l0I/S4kA8hov63xIBzaQevhfRzcv4uD+3bTePggzY2HaW48TGtTI20tjXhLE22tLdDWAq3NuLdh3oq3tYI7eBsQfvRQlHv4kQ/OAUKXU5q3D5K/PLP28fKh84YTxM4ZnFuYIixAif17n56TG5PX7dTGqLvPAGZAqOXSme8tUZCRAxk5pOZCT0I/ItJ1RHLZ4hagsN3ywPC6DseYWRqhv+u7olGgiIhEJpJAXw6UmlmxmaUDU4HZx42ZDdwZfn4r8Kr65yIineuULZdwT/xe4CVCly3+0t0rzexhoMLdZwP/AfzWzKqA3YRCX0REOlFEPXR3nwPMOW7dQ+2eHwU+Fd3SRETkdOir/yIiCUKBLiKSIBToIiIJQoEuIpIgAptt0cwagE1n+MvzOe5bqEkiGfc7GfcZknO/k3Gf4fT3e5C7F3S0IbBAPxtmVnGiyWkSWTLudzLuMyTnfifjPkN091stFxGRBKFAFxFJEPEa6DOCLiAgybjfybjPkJz7nYz7DFHc77jsoYuIyEfF6xm6iIgcR4EuIpIg4i7QzWyyma0zsyozeyDoemLBzArNbL6ZrTazSjO7L7w+18xeNrMN4cfeQdcabWaWamZvm9nz4eViM1sWPt6/D0/hnFDMrJeZzTKztWa2xswuT5Jj/dXwn+/3zOwpM8tMtONtZr80sx1m9l67dR0eWwv5f+F9f9fMLjnd94urQA/fsPox4DqgDJhmZmXBVhUTLcDX3b0MGAt8ObyfDwDz3L0UmBdeTjT3AWvaLX8f+Dd3HwLsAe4OpKrY+gnwZ3cfDlxEaP8T+lib2QDgK0C5u19AaGruqSTe8f5PYPJx6050bK8DSsM/04HHT/fN4irQaXfDandvAo7dsDqhuPtWd38r/PwAob/gAwjt66/Dw34NfCKYCmPDzAYCNwBPhJcNuJrQjcchMfe5JzCB0D0FcPcmd99Lgh/rsDSge/guZ1nAVhLseLv7G4TuEdHeiY7tFOA3HrIU6GVm557O+8VboHd0w+oBAdXSKcysCBgFLAP6uvvW8KZtQN+AyoqVHwN/z7E7RIduNL7X3VvCy4l4vIuBBuBX4VbTE2aWTYIfa3ffAjwCbCYU5PuAFST+8YYTH9uzzrd4C/SkYmY5wB+A+919f/tt4Vv8Jcw1p2Z2I7DD3VcEXUsnSwMuAR5391HAIY5rryTasQYI942nEPoHrT+QzUdbEwkv2sc23gI9khtWJwQz60YozH/n7s+EV28/9l+w8OOOoOqLgXHATWZWS6iVdjWh3nKv8H/JIZpEZDoAAAE4SURBVDGPdz1Q7+7LwsuzCAV8Ih9rgGuAje7e4O7NwDOE/gwk+vGGEx/bs863eAv0SG5YHffCveP/ANa4+6PtNrW/GfedwJ86u7ZYcfcH3X2guxcROq6vuvtngPmEbjwOCbbPAO6+Dagzs2HhVZOA1STwsQ7bDIw1s6zwn/dj+53QxzvsRMd2NvDX4atdxgL72rVmIuPucfUDXA+sB6qBbwZdT4z2cTyh/4a9C6wM/1xPqKc8D9gAvALkBl1rjPb/KuD58PMS4E2gCvhvICPo+mKwvxcDFeHj/SzQOxmONfAdYC3wHvBbICPRjjfwFKHPCJoJ/W/s7hMdW8AIXcVXDawidAXQab2fvvovIpIg4q3lIiIiJ6BAFxFJEAp0EZEEoUAXEUkQCnQRkQShQBcRSRAKdBGRBPE/wyCZVIXZVjgAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "1/1 [==============================] - 0s 15ms/step - loss: 8.2850e-06 - accuracy: 1.0000\n",
            "loss= 8.284978321171366e-06\n",
            "acc= 1.0\n",
            "WARNING:tensorflow:5 out of the last 5 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7f1aa01c1050> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kDUU5YAZksVp"
      },
      "source": [
        "## CNN"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1D1n9lG9k2ID",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "99d80df4-bf4f-4c21-f137-04876fa92e2d"
      },
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras.layers import Dense, RepeatVector, Flatten\n",
        "from tensorflow.keras.layers import Bidirectional, LSTM, GRU\n",
        "from tensorflow.keras.layers import Conv2D, MaxPooling2D, Conv1D, MaxPooling1D, Input, Reshape\n",
        "\n",
        "input_shape = train_x.shape[1:]\n",
        "# input_shape = (5,10)\n",
        "\n",
        "model = keras.Sequential()\n",
        "model.add(Conv1D(32, 3, padding=\"same\", input_shape=input_shape))\n",
        "model.add(Conv1D(32, 3, padding=\"same\"))\n",
        "model.add(MaxPooling1D())\n",
        "model.add(Conv1D(32, 3, padding=\"same\"))\n",
        "model.add(Conv1D(32, 3, padding=\"same\"))\n",
        "model.add(MaxPooling1D())\n",
        "model.add(Flatten())\n",
        "model.add(Dense(CATEGORY_NUM, activation=\"softmax\"))\n",
        "model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
        "model.summary()\n",
        "\n",
        "history = model.fit(train_x, train_y, epochs=100, verbose=1, validation_split=0.1, batch_size=128)\n",
        "\n",
        "\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.show()\n",
        "\n",
        "\n",
        "\n",
        "loss, acc = model.evaluate(test_x, test_y)\n",
        "print(\"loss=\", loss)\n",
        "print(\"acc=\", acc)\n",
        "\n",
        "\n",
        "\n",
        "y_ = model.predict(test_x)\n",
        "predicted = np.argmax(y_, axis=-1)\n",
        "\n",
        "plt.plot(test_y, \"o\")\n",
        "plt.plot(predicted, \".\")\n",
        "plt.show()"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_5\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv1d (Conv1D)              (None, 5, 32)             992       \n",
            "_________________________________________________________________\n",
            "conv1d_1 (Conv1D)            (None, 5, 32)             3104      \n",
            "_________________________________________________________________\n",
            "max_pooling1d (MaxPooling1D) (None, 2, 32)             0         \n",
            "_________________________________________________________________\n",
            "conv1d_2 (Conv1D)            (None, 2, 32)             3104      \n",
            "_________________________________________________________________\n",
            "conv1d_3 (Conv1D)            (None, 2, 32)             3104      \n",
            "_________________________________________________________________\n",
            "max_pooling1d_1 (MaxPooling1 (None, 1, 32)             0         \n",
            "_________________________________________________________________\n",
            "flatten_5 (Flatten)          (None, 32)                0         \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 2)                 66        \n",
            "=================================================================\n",
            "Total params: 10,370\n",
            "Trainable params: 10,370\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "1/1 [==============================] - 0s 22ms/step - loss: 2.9328e-04 - accuracy: 1.0000\n",
            "loss= 0.00029327557422220707\n",
            "acc= 1.0\n",
            "WARNING:tensorflow:6 out of the last 6 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7f1ab619e9e0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FUjeezafksMt"
      },
      "source": [
        "## CNN + LSTM"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "emE8KhR9BF-Z",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 891
        },
        "outputId": "83dd0b1e-d927-45ba-d951-00dd4bb6d7ff"
      },
      "source": [
        "from tensorflow import keras\n",
        "from tensorflow.keras.layers import Dense, RepeatVector, Flatten\n",
        "from tensorflow.keras.layers import Bidirectional, LSTM, GRU\n",
        "from tensorflow.keras.layers import Conv2D, MaxPooling2D, Conv1D, MaxPooling1D, Input, Reshape\n",
        "\n",
        "input_shape = train_x.shape[1:]\n",
        "# input_shape = (5,10)\n",
        "\n",
        "model = keras.Sequential()\n",
        "model.add(Conv1D(32, 3, padding=\"same\", input_shape=input_shape))\n",
        "model.add(Conv1D(32, 3, padding=\"same\"))\n",
        "model.add(LSTM(128))\n",
        "model.add(Flatten())\n",
        "model.add(Dense(CATEGORY_NUM, activation=\"softmax\"))\n",
        "model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
        "model.summary()\n",
        "\n",
        "history = model.fit(train_x, train_y, epochs=100, verbose=1, validation_split=0.1, batch_size=128)\n",
        "\n",
        "\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.show()\n",
        "\n",
        "\n",
        "\n",
        "loss, acc = model.evaluate(test_x, test_y)\n",
        "print(\"loss=\", loss)\n",
        "print(\"acc=\", acc)\n",
        "\n",
        "\n",
        "\n",
        "y_ = model.predict(test_x)\n",
        "predicted = np.argmax(y_, axis=-1)\n",
        "\n",
        "plt.plot(test_y, \"o\")\n",
        "plt.plot(predicted, \".\")\n",
        "plt.show()"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_6\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv1d_4 (Conv1D)            (None, 5, 32)             992       \n",
            "_________________________________________________________________\n",
            "conv1d_5 (Conv1D)            (None, 5, 32)             3104      \n",
            "_________________________________________________________________\n",
            "lstm_4 (LSTM)                (None, 128)               82432     \n",
            "_________________________________________________________________\n",
            "flatten_6 (Flatten)          (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_6 (Dense)              (None, 2)                 258       \n",
            "=================================================================\n",
            "Total params: 86,786\n",
            "Trainable params: 86,786\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "1/1 [==============================] - 0s 15ms/step - loss: 1.8739e-05 - accuracy: 1.0000\n",
            "loss= 1.8739126971922815e-05\n",
            "acc= 1.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BFJ-cQY9l62C"
      },
      "source": [
        ""
      ],
      "execution_count": 27,
      "outputs": []
    }
  ]
}