{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 28,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "gfxHFbMm-gBz",
        "outputId": "b5044086-2cd2-4783-8693-bdfab2cf7177"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 1.2253\n",
            "Epoch 2/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 1.0474\n",
            "Epoch 3/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.9757\n",
            "Epoch 4/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.9206\n",
            "Epoch 5/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.8931\n",
            "Epoch 6/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.8738\n",
            "Epoch 7/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.8561\n",
            "Epoch 8/50\n",
            "5/5 [==============================] - 0s 2ms/step - loss: 0.8402\n",
            "Epoch 9/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.8245\n",
            "Epoch 10/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.8095\n",
            "Epoch 11/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.7947\n",
            "Epoch 12/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.7765\n",
            "Epoch 13/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.7634\n",
            "Epoch 14/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.7473\n",
            "Epoch 15/50\n",
            "5/5 [==============================] - 0s 2ms/step - loss: 0.7307\n",
            "Epoch 16/50\n",
            "5/5 [==============================] - 0s 4ms/step - loss: 0.7162\n",
            "Epoch 17/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.7033\n",
            "Epoch 18/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.6876\n",
            "Epoch 19/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.6731\n",
            "Epoch 20/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.6601\n",
            "Epoch 21/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.6497\n",
            "Epoch 22/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.6377\n",
            "Epoch 23/50\n",
            "5/5 [==============================] - 0s 2ms/step - loss: 0.6232\n",
            "Epoch 24/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.6111\n",
            "Epoch 25/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.5984\n",
            "Epoch 26/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.5885\n",
            "Epoch 27/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.5786\n",
            "Epoch 28/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.5676\n",
            "Epoch 29/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.5610\n",
            "Epoch 30/50\n",
            "5/5 [==============================] - 0s 2ms/step - loss: 0.5516\n",
            "Epoch 31/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.5405\n",
            "Epoch 32/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.5358\n",
            "Epoch 33/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.5250\n",
            "Epoch 34/50\n",
            "5/5 [==============================] - 0s 4ms/step - loss: 0.5159\n",
            "Epoch 35/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.5110\n",
            "Epoch 36/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.5017\n",
            "Epoch 37/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4964\n",
            "Epoch 38/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4906\n",
            "Epoch 39/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4852\n",
            "Epoch 40/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4744\n",
            "Epoch 41/50\n",
            "5/5 [==============================] - 0s 5ms/step - loss: 0.4691\n",
            "Epoch 42/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4607\n",
            "Epoch 43/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4560\n",
            "Epoch 44/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4504\n",
            "Epoch 45/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4443\n",
            "Epoch 46/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4421\n",
            "Epoch 47/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4346\n",
            "Epoch 48/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4297\n",
            "Epoch 49/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4222\n",
            "Epoch 50/50\n",
            "5/5 [==============================] - 0s 3ms/step - loss: 0.4172\n",
            "5/5 [==============================] - 0s 2ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7c17f5ffbf70>]"
            ]
          },
          "metadata": {},
          "execution_count": 28
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "from tensorflow.keras import Sequential\n",
        "from tensorflow.keras.layers import Dense\n",
        "\n",
        "from sklearn.datasets import load_iris\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "\n",
        "\n",
        "data = load_iris()\n",
        "x_train, y_train = data.data, data.target\n",
        "\n",
        "model = Sequential()\n",
        "model.add(Dense(32,activation='relu', input_dim= 4))\n",
        "model.add(Dense(8, activation='relu'))\n",
        "model.add(Dense(3, activation = 'softmax'))\n",
        "\n",
        "model.compile(loss=\"sparse_categorical_crossentropy\", optimizer='SGD')\n",
        "\n",
        "history = model.fit(x_train,y_train, epochs=50)\n",
        "y_hat = model.predict(x_train)\n",
        "classes = y_hat.argmax(axis=1)\n",
        "acc = sum(y_train == classes)/len(x_train)\n",
        "\n",
        "loss = history.history['loss']\n",
        "os_x = np.arange(50)+1\n",
        "plt.plot(os_x, loss)\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "dtyZaHKkNPhd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "tqxyYzPwMKQq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "iC5tAMk4MTjs"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [],
      "metadata": {
        "id": "7bOY3GyZ-s1L"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "LQwJElLlL2vD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "cQki26wGL4Hu"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}