python_tinyML/01-linear_regression_loss_function.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "01-linear-regression-loss-function.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyN4lrgjcJOirfieEi5iPMXf",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ferris77/my_notebooks/blob/main/python_tinyML/01-linear_regression_loss_function.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "avBxfkxqNGq8"
      },
      "source": [
        "Este notebook foi feito com base no material fornecido em aula durante o cuso Fundamentals of TinyML, na plataforma edX, como o único propósito de servir como meu guia de meus estudos pessoais.\n",
        "Link para acesso ao curso: https://courses.edx.org/dashboard/programs/f7868191-7d7f-4292-b117-64549f1f483a/"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AVgFQmmeHpkW"
      },
      "source": [
        "Machine Learning envolve uma mudança de paradigma: Enviamos dados + respostas para o modelo e ele nos retorna a regra que leva os dados às respostas.\n",
        "Neste primeiro exemplo, vamos explorar a funcionalidade para buscar a regra (função) que leva os valores X em / f(x) = y\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Pzh2LTDMJjsz"
      },
      "source": [
        "Vamos utilizar como exemplo um conjunto de pares ordenados (x,y). Nosso objetivo será tentar encontrar uma função que correlacione o x com o y f(x) = y. Podemos plotar estes valores para ver como eles estão distribuidos no plano cartesiano:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "CBWdtuPwJ1OZ",
        "outputId": "f4193cc7-2f0b-4bfa-d962-4528958bcf15"
      },
      "source": [
        "import matplotlib\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "\n",
        "# Dados que iremos plotar\n",
        "x = [-1, 0, 1, 2, 3, 4]\n",
        "y = [-3, -1, 1, 3, 5, 7]\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "ax.plot(x, y)\n",
        "\n",
        "ax.set(xlabel='X', ylabel='Y',\n",
        "       title='f(x) = y')\n",
        "ax.grid()\n",
        "\n",
        "fig.savefig(\"test.png\")\n",
        "plt.show()"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kNk7rIDpKZpe"
      },
      "source": [
        "Do desenho acima, ficou claro que a regra é uma reta. Vamos usar uma abordagem com ML para determinar esta função."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lDOso0FQHkWG",
        "outputId": "a1da3d3c-49ee-41ac-b56a-b6d2cfc21583"
      },
      "source": [
        "import math\n",
        "\n",
        "x = [-1, 0, 1, 2, 3, 4]\n",
        "y = [-3, -1, 1, 3, 5, 7]\n",
        "\n",
        "#vamos supor que o modelo é linear e estimar uma curva que encaixe o Y em função do valor de X\n",
        "# y = a * x + b\n",
        "\n",
        "a = 3\n",
        "b = -1\n",
        "\n",
        "#vamos armazenar os dados estimados para comparação posterior\n",
        "Y_estimado = []\n",
        "\n",
        "for item in x:\n",
        "  Y_calculado = a*item+b\n",
        "  Y_estimado.append(Y_calculado)\n",
        "\n",
        "print('Y real:', y)\n",
        "\n",
        "print('Y estimado:', Y_estimado)\n",
        "\n",
        "# vamos definir a função de erro como sendo a diferença entre o Y que desejamos e o que calculamos\n",
        "# vamos somar o erro de cada par de Y (calculado e esperado) para medir a precisão do modelo\n",
        "# como podemos ter valores positivos e negativos, vamos elevar todas diferenças ao quadrado para sempre somarmos números positivos\n",
        "\n",
        "erro_quadratico_total = 0\n",
        "for i in range(0,len(y)):\n",
        "  erro_quadratico = (y[i] - Y_estimado[i])**2\n",
        "  erro_quadratico_total = erro_quadratico_total + erro_quadratico\n",
        "\n",
        "print(f'Erro quadratico total {str(math.sqrt(erro_quadratico_total)):.4s}%', )"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Y real: [-3, -1, 1, 3, 5, 7]\n",
            "Y estimado: [-4, -1, 2, 5, 8, 11]\n",
            "Erro quadratico total 5.56%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "G7GI4dDLK_Bq"
      },
      "source": [
        "Vamos usar o código fornecido em aula para definir uma função que calcula o erro entre o valor esperado de Y e o valor real. Se minimizarmos esta função, teremos a melhor curva para nosso modelo (no caso, uma reta). Para tanto, iremos definir um palpite inicial (um chute), e vamos ajustando os parâmetros a e b de acordo com o erro. Enquanto o erro estiver fora de uma tolerância mínima, vamos caminhar ao longo do eixo x com pequenos incrementos. Como a função de erro quadrático é uma parábola, podemos usar o gradiente desta funçaõ para caminhar em direção ao valor mínimo. Dessa forma, iremos calcular a função gradiente do erro quadrático e avaliá-la continuamente pelo eixo x. De acordo com o valor do gradiente para certo x, incrementamos o valor de x em um pequeno passo, até chegar no valor mínimo do erro."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ycX2mzBdPi5r"
      },
      "source": [
        "# First import the functions we will need\n",
        "from __future__ import absolute_import, division, print_function, unicode_literals\n",
        "\n",
        "try:\n",
        "  # %tensorflow_version only exists in Colab.\n",
        "  %tensorflow_version 2.x\n",
        "except Exception:\n",
        "  pass\n",
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c6sGE8icLKbt"
      },
      "source": [
        "# Nosso palpite inicial\n",
        "a_inicial = 10.0\n",
        "b_inicial = 10.0\n",
        "\n",
        "# Vamos definir uma função de erro\n",
        "def loss(y_esperado, y_conhecido):\n",
        "  return tf.reduce_mean(tf.square(y_esperado - y_conhecido))\n",
        "\n",
        "# Vamos definir nosso procedimento de trieno\n",
        "def train(model, inputs, outputs, learning_rate):\n",
        "  with tf.GradientTape() as t:\n",
        "    current_loss = loss(model(inputs), outputs)\n",
        "  # aqui iremos diferenciar o modelo (calcular o gradiente)\n",
        "  da, db = t.gradient(current_loss, [model.a, model.b])\n",
        "  # aqui iremos atualizar o modelo com base na taxa de aprendizado (learning_rate)\n",
        "  model.a.assign_sub(learning_rate * da)\n",
        "  model.b.assign_sub(learning_rate * db)\n",
        "  return current_loss\n",
        "\n",
        "# Vamos definir nosso modelo de regressão linear\n",
        "class Model(object):\n",
        "  def __init__(self):\n",
        "    # inicializar com os valores de chute\n",
        "    self.a = tf.Variable(a_inicial)\n",
        "    self.b = tf.Variable(b_inicial)\n",
        "\n",
        "  def __call__(self, x):\n",
        "    return self.a * x + self.b"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2bWQU7E9OSOA"
      },
      "source": [
        "Agora, vamos treinar o modelo que criamos:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fMbGJir2OU4x",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "90ce99d9-d346-4017-8d81-8dd3e64f5b04"
      },
      "source": [
        "# vamos definir os valores de entrada para x, y e a taxa de aprendizado\n",
        "xs = [-1.0, 0.0, 1.0, 2.0, 3.0, 4.0]\n",
        "ys = [-3.0, -1.0, 1.0, 3.0, 5.0, 7.0]\n",
        "LEARNING_RATE=0.09\n",
        "\n",
        "# Instanciamos o modelo\n",
        "model = Model()\n",
        "\n",
        "# armazenar os valores de a e b para vermos depois\n",
        "lista_a, lista_b = [], []\n",
        "epochs = range(50)\n",
        "losses = []\n",
        "for epoch in epochs:\n",
        "  lista_a.append(model.a.numpy())\n",
        "  lista_b.append(model.b.numpy())\n",
        "  current_loss = train(model, xs, ys, learning_rate=LEARNING_RATE)\n",
        "  losses.append(current_loss)\n",
        "  print('Epoch %2d: a=%1.2f b=%1.2f, loss=%2.5f' %\n",
        "        (epoch, lista_a[-1], lista_b[-1], current_loss))"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch  0: a=10.00 b=10.00, loss=715.66669\n",
            "Epoch  1: a=-0.41 b=5.86, loss=27.47032\n",
            "Epoch  2: a=-0.02 b=5.28, loss=22.43888\n",
            "Epoch  3: a=0.16 b=4.69, loss=18.46284\n",
            "Epoch  4: a=0.33 b=4.16, loss=15.19137\n",
            "Epoch  5: a=0.49 b=3.68, loss=12.49958\n",
            "Epoch  6: a=0.63 b=3.25, loss=10.28476\n",
            "Epoch  7: a=0.76 b=2.85, loss=8.46238\n",
            "Epoch  8: a=0.87 b=2.50, loss=6.96291\n",
            "Epoch  9: a=0.98 b=2.17, loss=5.72914\n",
            "Epoch 10: a=1.07 b=1.88, loss=4.71398\n",
            "Epoch 11: a=1.16 b=1.61, loss=3.87870\n",
            "Epoch 12: a=1.24 b=1.37, loss=3.19143\n",
            "Epoch 13: a=1.31 b=1.15, loss=2.62593\n",
            "Epoch 14: a=1.37 b=0.95, loss=2.16064\n",
            "Epoch 15: a=1.43 b=0.77, loss=1.77779\n",
            "Epoch 16: a=1.48 b=0.60, loss=1.46278\n",
            "Epoch 17: a=1.53 b=0.45, loss=1.20359\n",
            "Epoch 18: a=1.57 b=0.32, loss=0.99032\n",
            "Epoch 19: a=1.61 b=0.20, loss=0.81484\n",
            "Epoch 20: a=1.65 b=0.08, loss=0.67046\n",
            "Epoch 21: a=1.68 b=-0.02, loss=0.55166\n",
            "Epoch 22: a=1.71 b=-0.11, loss=0.45391\n",
            "Epoch 23: a=1.74 b=-0.19, loss=0.37348\n",
            "Epoch 24: a=1.76 b=-0.27, loss=0.30730\n",
            "Epoch 25: a=1.79 b=-0.33, loss=0.25285\n",
            "Epoch 26: a=1.81 b=-0.40, loss=0.20805\n",
            "Epoch 27: a=1.82 b=-0.45, loss=0.17118\n",
            "Epoch 28: a=1.84 b=-0.50, loss=0.14085\n",
            "Epoch 29: a=1.85 b=-0.55, loss=0.11589\n",
            "Epoch 30: a=1.87 b=-0.59, loss=0.09536\n",
            "Epoch 31: a=1.88 b=-0.63, loss=0.07846\n",
            "Epoch 32: a=1.89 b=-0.66, loss=0.06456\n",
            "Epoch 33: a=1.90 b=-0.69, loss=0.05312\n",
            "Epoch 34: a=1.91 b=-0.72, loss=0.04371\n",
            "Epoch 35: a=1.92 b=-0.75, loss=0.03596\n",
            "Epoch 36: a=1.93 b=-0.77, loss=0.02959\n",
            "Epoch 37: a=1.93 b=-0.79, loss=0.02435\n",
            "Epoch 38: a=1.94 b=-0.81, loss=0.02003\n",
            "Epoch 39: a=1.95 b=-0.83, loss=0.01648\n",
            "Epoch 40: a=1.95 b=-0.85, loss=0.01356\n",
            "Epoch 41: a=1.95 b=-0.86, loss=0.01116\n",
            "Epoch 42: a=1.96 b=-0.87, loss=0.00918\n",
            "Epoch 43: a=1.96 b=-0.88, loss=0.00756\n",
            "Epoch 44: a=1.97 b=-0.90, loss=0.00622\n",
            "Epoch 45: a=1.97 b=-0.91, loss=0.00511\n",
            "Epoch 46: a=1.97 b=-0.91, loss=0.00421\n",
            "Epoch 47: a=1.97 b=-0.92, loss=0.00346\n",
            "Epoch 48: a=1.98 b=-0.93, loss=0.00285\n",
            "Epoch 49: a=1.98 b=-0.94, loss=0.00234\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vq5NgfPvRLZx",
        "outputId": "14eacf25-a5ae-4980-8226-7ee7ef119f27"
      },
      "source": [
        "print(f'Para x=10, nosso modelo retorna y={lista_a[-1]*10 + lista_b[-1]:.2f}')"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Para x=10, nosso modelo retorna y=18.86\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qZLp8KjiQz8t"
      },
      "source": [
        "Vamos utilizar o TensorFlow para criar um modelo de aprendizado com redes neurais e comparar o valor com nossa análise acima. Uma rede neural com apenas um neurônio pode fazer isso, uma vez que, neste caso, buscamos uma função que aprenda dois parâmetros (a e b). Caso tenhamos mais neurônios em camadas podemos amplificar esta capacidade de associar parâmetros, aprendendo padrões mais complexos."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gI9RHJ8GRBC4"
      },
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "from tensorflow import keras"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D2B5cKYnYPrF"
      },
      "source": [
        "Estamos utilizando uma rede neural com neurônios densamente conectados (Dense layer). Há também outros tipos: convolucionais, recorrentes, etc. Há layers que não aprendem parâmetros, mas afetam as demais layers. Podem ser layers para reduzir a densidade das conexões (dropouts), reduzir a quantidade de dados fluindo pelas redes (pooling) e outras."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TPaSjvRURCpM",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1ed07b8e-5560-47e7-b725-9ea926044e86"
      },
      "source": [
        "model = tf.keras.Sequential([keras.layers.Dense(units=1, input_shape=[1])])\n",
        "model.compile(optimizer='sgd', loss='mean_squared_error')\n",
        "\n",
        "xs = np.array([-1.0,  0.0, 1.0, 2.0, 3.0, 4.0], dtype=float)\n",
        "ys = np.array([-3.0, -1.0, 1.0, 3.0, 5.0, 7.0], dtype=float)\n",
        "\n",
        "model.fit(xs, ys, epochs=500)"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/500\n",
            "1/1 [==============================] - 0s 368ms/step - loss: 13.5659\n",
            "Epoch 2/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 10.9052\n",
            "Epoch 3/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 8.8071\n",
            "Epoch 4/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 7.1517\n",
            "Epoch 5/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 5.8448\n",
            "Epoch 6/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.8121\n",
            "Epoch 7/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.9952\n",
            "Epoch 8/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 3.3483\n",
            "Epoch 9/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 2.8350\n",
            "Epoch 10/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 2.4271\n",
            "Epoch 11/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.1022\n",
            "Epoch 12/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.8426\n",
            "Epoch 13/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.6344\n",
            "Epoch 14/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.4669\n",
            "Epoch 15/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.3313\n",
            "Epoch 16/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.2211\n",
            "Epoch 17/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.1308\n",
            "Epoch 18/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.0562\n",
            "Epoch 19/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.9941\n",
            "Epoch 20/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.9419\n",
            "Epoch 21/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.8976\n",
            "Epoch 22/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.8595\n",
            "Epoch 23/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.8264\n",
            "Epoch 24/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.7972\n",
            "Epoch 25/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.7713\n",
            "Epoch 26/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.7479\n",
            "Epoch 27/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.7266\n",
            "Epoch 28/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.7070\n",
            "Epoch 29/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.6888\n",
            "Epoch 30/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.6718\n",
            "Epoch 31/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.6557\n",
            "Epoch 32/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.6405\n",
            "Epoch 33/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.6259\n",
            "Epoch 34/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.6119\n",
            "Epoch 35/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.5985\n",
            "Epoch 36/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.5855\n",
            "Epoch 37/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.5729\n",
            "Epoch 38/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.5608\n",
            "Epoch 39/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.5489\n",
            "Epoch 40/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.5374\n",
            "Epoch 41/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.5261\n",
            "Epoch 42/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.5152\n",
            "Epoch 43/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.5044\n",
            "Epoch 44/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.4940\n",
            "Epoch 45/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.4838\n",
            "Epoch 46/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.4738\n",
            "Epoch 47/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.4640\n",
            "Epoch 48/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.4544\n",
            "Epoch 49/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.4450\n",
            "Epoch 50/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.4359\n",
            "Epoch 51/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.4269\n",
            "Epoch 52/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.4181\n",
            "Epoch 53/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.4095\n",
            "Epoch 54/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.4011\n",
            "Epoch 55/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.3929\n",
            "Epoch 56/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.3848\n",
            "Epoch 57/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.3769\n",
            "Epoch 58/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.3691\n",
            "Epoch 59/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.3615\n",
            "Epoch 60/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.3541\n",
            "Epoch 61/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.3468\n",
            "Epoch 62/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.3397\n",
            "Epoch 63/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.3327\n",
            "Epoch 64/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.3259\n",
            "Epoch 65/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.3192\n",
            "Epoch 66/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.3126\n",
            "Epoch 67/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.3062\n",
            "Epoch 68/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2999\n",
            "Epoch 69/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2938\n",
            "Epoch 70/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.2877\n",
            "Epoch 71/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2818\n",
            "Epoch 72/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.2760\n",
            "Epoch 73/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2704\n",
            "Epoch 74/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2648\n",
            "Epoch 75/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2594\n",
            "Epoch 76/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.2540\n",
            "Epoch 77/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2488\n",
            "Epoch 78/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.2437\n",
            "Epoch 79/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.2387\n",
            "Epoch 80/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.2338\n",
            "Epoch 81/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2290\n",
            "Epoch 82/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2243\n",
            "Epoch 83/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2197\n",
            "Epoch 84/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2152\n",
            "Epoch 85/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.2108\n",
            "Epoch 86/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.2064\n",
            "Epoch 87/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.2022\n",
            "Epoch 88/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.1980\n",
            "Epoch 89/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1940\n",
            "Epoch 90/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1900\n",
            "Epoch 91/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1861\n",
            "Epoch 92/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.1823\n",
            "Epoch 93/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1785\n",
            "Epoch 94/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.1749\n",
            "Epoch 95/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1713\n",
            "Epoch 96/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1677\n",
            "Epoch 97/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 0.1643\n",
            "Epoch 98/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.1609\n",
            "Epoch 99/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.1576\n",
            "Epoch 100/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1544\n",
            "Epoch 101/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.1512\n",
            "Epoch 102/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.1481\n",
            "Epoch 103/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1451\n",
            "Epoch 104/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1421\n",
            "Epoch 105/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1392\n",
            "Epoch 106/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1363\n",
            "Epoch 107/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.1335\n",
            "Epoch 108/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1308\n",
            "Epoch 109/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.1281\n",
            "Epoch 110/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.1254\n",
            "Epoch 111/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.1229\n",
            "Epoch 112/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 0.1203\n",
            "Epoch 113/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.1179\n",
            "Epoch 114/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 0.1155\n",
            "Epoch 115/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1131\n",
            "Epoch 116/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1108\n",
            "Epoch 117/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.1085\n",
            "Epoch 118/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.1063\n",
            "Epoch 119/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1041\n",
            "Epoch 120/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.1019\n",
            "Epoch 121/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0998\n",
            "Epoch 122/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0978\n",
            "Epoch 123/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0958\n",
            "Epoch 124/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0938\n",
            "Epoch 125/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 0.0919\n",
            "Epoch 126/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0900\n",
            "Epoch 127/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0882\n",
            "Epoch 128/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0863\n",
            "Epoch 129/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0846\n",
            "Epoch 130/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0828\n",
            "Epoch 131/500\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.0811\n",
            "Epoch 132/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0795\n",
            "Epoch 133/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0778\n",
            "Epoch 134/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0762\n",
            "Epoch 135/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0747\n",
            "Epoch 136/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0731\n",
            "Epoch 137/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0716\n",
            "Epoch 138/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0702\n",
            "Epoch 139/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 0.0687\n",
            "Epoch 140/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0673\n",
            "Epoch 141/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 0.0659\n",
            "Epoch 142/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0646\n",
            "Epoch 143/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0632\n",
            "Epoch 144/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0619\n",
            "Epoch 145/500\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.0607\n",
            "Epoch 146/500\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 0.0594\n",
            "Epoch 147/500\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.0582\n",
            "Epoch 148/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0570\n",
            "Epoch 149/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0558\n",
            "Epoch 150/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0547\n",
            "Epoch 151/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0536\n",
            "Epoch 152/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0525\n",
            "Epoch 153/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0514\n",
            "Epoch 154/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0503\n",
            "Epoch 155/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0493\n",
            "Epoch 156/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0483\n",
            "Epoch 157/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0473\n",
            "Epoch 158/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0463\n",
            "Epoch 159/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0454\n",
            "Epoch 160/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0444\n",
            "Epoch 161/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0435\n",
            "Epoch 162/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0426\n",
            "Epoch 163/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0418\n",
            "Epoch 164/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0409\n",
            "Epoch 165/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0401\n",
            "Epoch 166/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0392\n",
            "Epoch 167/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0384\n",
            "Epoch 168/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0376\n",
            "Epoch 169/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0369\n",
            "Epoch 170/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0361\n",
            "Epoch 171/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0354\n",
            "Epoch 172/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0346\n",
            "Epoch 173/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0339\n",
            "Epoch 174/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0332\n",
            "Epoch 175/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 0.0326\n",
            "Epoch 176/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0319\n",
            "Epoch 177/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0312\n",
            "Epoch 178/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0306\n",
            "Epoch 179/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0300\n",
            "Epoch 180/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0293\n",
            "Epoch 181/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0287\n",
            "Epoch 182/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0281\n",
            "Epoch 183/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0276\n",
            "Epoch 184/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0270\n",
            "Epoch 185/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0265\n",
            "Epoch 186/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0259\n",
            "Epoch 187/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0254\n",
            "Epoch 188/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0249\n",
            "Epoch 189/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0243\n",
            "Epoch 190/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0238\n",
            "Epoch 191/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0234\n",
            "Epoch 192/500\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.0229\n",
            "Epoch 193/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0224\n",
            "Epoch 194/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0219\n",
            "Epoch 195/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0215\n",
            "Epoch 196/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0211\n",
            "Epoch 197/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0206\n",
            "Epoch 198/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 0.0202\n",
            "Epoch 199/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0198\n",
            "Epoch 200/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0194\n",
            "Epoch 201/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0190\n",
            "Epoch 202/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0186\n",
            "Epoch 203/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0182\n",
            "Epoch 204/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0178\n",
            "Epoch 205/500\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 0.0175\n",
            "Epoch 206/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0171\n",
            "Epoch 207/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0168\n",
            "Epoch 208/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0164\n",
            "Epoch 209/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0161\n",
            "Epoch 210/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0157\n",
            "Epoch 211/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0154\n",
            "Epoch 212/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0151\n",
            "Epoch 213/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0148\n",
            "Epoch 214/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0145\n",
            "Epoch 215/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0142\n",
            "Epoch 216/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0139\n",
            "Epoch 217/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0136\n",
            "Epoch 218/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0133\n",
            "Epoch 219/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0131\n",
            "Epoch 220/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0128\n",
            "Epoch 221/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0125\n",
            "Epoch 222/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0123\n",
            "Epoch 223/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0120\n",
            "Epoch 224/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0118\n",
            "Epoch 225/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0115\n",
            "Epoch 226/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0113\n",
            "Epoch 227/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0111\n",
            "Epoch 228/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0108\n",
            "Epoch 229/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0106\n",
            "Epoch 230/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0104\n",
            "Epoch 231/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0102\n",
            "Epoch 232/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0100\n",
            "Epoch 233/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0098\n",
            "Epoch 234/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0096\n",
            "Epoch 235/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0094\n",
            "Epoch 236/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0092\n",
            "Epoch 237/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0090\n",
            "Epoch 238/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0088\n",
            "Epoch 239/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0086\n",
            "Epoch 240/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0084\n",
            "Epoch 241/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0083\n",
            "Epoch 242/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0081\n",
            "Epoch 243/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 0.0079\n",
            "Epoch 244/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0078\n",
            "Epoch 245/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0076\n",
            "Epoch 246/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0075\n",
            "Epoch 247/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0073\n",
            "Epoch 248/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0072\n",
            "Epoch 249/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0070\n",
            "Epoch 250/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0069\n",
            "Epoch 251/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0067\n",
            "Epoch 252/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0066\n",
            "Epoch 253/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0064\n",
            "Epoch 254/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0063\n",
            "Epoch 255/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0062\n",
            "Epoch 256/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0061\n",
            "Epoch 257/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0059\n",
            "Epoch 258/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0058\n",
            "Epoch 259/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0057\n",
            "Epoch 260/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0056\n",
            "Epoch 261/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0055\n",
            "Epoch 262/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0054\n",
            "Epoch 263/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0052\n",
            "Epoch 264/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0051\n",
            "Epoch 265/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0050\n",
            "Epoch 266/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0049\n",
            "Epoch 267/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0048\n",
            "Epoch 268/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0047\n",
            "Epoch 269/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0046\n",
            "Epoch 270/500\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 0.0045\n",
            "Epoch 271/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0044\n",
            "Epoch 272/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0043\n",
            "Epoch 273/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0043\n",
            "Epoch 274/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0042\n",
            "Epoch 275/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0041\n",
            "Epoch 276/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0040\n",
            "Epoch 277/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0039\n",
            "Epoch 278/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0038\n",
            "Epoch 279/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0038\n",
            "Epoch 280/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0037\n",
            "Epoch 281/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0036\n",
            "Epoch 282/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0035\n",
            "Epoch 283/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0035\n",
            "Epoch 284/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0034\n",
            "Epoch 285/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0033\n",
            "Epoch 286/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0033\n",
            "Epoch 287/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0032\n",
            "Epoch 288/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0031\n",
            "Epoch 289/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0031\n",
            "Epoch 290/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0030\n",
            "Epoch 291/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0029\n",
            "Epoch 292/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0029\n",
            "Epoch 293/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0028\n",
            "Epoch 294/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0028\n",
            "Epoch 295/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0027\n",
            "Epoch 296/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0026\n",
            "Epoch 297/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0026\n",
            "Epoch 298/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0025\n",
            "Epoch 299/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0025\n",
            "Epoch 300/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0024\n",
            "Epoch 301/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0024\n",
            "Epoch 302/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0023\n",
            "Epoch 303/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0023\n",
            "Epoch 304/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0022\n",
            "Epoch 305/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0022\n",
            "Epoch 306/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0021\n",
            "Epoch 307/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0021\n",
            "Epoch 308/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0021\n",
            "Epoch 309/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0020\n",
            "Epoch 310/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0020\n",
            "Epoch 311/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0019\n",
            "Epoch 312/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0019\n",
            "Epoch 313/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0019\n",
            "Epoch 314/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0018\n",
            "Epoch 315/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0018\n",
            "Epoch 316/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0017\n",
            "Epoch 317/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0017\n",
            "Epoch 318/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0017\n",
            "Epoch 319/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0016\n",
            "Epoch 320/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0016\n",
            "Epoch 321/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0016\n",
            "Epoch 322/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0015\n",
            "Epoch 323/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0015\n",
            "Epoch 324/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0015\n",
            "Epoch 325/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0014\n",
            "Epoch 326/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0014\n",
            "Epoch 327/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0014\n",
            "Epoch 328/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0014\n",
            "Epoch 329/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0013\n",
            "Epoch 330/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0013\n",
            "Epoch 331/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0013\n",
            "Epoch 332/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0013\n",
            "Epoch 333/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0012\n",
            "Epoch 334/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0012\n",
            "Epoch 335/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0012\n",
            "Epoch 336/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0012\n",
            "Epoch 337/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 0.0011\n",
            "Epoch 338/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0011\n",
            "Epoch 339/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0011\n",
            "Epoch 340/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0011\n",
            "Epoch 341/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 0.0010\n",
            "Epoch 342/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0010\n",
            "Epoch 343/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 9.9607e-04\n",
            "Epoch 344/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 9.7561e-04\n",
            "Epoch 345/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 9.5557e-04\n",
            "Epoch 346/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 9.3595e-04\n",
            "Epoch 347/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 9.1672e-04\n",
            "Epoch 348/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 8.9789e-04\n",
            "Epoch 349/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 8.7945e-04\n",
            "Epoch 350/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 8.6139e-04\n",
            "Epoch 351/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 8.4369e-04\n",
            "Epoch 352/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 8.2636e-04\n",
            "Epoch 353/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 8.0938e-04\n",
            "Epoch 354/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 7.9276e-04\n",
            "Epoch 355/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 7.7647e-04\n",
            "Epoch 356/500\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 7.6052e-04\n",
            "Epoch 357/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 7.4491e-04\n",
            "Epoch 358/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 7.2960e-04\n",
            "Epoch 359/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 7.1462e-04\n",
            "Epoch 360/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 6.9994e-04\n",
            "Epoch 361/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 6.8557e-04\n",
            "Epoch 362/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 6.7148e-04\n",
            "Epoch 363/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 6.5769e-04\n",
            "Epoch 364/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 6.4418e-04\n",
            "Epoch 365/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 6.3095e-04\n",
            "Epoch 366/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 6.1799e-04\n",
            "Epoch 367/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 6.0530e-04\n",
            "Epoch 368/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 5.9286e-04\n",
            "Epoch 369/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 5.8068e-04\n",
            "Epoch 370/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 5.6876e-04\n",
            "Epoch 371/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 5.5707e-04\n",
            "Epoch 372/500\n",
            "1/1 [==============================] - 0s 27ms/step - loss: 5.4563e-04\n",
            "Epoch 373/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 5.3442e-04\n",
            "Epoch 374/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.2344e-04\n",
            "Epoch 375/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 5.1270e-04\n",
            "Epoch 376/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.0216e-04\n",
            "Epoch 377/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 4.9185e-04\n",
            "Epoch 378/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 4.8175e-04\n",
            "Epoch 379/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 4.7185e-04\n",
            "Epoch 380/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 4.6216e-04\n",
            "Epoch 381/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 4.5267e-04\n",
            "Epoch 382/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 4.4337e-04\n",
            "Epoch 383/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 4.3426e-04\n",
            "Epoch 384/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 4.2534e-04\n",
            "Epoch 385/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.1661e-04\n",
            "Epoch 386/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 4.0805e-04\n",
            "Epoch 387/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.9967e-04\n",
            "Epoch 388/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.9145e-04\n",
            "Epoch 389/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.8341e-04\n",
            "Epoch 390/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.7554e-04\n",
            "Epoch 391/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.6783e-04\n",
            "Epoch 392/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 3.6027e-04\n",
            "Epoch 393/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.5287e-04\n",
            "Epoch 394/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.4562e-04\n",
            "Epoch 395/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.3852e-04\n",
            "Epoch 396/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.3157e-04\n",
            "Epoch 397/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.2476e-04\n",
            "Epoch 398/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.1809e-04\n",
            "Epoch 399/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.1156e-04\n",
            "Epoch 400/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.0516e-04\n",
            "Epoch 401/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 2.9889e-04\n",
            "Epoch 402/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.9275e-04\n",
            "Epoch 403/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.8673e-04\n",
            "Epoch 404/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.8084e-04\n",
            "Epoch 405/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.7507e-04\n",
            "Epoch 406/500\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 2.6943e-04\n",
            "Epoch 407/500\n",
            "1/1 [==============================] - 0s 19ms/step - loss: 2.6389e-04\n",
            "Epoch 408/500\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 2.5847e-04\n",
            "Epoch 409/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.5316e-04\n",
            "Epoch 410/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.4796e-04\n",
            "Epoch 411/500\n",
            "1/1 [==============================] - 0s 55ms/step - loss: 2.4287e-04\n",
            "Epoch 412/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.3788e-04\n",
            "Epoch 413/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.3299e-04\n",
            "Epoch 414/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 2.2820e-04\n",
            "Epoch 415/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.2352e-04\n",
            "Epoch 416/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 2.1893e-04\n",
            "Epoch 417/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 2.1443e-04\n",
            "Epoch 418/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 2.1002e-04\n",
            "Epoch 419/500\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 2.0571e-04\n",
            "Epoch 420/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.0148e-04\n",
            "Epoch 421/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 1.9734e-04\n",
            "Epoch 422/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.9329e-04\n",
            "Epoch 423/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 1.8932e-04\n",
            "Epoch 424/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 1.8543e-04\n",
            "Epoch 425/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.8162e-04\n",
            "Epoch 426/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.7789e-04\n",
            "Epoch 427/500\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 1.7424e-04\n",
            "Epoch 428/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.7066e-04\n",
            "Epoch 429/500\n",
            "1/1 [==============================] - 0s 21ms/step - loss: 1.6716e-04\n",
            "Epoch 430/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.6372e-04\n",
            "Epoch 431/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.6036e-04\n",
            "Epoch 432/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.5706e-04\n",
            "Epoch 433/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 1.5384e-04\n",
            "Epoch 434/500\n",
            "1/1 [==============================] - 0s 16ms/step - loss: 1.5068e-04\n",
            "Epoch 435/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 1.4758e-04\n",
            "Epoch 436/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.4455e-04\n",
            "Epoch 437/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.4158e-04\n",
            "Epoch 438/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 1.3868e-04\n",
            "Epoch 439/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.3583e-04\n",
            "Epoch 440/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 1.3304e-04\n",
            "Epoch 441/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.3030e-04\n",
            "Epoch 442/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.2763e-04\n",
            "Epoch 443/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 1.2501e-04\n",
            "Epoch 444/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.2244e-04\n",
            "Epoch 445/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.1993e-04\n",
            "Epoch 446/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.1746e-04\n",
            "Epoch 447/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.1505e-04\n",
            "Epoch 448/500\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 1.1269e-04\n",
            "Epoch 449/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.1037e-04\n",
            "Epoch 450/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.0810e-04\n",
            "Epoch 451/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 1.0588e-04\n",
            "Epoch 452/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 1.0371e-04\n",
            "Epoch 453/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.0158e-04\n",
            "Epoch 454/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 9.9490e-05\n",
            "Epoch 455/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 9.7448e-05\n",
            "Epoch 456/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 9.5446e-05\n",
            "Epoch 457/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 9.3486e-05\n",
            "Epoch 458/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 9.1565e-05\n",
            "Epoch 459/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 8.9684e-05\n",
            "Epoch 460/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 8.7842e-05\n",
            "Epoch 461/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 8.6038e-05\n",
            "Epoch 462/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 8.4270e-05\n",
            "Epoch 463/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 8.2540e-05\n",
            "Epoch 464/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 8.0844e-05\n",
            "Epoch 465/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 7.9183e-05\n",
            "Epoch 466/500\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 7.7556e-05\n",
            "Epoch 467/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 7.5963e-05\n",
            "Epoch 468/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 7.4403e-05\n",
            "Epoch 469/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 7.2875e-05\n",
            "Epoch 470/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 7.1378e-05\n",
            "Epoch 471/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 6.9911e-05\n",
            "Epoch 472/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 6.8476e-05\n",
            "Epoch 473/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 6.7068e-05\n",
            "Epoch 474/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 6.5691e-05\n",
            "Epoch 475/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 6.4341e-05\n",
            "Epoch 476/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 6.3021e-05\n",
            "Epoch 477/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 6.1725e-05\n",
            "Epoch 478/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 6.0458e-05\n",
            "Epoch 479/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.9215e-05\n",
            "Epoch 480/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 5.7999e-05\n",
            "Epoch 481/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 5.6807e-05\n",
            "Epoch 482/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 5.5641e-05\n",
            "Epoch 483/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 5.4498e-05\n",
            "Epoch 484/500\n",
            "1/1 [==============================] - 0s 32ms/step - loss: 5.3378e-05\n",
            "Epoch 485/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 5.2282e-05\n",
            "Epoch 486/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 5.1208e-05\n",
            "Epoch 487/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.0156e-05\n",
            "Epoch 488/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 4.9126e-05\n",
            "Epoch 489/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 4.8117e-05\n",
            "Epoch 490/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 4.7128e-05\n",
            "Epoch 491/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.6160e-05\n",
            "Epoch 492/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 4.5212e-05\n",
            "Epoch 493/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.4283e-05\n",
            "Epoch 494/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 4.3373e-05\n",
            "Epoch 495/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.2482e-05\n",
            "Epoch 496/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 4.1610e-05\n",
            "Epoch 497/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.0755e-05\n",
            "Epoch 498/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.9918e-05\n",
            "Epoch 499/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 3.9098e-05\n",
            "Epoch 500/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 3.8294e-05\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f05d986fad0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OGMJRq4eRJJm",
        "outputId": "aea3ec55-9f61-492b-9cb4-4a2cba961b65"
      },
      "source": [
        "print(f'Para x=10, nosso novo modelo retorna y={model.predict([10.0]).ravel()[0]:.2f}')"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Para x=10, nosso novo modelo retorna y=18.98\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "umQ2Pfobum3z"
      },
      "source": [
        "Vamos agora analisar a mesma rede, porém, com dois neurônios. Vamos analisar também, individualmente, como cada neurônio afeta o resultado final do modelo."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nVYzrfGouxS5"
      },
      "source": [
        "Com um neurônio apenas, teremos no final um modelo treinado que retorna uma função y = a*x + b, sendo a e b os pesos do neurônio. Caos adicionemos mais uma neurônio a rede, termos uma saída na forma y = a1*x1 + b1 + a2*x2 + b2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4fELBrgrRpvE",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "bb166762-e5d2-4137-b854-874d943a47cf"
      },
      "source": [
        "my_layer_1 = keras.layers.Dense(units=2, input_shape=[1])\n",
        "\n",
        "my_layer_2 = keras.layers.Dense(units=1)\n",
        "\n",
        "model = tf.keras.Sequential([my_layer_1, my_layer_2])\n",
        "\n",
        "model.compile(optimizer='sgd', loss='mean_squared_error')\n",
        "\n",
        " \n",
        "\n",
        "xs = np.array([-1.0,  0.0, 1.0, 2.0, 3.0, 4.0], dtype=float)\n",
        "\n",
        "ys = np.array([-3.0, -1.0, 1.0, 3.0, 5.0, 7.0], dtype=float)\n",
        "\n",
        " \n",
        "\n",
        "model.fit(xs, ys, epochs=500)"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/500\n",
            "1/1 [==============================] - 1s 610ms/step - loss: 4.7926\n",
            "Epoch 2/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 3.3067\n",
            "Epoch 3/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.2197\n",
            "Epoch 4/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.5070\n",
            "Epoch 5/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.0833\n",
            "Epoch 6/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.8479\n",
            "Epoch 7/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.7193\n",
            "Epoch 8/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.6450\n",
            "Epoch 9/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.5966\n",
            "Epoch 10/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.5602\n",
            "Epoch 11/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.5295\n",
            "Epoch 12/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 0.5018\n",
            "Epoch 13/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.4760\n",
            "Epoch 14/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.4516\n",
            "Epoch 15/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.4284\n",
            "Epoch 16/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.4063\n",
            "Epoch 17/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.3853\n",
            "Epoch 18/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.3652\n",
            "Epoch 19/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.3460\n",
            "Epoch 20/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.3278\n",
            "Epoch 21/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.3104\n",
            "Epoch 22/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.2939\n",
            "Epoch 23/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.2781\n",
            "Epoch 24/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2631\n",
            "Epoch 25/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.2488\n",
            "Epoch 26/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.2353\n",
            "Epoch 27/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.2224\n",
            "Epoch 28/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.2102\n",
            "Epoch 29/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.1985\n",
            "Epoch 30/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.1875\n",
            "Epoch 31/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.1770\n",
            "Epoch 32/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.1671\n",
            "Epoch 33/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.1577\n",
            "Epoch 34/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.1488\n",
            "Epoch 35/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.1403\n",
            "Epoch 36/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.1323\n",
            "Epoch 37/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.1247\n",
            "Epoch 38/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.1175\n",
            "Epoch 39/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.1107\n",
            "Epoch 40/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.1043\n",
            "Epoch 41/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0982\n",
            "Epoch 42/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0925\n",
            "Epoch 43/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0871\n",
            "Epoch 44/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 0.0819\n",
            "Epoch 45/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0771\n",
            "Epoch 46/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0725\n",
            "Epoch 47/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0682\n",
            "Epoch 48/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0641\n",
            "Epoch 49/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0603\n",
            "Epoch 50/500\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 0.0567\n",
            "Epoch 51/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0533\n",
            "Epoch 52/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0501\n",
            "Epoch 53/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0470\n",
            "Epoch 54/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0442\n",
            "Epoch 55/500\n",
            "1/1 [==============================] - 0s 20ms/step - loss: 0.0415\n",
            "Epoch 56/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0390\n",
            "Epoch 57/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0366\n",
            "Epoch 58/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 0.0343\n",
            "Epoch 59/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0322\n",
            "Epoch 60/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0302\n",
            "Epoch 61/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0284\n",
            "Epoch 62/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0266\n",
            "Epoch 63/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0250\n",
            "Epoch 64/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0234\n",
            "Epoch 65/500\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 0.0220\n",
            "Epoch 66/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0206\n",
            "Epoch 67/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0193\n",
            "Epoch 68/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0181\n",
            "Epoch 69/500\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0170\n",
            "Epoch 70/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0159\n",
            "Epoch 71/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0149\n",
            "Epoch 72/500\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 0.0140\n",
            "Epoch 73/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0131\n",
            "Epoch 74/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0122\n",
            "Epoch 75/500\n",
            "1/1 [==============================] - 0s 17ms/step - loss: 0.0115\n",
            "Epoch 76/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0107\n",
            "Epoch 77/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0101\n",
            "Epoch 78/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0094\n",
            "Epoch 79/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0088\n",
            "Epoch 80/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0083\n",
            "Epoch 81/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 0.0077\n",
            "Epoch 82/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0072\n",
            "Epoch 83/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0068\n",
            "Epoch 84/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0063\n",
            "Epoch 85/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0059\n",
            "Epoch 86/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0056\n",
            "Epoch 87/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0052\n",
            "Epoch 88/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0049\n",
            "Epoch 89/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0046\n",
            "Epoch 90/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0043\n",
            "Epoch 91/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0040\n",
            "Epoch 92/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 0.0037\n",
            "Epoch 93/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0035\n",
            "Epoch 94/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0033\n",
            "Epoch 95/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 0.0031\n",
            "Epoch 96/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0029\n",
            "Epoch 97/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0027\n",
            "Epoch 98/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0025\n",
            "Epoch 99/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0023\n",
            "Epoch 100/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0022\n",
            "Epoch 101/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0020\n",
            "Epoch 102/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0019\n",
            "Epoch 103/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0018\n",
            "Epoch 104/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0017\n",
            "Epoch 105/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0016\n",
            "Epoch 106/500\n",
            "1/1 [==============================] - 0s 19ms/step - loss: 0.0015\n",
            "Epoch 107/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0014\n",
            "Epoch 108/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 0.0013\n",
            "Epoch 109/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 0.0012\n",
            "Epoch 110/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 0.0011\n",
            "Epoch 111/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 0.0010\n",
            "Epoch 112/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 9.7243e-04\n",
            "Epoch 113/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 9.0889e-04\n",
            "Epoch 114/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 8.4948e-04\n",
            "Epoch 115/500\n",
            "1/1 [==============================] - 0s 15ms/step - loss: 7.9392e-04\n",
            "Epoch 116/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 7.4198e-04\n",
            "Epoch 117/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 6.9342e-04\n",
            "Epoch 118/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 6.4801e-04\n",
            "Epoch 119/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 6.0557e-04\n",
            "Epoch 120/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 5.6589e-04\n",
            "Epoch 121/500\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 5.2880e-04\n",
            "Epoch 122/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 4.9414e-04\n",
            "Epoch 123/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 4.6172e-04\n",
            "Epoch 124/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 4.3143e-04\n",
            "Epoch 125/500\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 4.0312e-04\n",
            "Epoch 126/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 3.7666e-04\n",
            "Epoch 127/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.5193e-04\n",
            "Epoch 128/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 3.2882e-04\n",
            "Epoch 129/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 3.0722e-04\n",
            "Epoch 130/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 2.8703e-04\n",
            "Epoch 131/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.6816e-04\n",
            "Epoch 132/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.5054e-04\n",
            "Epoch 133/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.3406e-04\n",
            "Epoch 134/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 2.1867e-04\n",
            "Epoch 135/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.0428e-04\n",
            "Epoch 136/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.9084e-04\n",
            "Epoch 137/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 1.7828e-04\n",
            "Epoch 138/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.6655e-04\n",
            "Epoch 139/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 1.5559e-04\n",
            "Epoch 140/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 1.4534e-04\n",
            "Epoch 141/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 1.3577e-04\n",
            "Epoch 142/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.2683e-04\n",
            "Epoch 143/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 1.1847e-04\n",
            "Epoch 144/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 1.1067e-04\n",
            "Epoch 145/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.0338e-04\n",
            "Epoch 146/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 9.6563e-05\n",
            "Epoch 147/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 9.0199e-05\n",
            "Epoch 148/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 8.4253e-05\n",
            "Epoch 149/500\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 7.8698e-05\n",
            "Epoch 150/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 7.3509e-05\n",
            "Epoch 151/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 6.8663e-05\n",
            "Epoch 152/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 6.4135e-05\n",
            "Epoch 153/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 5.9905e-05\n",
            "Epoch 154/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.5952e-05\n",
            "Epoch 155/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 5.2261e-05\n",
            "Epoch 156/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.8812e-05\n",
            "Epoch 157/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.5592e-05\n",
            "Epoch 158/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 4.2583e-05\n",
            "Epoch 159/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.9772e-05\n",
            "Epoch 160/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.7147e-05\n",
            "Epoch 161/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 3.4695e-05\n",
            "Epoch 162/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.2405e-05\n",
            "Epoch 163/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.0265e-05\n",
            "Epoch 164/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.8268e-05\n",
            "Epoch 165/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.6401e-05\n",
            "Epoch 166/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.4657e-05\n",
            "Epoch 167/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.3029e-05\n",
            "Epoch 168/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.1508e-05\n",
            "Epoch 169/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.0087e-05\n",
            "Epoch 170/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.8761e-05\n",
            "Epoch 171/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 1.7522e-05\n",
            "Epoch 172/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.6365e-05\n",
            "Epoch 173/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.5283e-05\n",
            "Epoch 174/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.4273e-05\n",
            "Epoch 175/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.3331e-05\n",
            "Epoch 176/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.2450e-05\n",
            "Epoch 177/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.1627e-05\n",
            "Epoch 178/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 1.0858e-05\n",
            "Epoch 179/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.0141e-05\n",
            "Epoch 180/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 9.4705e-06\n",
            "Epoch 181/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 8.8445e-06\n",
            "Epoch 182/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 8.2599e-06\n",
            "Epoch 183/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 7.7143e-06\n",
            "Epoch 184/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 7.2041e-06\n",
            "Epoch 185/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 6.7281e-06\n",
            "Epoch 186/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 6.2829e-06\n",
            "Epoch 187/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 5.8674e-06\n",
            "Epoch 188/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.4799e-06\n",
            "Epoch 189/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.1173e-06\n",
            "Epoch 190/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 4.7789e-06\n",
            "Epoch 191/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 4.4628e-06\n",
            "Epoch 192/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 4.1679e-06\n",
            "Epoch 193/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 3.8925e-06\n",
            "Epoch 194/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.6348e-06\n",
            "Epoch 195/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.3949e-06\n",
            "Epoch 196/500\n",
            "1/1 [==============================] - 0s 25ms/step - loss: 3.1702e-06\n",
            "Epoch 197/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.9605e-06\n",
            "Epoch 198/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.7649e-06\n",
            "Epoch 199/500\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 2.5822e-06\n",
            "Epoch 200/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.4111e-06\n",
            "Epoch 201/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.2516e-06\n",
            "Epoch 202/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.1026e-06\n",
            "Epoch 203/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.9638e-06\n",
            "Epoch 204/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.8338e-06\n",
            "Epoch 205/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.7124e-06\n",
            "Epoch 206/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.5995e-06\n",
            "Epoch 207/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.4935e-06\n",
            "Epoch 208/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.3946e-06\n",
            "Epoch 209/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.3024e-06\n",
            "Epoch 210/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.2164e-06\n",
            "Epoch 211/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.1359e-06\n",
            "Epoch 212/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.0608e-06\n",
            "Epoch 213/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 9.9059e-07\n",
            "Epoch 214/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 9.2507e-07\n",
            "Epoch 215/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 8.6379e-07\n",
            "Epoch 216/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 8.0669e-07\n",
            "Epoch 217/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 7.5330e-07\n",
            "Epoch 218/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 7.0356e-07\n",
            "Epoch 219/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 6.5701e-07\n",
            "Epoch 220/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 6.1348e-07\n",
            "Epoch 221/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 5.7297e-07\n",
            "Epoch 222/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 5.3505e-07\n",
            "Epoch 223/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.9954e-07\n",
            "Epoch 224/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 4.6662e-07\n",
            "Epoch 225/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.3579e-07\n",
            "Epoch 226/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 4.0691e-07\n",
            "Epoch 227/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.7991e-07\n",
            "Epoch 228/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.5483e-07\n",
            "Epoch 229/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.3134e-07\n",
            "Epoch 230/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 3.0944e-07\n",
            "Epoch 231/500\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 2.8895e-07\n",
            "Epoch 232/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.6989e-07\n",
            "Epoch 233/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.5213e-07\n",
            "Epoch 234/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.3533e-07\n",
            "Epoch 235/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.1975e-07\n",
            "Epoch 236/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.0524e-07\n",
            "Epoch 237/500\n",
            "1/1 [==============================] - 0s 12ms/step - loss: 1.9176e-07\n",
            "Epoch 238/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 1.7906e-07\n",
            "Epoch 239/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.6717e-07\n",
            "Epoch 240/500\n",
            "1/1 [==============================] - 0s 25ms/step - loss: 1.5610e-07\n",
            "Epoch 241/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.4578e-07\n",
            "Epoch 242/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 1.3612e-07\n",
            "Epoch 243/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.2710e-07\n",
            "Epoch 244/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 1.1873e-07\n",
            "Epoch 245/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.1085e-07\n",
            "Epoch 246/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 1.0353e-07\n",
            "Epoch 247/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 9.6682e-08\n",
            "Epoch 248/500\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 9.0313e-08\n",
            "Epoch 249/500\n",
            "1/1 [==============================] - 0s 22ms/step - loss: 8.4323e-08\n",
            "Epoch 250/500\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 7.8683e-08\n",
            "Epoch 251/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 7.3520e-08\n",
            "Epoch 252/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 6.8663e-08\n",
            "Epoch 253/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 6.4089e-08\n",
            "Epoch 254/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.9895e-08\n",
            "Epoch 255/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.5955e-08\n",
            "Epoch 256/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 5.2225e-08\n",
            "Epoch 257/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 4.8820e-08\n",
            "Epoch 258/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 4.5534e-08\n",
            "Epoch 259/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 4.2539e-08\n",
            "Epoch 260/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.9713e-08\n",
            "Epoch 261/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 3.7059e-08\n",
            "Epoch 262/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 3.4641e-08\n",
            "Epoch 263/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 3.2356e-08\n",
            "Epoch 264/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.0205e-08\n",
            "Epoch 265/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.8220e-08\n",
            "Epoch 266/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.6341e-08\n",
            "Epoch 267/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.4588e-08\n",
            "Epoch 268/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 2.2966e-08\n",
            "Epoch 269/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.1464e-08\n",
            "Epoch 270/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.0038e-08\n",
            "Epoch 271/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 1.8709e-08\n",
            "Epoch 272/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 1.7478e-08\n",
            "Epoch 273/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.6296e-08\n",
            "Epoch 274/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.5221e-08\n",
            "Epoch 275/500\n",
            "1/1 [==============================] - 0s 14ms/step - loss: 1.4246e-08\n",
            "Epoch 276/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.3274e-08\n",
            "Epoch 277/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.2407e-08\n",
            "Epoch 278/500\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 1.1592e-08\n",
            "Epoch 279/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 1.0807e-08\n",
            "Epoch 280/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 1.0106e-08\n",
            "Epoch 281/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 9.4422e-09\n",
            "Epoch 282/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 8.8220e-09\n",
            "Epoch 283/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 8.2309e-09\n",
            "Epoch 284/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 7.6752e-09\n",
            "Epoch 285/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 7.1651e-09\n",
            "Epoch 286/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 6.6898e-09\n",
            "Epoch 287/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 6.2537e-09\n",
            "Epoch 288/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 5.8289e-09\n",
            "Epoch 289/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.4626e-09\n",
            "Epoch 290/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 5.0942e-09\n",
            "Epoch 291/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 4.7557e-09\n",
            "Epoch 292/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 4.4343e-09\n",
            "Epoch 293/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 4.1468e-09\n",
            "Epoch 294/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.8849e-09\n",
            "Epoch 295/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.6185e-09\n",
            "Epoch 296/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.3933e-09\n",
            "Epoch 297/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 3.1663e-09\n",
            "Epoch 298/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.9533e-09\n",
            "Epoch 299/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.7549e-09\n",
            "Epoch 300/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.5714e-09\n",
            "Epoch 301/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4097e-09\n",
            "Epoch 302/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.2460e-09\n",
            "Epoch 303/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.0944e-09\n",
            "Epoch 304/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.9605e-09\n",
            "Epoch 305/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.8266e-09\n",
            "Epoch 306/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.7116e-09\n",
            "Epoch 307/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.6005e-09\n",
            "Epoch 308/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.4901e-09\n",
            "Epoch 309/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.3910e-09\n",
            "Epoch 310/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.3044e-09\n",
            "Epoch 311/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.2179e-09\n",
            "Epoch 312/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.1361e-09\n",
            "Epoch 313/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.0545e-09\n",
            "Epoch 314/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 9.9004e-10\n",
            "Epoch 315/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 9.1928e-10\n",
            "Epoch 316/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 8.5871e-10\n",
            "Epoch 317/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 8.0300e-10\n",
            "Epoch 318/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 7.4924e-10\n",
            "Epoch 319/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 6.9829e-10\n",
            "Epoch 320/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 6.5027e-10\n",
            "Epoch 321/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 6.1551e-10\n",
            "Epoch 322/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 5.6723e-10\n",
            "Epoch 323/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.3426e-10\n",
            "Epoch 324/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 4.9923e-10\n",
            "Epoch 325/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 4.6618e-10\n",
            "Epoch 326/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 4.3649e-10\n",
            "Epoch 327/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 4.1203e-10\n",
            "Epoch 328/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.8136e-10\n",
            "Epoch 329/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.5619e-10\n",
            "Epoch 330/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 3.3590e-10\n",
            "Epoch 331/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.1067e-10\n",
            "Epoch 332/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 2.9350e-10\n",
            "Epoch 333/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.7206e-10\n",
            "Epoch 334/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.5490e-10\n",
            "Epoch 335/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.3700e-10\n",
            "Epoch 336/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 2.2314e-10\n",
            "Epoch 337/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.0828e-10\n",
            "Epoch 338/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.9369e-10\n",
            "Epoch 339/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.8230e-10\n",
            "Epoch 340/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.7046e-10\n",
            "Epoch 341/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.5756e-10\n",
            "Epoch 342/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.4686e-10\n",
            "Epoch 343/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.3725e-10\n",
            "Epoch 344/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.2531e-10\n",
            "Epoch 345/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.1742e-10\n",
            "Epoch 346/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.1161e-10\n",
            "Epoch 347/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 1.0328e-10\n",
            "Epoch 348/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 9.7979e-11\n",
            "Epoch 349/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 9.0914e-11\n",
            "Epoch 350/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 8.3821e-11\n",
            "Epoch 351/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 7.9595e-11\n",
            "Epoch 352/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 7.1997e-11\n",
            "Epoch 353/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 6.8811e-11\n",
            "Epoch 354/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 6.3309e-11\n",
            "Epoch 355/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 6.0181e-11\n",
            "Epoch 356/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.5991e-11\n",
            "Epoch 357/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 5.3104e-11\n",
            "Epoch 358/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.8184e-11\n",
            "Epoch 359/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 4.5363e-11\n",
            "Epoch 360/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.2085e-11\n",
            "Epoch 361/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.9864e-11\n",
            "Epoch 362/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.8535e-11\n",
            "Epoch 363/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.5589e-11\n",
            "Epoch 364/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 3.3007e-11\n",
            "Epoch 365/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 3.0650e-11\n",
            "Epoch 366/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.9066e-11\n",
            "Epoch 367/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.8062e-11\n",
            "Epoch 368/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.5134e-11\n",
            "Epoch 369/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.3524e-11\n",
            "Epoch 370/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.2141e-11\n",
            "Epoch 371/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.0753e-11\n",
            "Epoch 372/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.9069e-11\n",
            "Epoch 373/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 1.8569e-11\n",
            "Epoch 374/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 1.8328e-11\n",
            "Epoch 375/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.6163e-11\n",
            "Epoch 376/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.5254e-11\n",
            "Epoch 377/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.4448e-11\n",
            "Epoch 378/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.3730e-11\n",
            "Epoch 379/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 1.2686e-11\n",
            "Epoch 380/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.2115e-11\n",
            "Epoch 381/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.1459e-11\n",
            "Epoch 382/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.0462e-11\n",
            "Epoch 383/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 1.0009e-11\n",
            "Epoch 384/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 1.0222e-11\n",
            "Epoch 385/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 9.8013e-12\n",
            "Epoch 386/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 8.9363e-12\n",
            "Epoch 387/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 8.9274e-12\n",
            "Epoch 388/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 7.5910e-12\n",
            "Epoch 389/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 7.4465e-12\n",
            "Epoch 390/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 7.4796e-12\n",
            "Epoch 391/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 6.6892e-12\n",
            "Epoch 392/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 6.5394e-12\n",
            "Epoch 393/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 6.1794e-12\n",
            "Epoch 394/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 6.0917e-12\n",
            "Epoch 395/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 5.5949e-12\n",
            "Epoch 396/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 4.8175e-12\n",
            "Epoch 397/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.9359e-12\n",
            "Epoch 398/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 4.7867e-12\n",
            "Epoch 399/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.4391e-12\n",
            "Epoch 400/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 4.2585e-12\n",
            "Epoch 401/500\n",
            "1/1 [==============================] - 0s 10ms/step - loss: 3.6575e-12\n",
            "Epoch 402/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 3.5622e-12\n",
            "Epoch 403/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.4153e-12\n",
            "Epoch 404/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 3.1311e-12\n",
            "Epoch 405/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.9938e-12\n",
            "Epoch 406/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.9938e-12\n",
            "Epoch 407/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.6645e-12\n",
            "Epoch 408/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.5864e-12\n",
            "Epoch 409/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.5580e-12\n",
            "Epoch 410/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.4774e-12\n",
            "Epoch 411/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.4774e-12\n",
            "Epoch 412/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.4774e-12\n",
            "Epoch 413/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.4774e-12\n",
            "Epoch 414/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 415/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 416/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.4774e-12\n",
            "Epoch 417/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 2.4774e-12\n",
            "Epoch 418/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.4774e-12\n",
            "Epoch 419/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 420/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 421/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 422/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 423/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 424/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 425/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 426/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 2.4774e-12\n",
            "Epoch 427/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.4774e-12\n",
            "Epoch 428/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 2.4774e-12\n",
            "Epoch 429/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 430/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 2.4774e-12\n",
            "Epoch 431/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 432/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 2.4774e-12\n",
            "Epoch 433/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 434/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.4774e-12\n",
            "Epoch 435/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 436/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 437/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 2.4774e-12\n",
            "Epoch 438/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.4774e-12\n",
            "Epoch 439/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 440/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 441/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.4774e-12\n",
            "Epoch 442/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 443/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.4774e-12\n",
            "Epoch 444/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.4774e-12\n",
            "Epoch 445/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 446/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.4774e-12\n",
            "Epoch 447/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 448/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.4774e-12\n",
            "Epoch 449/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.4774e-12\n",
            "Epoch 450/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 451/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 2.4774e-12\n",
            "Epoch 452/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.4774e-12\n",
            "Epoch 453/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 454/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 455/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.4774e-12\n",
            "Epoch 456/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.4774e-12\n",
            "Epoch 457/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 458/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.4774e-12\n",
            "Epoch 459/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 2.4774e-12\n",
            "Epoch 460/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 461/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 462/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 463/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 464/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 465/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 466/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 467/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.4774e-12\n",
            "Epoch 468/500\n",
            "1/1 [==============================] - 0s 2ms/step - loss: 2.4774e-12\n",
            "Epoch 469/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 470/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.4774e-12\n",
            "Epoch 471/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 472/500\n",
            "1/1 [==============================] - 0s 11ms/step - loss: 2.4774e-12\n",
            "Epoch 473/500\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 2.4774e-12\n",
            "Epoch 474/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 475/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 2.4774e-12\n",
            "Epoch 476/500\n",
            "1/1 [==============================] - 0s 13ms/step - loss: 2.4774e-12\n",
            "Epoch 477/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 478/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.4774e-12\n",
            "Epoch 479/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 480/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.4774e-12\n",
            "Epoch 481/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.4774e-12\n",
            "Epoch 482/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.4774e-12\n",
            "Epoch 483/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 2.4774e-12\n",
            "Epoch 484/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.4774e-12\n",
            "Epoch 485/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.4774e-12\n",
            "Epoch 486/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.4774e-12\n",
            "Epoch 487/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.4774e-12\n",
            "Epoch 488/500\n",
            "1/1 [==============================] - 0s 8ms/step - loss: 2.4774e-12\n",
            "Epoch 489/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 490/500\n",
            "1/1 [==============================] - 0s 4ms/step - loss: 2.4774e-12\n",
            "Epoch 491/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.4774e-12\n",
            "Epoch 492/500\n",
            "1/1 [==============================] - 0s 3ms/step - loss: 2.4774e-12\n",
            "Epoch 493/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 494/500\n",
            "1/1 [==============================] - 0s 6ms/step - loss: 2.4774e-12\n",
            "Epoch 495/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.4774e-12\n",
            "Epoch 496/500\n",
            "1/1 [==============================] - 0s 5ms/step - loss: 2.4774e-12\n",
            "Epoch 497/500\n",
            "1/1 [==============================] - 0s 9ms/step - loss: 2.4774e-12\n",
            "Epoch 498/500\n",
            "1/1 [==============================] - 0s 22ms/step - loss: 2.4774e-12\n",
            "Epoch 499/500\n",
            "1/1 [==============================] - 0s 19ms/step - loss: 2.4774e-12\n",
            "Epoch 500/500\n",
            "1/1 [==============================] - 0s 7ms/step - loss: 2.4774e-12\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f05d7159450>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qPFJXCebv-Oi",
        "outputId": "3467eca7-3af0-4beb-b637-faf7fe8f97da"
      },
      "source": [
        "print(my_layer_1.get_weights())"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[array([[1.2162362 , 0.88762665]], dtype=float32), array([-0.37960783, -0.28105885], dtype=float32)]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mgeHHVKowLrS",
        "outputId": "2e59eb7f-fb6d-4a81-8f4b-7779536d7f1b"
      },
      "source": [
        "print(my_layer_2.get_weights())"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[array([[1.0675431],\n",
            "       [0.7904388]], dtype=float32), array([-0.37259063], dtype=float32)]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zgPETauzw7Ao",
        "outputId": "ec89ac68-ddb9-4d0e-8942-8f1b98e8b24c"
      },
      "source": [
        "value_to_predict = 10.0\n",
        "\n",
        "layer1_w1 = (my_layer_1.get_weights()[0][0][0])\n",
        "\n",
        "layer1_w2 = (my_layer_1.get_weights()[0][0][1])\n",
        "\n",
        "layer1_b1 = (my_layer_1.get_weights()[1][0])\n",
        "\n",
        "layer1_b2 = (my_layer_1.get_weights()[1][1])\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "layer2_w1 = (my_layer_2.get_weights()[0][0])\n",
        "\n",
        "layer2_w2 = (my_layer_2.get_weights()[0][1])\n",
        "\n",
        "layer2_b = (my_layer_2.get_weights()[1][0])\n",
        "\n",
        " \n",
        "\n",
        "neuron1_output = (layer1_w1 * value_to_predict) + layer1_b1\n",
        "\n",
        "neuron2_output = (layer1_w2 * value_to_predict) + layer1_b2\n",
        "\n",
        " \n",
        "\n",
        "neuron3_output = (layer2_w1 * neuron1_output) + (layer2_w2 * neuron2_output) + layer2_b\n",
        "\n",
        " \n",
        "print(f'Para x=10, nosso novo modelo retorna y={neuron3_output}')"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Para x=10, nosso novo modelo retorna y=[18.999992]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4Pfa7zBCyzfn",
        "outputId": "e4aff863-a57e-42a0-8048-dd7d5bf75dd3"
      },
      "source": [
        "print(model.summary())"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_1 (Dense)              (None, 2)                 4         \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 1)                 3         \n",
            "=================================================================\n",
            "Total params: 7\n",
            "Trainable params: 7\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}