Image_Classification_Custom_model.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Image_Classification_Custom_model.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyOpMBUSVBsjDRGFYvrTiLp/",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/KevinTheRainmaker/AI_Tutorials/blob/main/Image_Classification_Custom_model.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "FgjVxGuGCN3X"
      },
      "outputs": [],
      "source": [
        "# Install Model maker\n",
        "!pip install -q tflite-model-maker &> /dev/null"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Imports and check that we are using TF2.x\n",
        "import numpy as np\n",
        "import os\n",
        "\n",
        "from tflite_model_maker import configs\n",
        "from tflite_model_maker import ExportFormat\n",
        "from tflite_model_maker import model_spec\n",
        "from tflite_model_maker import image_classifier\n",
        "from tflite_model_maker.image_classifier import DataLoader\n",
        "\n",
        "import tensorflow as tf\n",
        "assert tf.__version__.startswith('2')\n",
        "tf.get_logger().setLevel('ERROR')"
      ],
      "metadata": {
        "id": "7yPR3jXSCZRx"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Data Load"
      ],
      "metadata": {
        "id": "ZFfBSeXBC38q"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "data_path = tf.keras.utils.get_file(\n",
        "      'flower_photos',\n",
        "      'https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz',\n",
        "      untar=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "y5E8TipAC1__",
        "outputId": "03fb358b-7d19-42f9-fec5-a8bc73849fae"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz\n",
            "228818944/228813984 [==============================] - 2s 0us/step\n",
            "228827136/228813984 [==============================] - 2s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "data = DataLoader.from_folder(data_path)\n",
        "train_data, test_data = data.split(0.9)"
      ],
      "metadata": {
        "id": "Id_0kg4TC5Jn"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Model"
      ],
      "metadata": {
        "id": "Wany9vCYEPuV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model = image_classifier.create(train_data)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1m75qGFYDALP",
        "outputId": "0f63c0af-1362-4e50-982c-992904a23ff3"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " hub_keras_layer_v1v2 (HubKe  (None, 1280)             3413024   \n",
            " rasLayerV1V2)                                                   \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 1280)              0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 5)                 6405      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 3,419,429\n",
            "Trainable params: 6,405\n",
            "Non-trainable params: 3,413,024\n",
            "_________________________________________________________________\n",
            "None\n",
            "Epoch 1/5\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/optimizer_v2/gradient_descent.py:102: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
            "  super(SGD, self).__init__(name, **kwargs)\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "103/103 [==============================] - 125s 1s/step - loss: 0.8550 - accuracy: 0.7843\n",
            "Epoch 2/5\n",
            "103/103 [==============================] - 118s 1s/step - loss: 0.6588 - accuracy: 0.9005\n",
            "Epoch 3/5\n",
            "103/103 [==============================] - 118s 1s/step - loss: 0.6198 - accuracy: 0.9154\n",
            "Epoch 4/5\n",
            "103/103 [==============================] - 118s 1s/step - loss: 0.6043 - accuracy: 0.9245\n",
            "Epoch 5/5\n",
            "103/103 [==============================] - 119s 1s/step - loss: 0.5894 - accuracy: 0.9339\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Testing"
      ],
      "metadata": {
        "id": "QsimJgRqEasG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "loss, accuracy = model.evaluate(test_data)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "O8wHaRvQEPMr",
        "outputId": "9fa9c672-4c6b-4df9-e3d0-7e5e5b17d891"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "12/12 [==============================] - 18s 1s/step - loss: 0.5939 - accuracy: 0.9264\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Export model"
      ],
      "metadata": {
        "id": "yb0vC3TGEkRG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.export(export_dir='/mm_flowers')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xIhmLJYdEcGo",
        "outputId": "919b7446-b24a-4d83-befb-a72c3a36d872"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/lite/python/convert.py:746: UserWarning: Statistics for quantized inputs were expected, but not specified; continuing anyway.\n",
            "  warnings.warn(\"Statistics for quantized inputs were expected, but not \"\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "kSogRn1rEmN-"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}