One_Shot_Learning_1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "One-Shot-Learning-1.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.4"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "bA1XMN2df42k"
      },
      "source": [
        "# Few Shot Learning\n",
        "\n",
        "---\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "tRLFL09MRb8H"
      },
      "source": [
        "# Omniglot data set for one-shot learning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "U4dwjHmSSEKu"
      },
      "source": [
        "The Omniglot data set is designed for developing more human-like learning algorithms. It contains 1623 different handwritten characters from 50 different alphabets. Each of the 1623 characters was drawn online via Amazon's Mechanical Turk by 20 different people. Each image is paired with stroke data, a sequences of [x,y,t] coordinates with time (t) in milliseconds.\n",
        "\n",
        "### Contents\n",
        "The Omniglot data set contains 50 alphabets. We split these into a background set of 30 alphabets and an evaluation set of 20 alphabets. To compare with the results with the paper, only the background set should be used to learn general knowledge about characters (e.g., feature learning, meta-learning, or hyperparameter inference). One-shot learning results are reported using alphabets from the evaluation set.\n",
        "\n",
        "Two more challenging \"minimal\" splits contain only five background alphabets, denoted as \"background small 1\" and \"background small 2\". This is a closer approximation to the experience that a human adult might have for characters in general. For the goal of building human-level AI systems with minimal training, given a rough estimate of what \"minimal\" means for people, there is a need to explore settings with fewer training examples per class and fewer background classes for learning to learn.\n",
        "\n",
        "- Link to paper: https://science.sciencemag.org/content/350/6266/1332\n",
        "- Link to dataset: https://github.com/brendenlake/omniglot"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "ZOk8Eu4_t70R"
      },
      "source": [
        "Firstly, let's select TensorFlow version 2.x in colab"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "H6RZUm0p4wYJ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "3c4f6996-6ed3-4c44-eaea-f34572ab1531"
      },
      "source": [
        "%tensorflow_version 2.x\n",
        "import tensorflow\n",
        "tensorflow.__version__"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'2.3.0'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "TWi96z-8SyX0",
        "colab": {}
      },
      "source": [
        "# Initialize the random number generator\n",
        "import random\n",
        "random.seed(0)\n",
        "\n",
        "# Ignore the warnings\n",
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "OKVkkLaLdPGU"
      },
      "source": [
        "### Load the dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "JK0ZRJoxhrFe"
      },
      "source": [
        "As we are using google colab, we need to mount the google drive to load the data file"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "dv6RCAcGLd4w",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 124
        },
        "outputId": "3a4cb429-0c18-47fb-dba4-38fc987b8654"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive/')"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive/\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "3gqIEcuK-gqQ"
      },
      "source": [
        "Add path to the folder where your dataset files are"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "fzXrTG6fLqw4",
        "colab": {}
      },
      "source": [
        "project_path = '/content/drive/My Drive/Colab Notebooks/databases/datasets-master/Omniglot Dataset/'"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "1FOfBTnoBjYi"
      },
      "source": [
        "Let's load the dataset now"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "PrXGB_RORiXN",
        "colab": {}
      },
      "source": [
        "training_zip_path = project_path + 'images_background-2.zip'\n",
        "validation_zip_path = project_path + 'images_evaluation-2.zip'"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "2uX9VNVfR2p3"
      },
      "source": [
        "Extract the content of the zip files"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "qFiZc7Rbw7mm",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "690a89ac-d75a-427b-c1d0-878ddaca3c87"
      },
      "source": [
        "from zipfile import ZipFile\n",
        "\n",
        "with ZipFile(training_zip_path, 'r') as z:\n",
        "  z.extractall()\n",
        "print(\"Training zip extraction done!\")\n",
        "\n",
        "with ZipFile(validation_zip_path, 'r') as z:\n",
        "  z.extractall()\n",
        "print(\"Validation zip extraction done!\")"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training zip extraction done!\n",
            "Validation zip extraction done!\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "X6uE4alZbecJ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 320
        },
        "outputId": "ff6a04dc-e0f8-4647-afc5-907a5dfae6a3"
      },
      "source": [
        "import cv2\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "\n",
        "img = cv2.imread('images_background/Sanskrit/character11/0861_05.png')\n",
        "print(\"Each image in the data set has a same of {0}\".format(img.shape))\n",
        "flattened_img = img.flatten()\n",
        "print(\"The number of features in any image from the data set are: {0}\".format(flattened_img.shape[0]))\n",
        "plt.imshow(img)"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Each image in the data set has a same of (105, 105, 3)\n",
            "The number of features in any image from the data set are: 33075\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f944551d710>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        },
        {
          "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": "code",
      "metadata": {
        "colab_type": "code",
        "id": "iP2J9zEyfS8e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 320
        },
        "outputId": "ea449e44-1c33-4b8c-ed89-0a4a947a2856"
      },
      "source": [
        "img = cv2.imread('images_evaluation/Kannada/character11/1215_05.png')\n",
        "print(\"Each image in the data set has a same of {0}\".format(img.shape))\n",
        "flattened_img = img.flatten()\n",
        "print(\"The number of features in any image from the data set are: {0}\".format(flattened_img.shape[0]))\n",
        "plt.imshow(img)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Each image in the data set has a same of (105, 105, 3)\n",
            "The number of features in any image from the data set are: 33075\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f94469facf8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        },
        {
          "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": "code",
      "metadata": {
        "colab_type": "code",
        "id": "aWmlgm76becL",
        "colab": {}
      },
      "source": [
        "base_class_name = 'character'\n",
        "\n",
        "def gen_class_names(base_class_name):\n",
        "    classes = []\n",
        "    for i in range(1,21):\n",
        "        if i < 10:\n",
        "            classes.append(\"{0}0{1}\".format(base_class_name, i))\n",
        "        else:\n",
        "            classes.append(\"{0}{1}\".format(base_class_name, i))\n",
        "    return classes"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "lW--qMTZbecN",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 55
        },
        "outputId": "9934c6a9-ee7b-4795-abb2-076b750678d7"
      },
      "source": [
        "classes = gen_class_names(base_class_name)\n",
        "print(classes)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['character01', 'character02', 'character03', 'character04', 'character05', 'character06', 'character07', 'character08', 'character09', 'character10', 'character11', 'character12', 'character13', 'character14', 'character15', 'character16', 'character17', 'character18', 'character19', 'character20']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QmlKkFtYh9yc",
        "colab_type": "text"
      },
      "source": [
        ""
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Egn9OIADbecP",
        "colab": {}
      },
      "source": [
        "from sklearn.preprocessing import LabelBinarizer\n",
        "\n",
        "def generate_one_hot_encoding(classes):\n",
        "    encoder = LabelBinarizer()\n",
        "    transfomed_labels = encoder.fit_transform(classes)\n",
        "    return transfomed_labels"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "4OGiK8uKbecR",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 381
        },
        "outputId": "75ffe87b-a09c-41a3-fe39-df82c3ef72ae"
      },
      "source": [
        "labels = generate_one_hot_encoding(classes)\n",
        "print(labels)\n",
        "len(labels)"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n",
            " [0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n",
            " [0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n",
            " [0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n",
            " [0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n",
            " [0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n",
            " [0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0]\n",
            " [0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0]\n",
            " [0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0]\n",
            " [0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0]\n",
            " [0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0]\n",
            " [0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0]\n",
            " [0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0]\n",
            " [0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0]\n",
            " [0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0]\n",
            " [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0]\n",
            " [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0]\n",
            " [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0]\n",
            " [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0]\n",
            " [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "20"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "1lQgkL9xF-oy"
      },
      "source": [
        "### Let's define our model now\n",
        "\n",
        "(https://drive.google.com/file/d/1PH50n205Bkv9yfoeo0564htgPLVynElI/view?usp=sharing)\n",
        "\n",
        "https://drive.google.com/file/d/15w-vpbOrc-rm-Xsg3kGqYUAq_eWmEmq5/view?usp=sharing"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "RWNSn9Ipbecd",
        "colab": {}
      },
      "source": [
        "from tensorflow.keras.models import Sequential, Model\n",
        "from tensorflow.keras.layers import Input, Conv2D, MaxPooling2D, Flatten, Dense, Lambda\n",
        "from tensorflow.keras import backend as K\n",
        "\n",
        "K.set_image_data_format('channels_last')\n",
        "\n",
        "def get_siamese_model(input_shape):\n",
        "    \"\"\"\n",
        "        Model architecture based on the one provided in: http://www.cs.utoronto.ca/~gkoch/files/msc-thesis.pdf\n",
        "    \"\"\"\n",
        "    left_input = Input(input_shape)\n",
        "    right_input = Input(input_shape)\n",
        "    model = Sequential()\n",
        "    model.add(Conv2D(64, (10,10), activation='relu', input_shape=input_shape))\n",
        "    model.add(MaxPooling2D())\n",
        "    model.add(Conv2D(128, (7,7), activation='relu'))\n",
        "    model.add(MaxPooling2D())\n",
        "    model.add(Conv2D(128, (4,4), activation='relu'))\n",
        "    model.add(MaxPooling2D())\n",
        "    model.add(Conv2D(256, (4,4), activation='relu'))\n",
        "    model.add(Flatten())\n",
        "    model.add(Dense(4096, activation='sigmoid'))\n",
        "    encoded_l = model(left_input)\n",
        "    encoded_r = model(right_input)\n",
        "    L1_layer = Lambda(lambda tensors:K.abs(tensors[0] - tensors[1]))\n",
        "    L1_distance = L1_layer([encoded_l, encoded_r])\n",
        "    prediction = Dense(1,activation='sigmoid')(L1_distance)\n",
        "    siamese_net = Model(inputs=[left_input,right_input],outputs=prediction)\n",
        "\n",
        "    return siamese_net"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "jRNNv1MIGQzs"
      },
      "source": [
        "### Create the model and print summary"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "ioLXDgqHbecg",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 364
        },
        "outputId": "99533dc3-c1d1-4f5a-a8a3-56e369bea588"
      },
      "source": [
        "model = get_siamese_model((105, 105, 1))\n",
        "model.summary()"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"functional_1\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_1 (InputLayer)            [(None, 105, 105, 1) 0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_2 (InputLayer)            [(None, 105, 105, 1) 0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 4096)         38947648    input_1[0][0]                    \n",
            "                                                                 input_2[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "lambda (Lambda)                 (None, 4096)         0           sequential[0][0]                 \n",
            "                                                                 sequential[1][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "dense_1 (Dense)                 (None, 1)            4097        lambda[0][0]                     \n",
            "==================================================================================================\n",
            "Total params: 38,951,745\n",
            "Trainable params: 38,951,745\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "0aeyqhjeGd_s"
      },
      "source": [
        "### Compile the model\n",
        "- Define optimizer\n",
        "- Define loss"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "AYLKoskFbeci",
        "colab": {}
      },
      "source": [
        "from tensorflow.keras.optimizers import Adam\n",
        "\n",
        "optimizer = Adam(lr = 0.00006)\n",
        "model.compile(loss=\"binary_crossentropy\",optimizer=optimizer)"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "JrQQFJl9AmME"
      },
      "source": [
        "#### Plot all 20 samples of a particular character of a language"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "ET1bZQsGbecm",
        "colab": {}
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import os\n",
        "\n",
        "def plot_images(path):\n",
        "    \"\"\"\n",
        "        Plot all 20 samples of a particular character of a language\n",
        "    \"\"\"\n",
        "    f, axarr = plt.subplots(5,4, figsize=(10,10))\n",
        "    images_list = []\n",
        "    for image in os.listdir(path):\n",
        "        image_path = os.path.join(path, image)\n",
        "        img = cv2.imread(image_path)\n",
        "        images_list.append(img)\n",
        "    for i in range(5):\n",
        "        for j in range(4):\n",
        "            axarr[i,j].imshow(images_list.pop())"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "dJKVKZspbeco",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 613
        },
        "outputId": "0fde8689-4af4-4890-d8a3-85e2e4447750"
      },
      "source": [
        "plot_images('images_background/Sanskrit/character02/')\n",
        "print(\"Sanskrit language, 20 samples of the third character.\")"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Sanskrit language, 20 samples of the third character.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 20 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "cEQ8oJmWAp5Q"
      },
      "source": [
        "Let's take one other language"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "_Msr3Q_cbecq",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 613
        },
        "outputId": "3c2de523-229b-4601-a178-197fc369e9c9"
      },
      "source": [
        "plot_images('images_background/Bengali/character03/')\n",
        "print(\"Bengali language, 20 samples of the seventh character.\")"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Bengali language, 20 samples of the seventh character.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 20 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "aKAQ7bEoA52a"
      },
      "source": [
        "### Load the training and validation pickles"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "URFz4aLebecw",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 159
        },
        "outputId": "51d55711-9cc7-42e0-e5d4-f178f38a3dd4"
      },
      "source": [
        "import pickle\n",
        "\n",
        "with open(os.path.join(project_path, \"train.pickle\"), \"rb\") as f:\n",
        "    (X, classes) = pickle.load(f)\n",
        "\n",
        "with open(os.path.join(project_path, \"val.pickle\"), \"rb\") as f:\n",
        "    (Xval, val_classes) = pickle.load(f)\n",
        "    \n",
        "print(\"Training alphabets: \\n\")\n",
        "print(len(list(classes.keys())))\n",
        "print(\"Validation alphabets:\", end=\"\\n\\n\")\n",
        "print(list(val_classes.keys()))\n",
        "print(X.shape)"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training alphabets: \n",
            "\n",
            "30\n",
            "Validation alphabets:\n",
            "\n",
            "['Angelic', 'Atemayar_Qelisayer', 'Atlantean', 'Aurek-Besh', 'Avesta', 'Ge_ez', 'Glagolitic', 'Gurmukhi', 'Kannada', 'Keble', 'Malayalam', 'Manipuri', 'Mongolian', 'Old_Church_Slavonic_(Cyrillic)', 'Oriya', 'Sylheti', 'Syriac_(Serto)', 'Tengwar', 'Tibetan', 'ULOG']\n",
            "(964, 20, 105, 105)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xifyAcMnRVa2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 55
        },
        "outputId": "e459142c-1142-4fa0-87e9-97858ca26f0f"
      },
      "source": [
        "print(list(classes.keys()))"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['Alphabet_of_the_Magi', 'Anglo-Saxon_Futhorc', 'Arcadian', 'Armenian', 'Asomtavruli_(Georgian)', 'Balinese', 'Bengali', 'Blackfoot_(Canadian_Aboriginal_Syllabics)', 'Braille', 'Burmese_(Myanmar)', 'Cyrillic', 'Early_Aramaic', 'Futurama', 'Grantha', 'Greek', 'Gujarati', 'Hebrew', 'Inuktitut_(Canadian_Aboriginal_Syllabics)', 'Japanese_(hiragana)', 'Japanese_(katakana)', 'Korean', 'Latin', 'Malay_(Jawi_-_Arabic)', 'Mkhedruli_(Georgian)', 'N_Ko', 'Ojibwe_(Canadian_Aboriginal_Syllabics)', 'Sanskrit', 'Syriac_(Estrangelo)', 'Tagalog', 'Tifinagh']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "W_S439qQbecy",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "a15e6b75-323b-4033-8819-c3655401063c"
      },
      "source": [
        "import numpy.random as rng\n",
        "import numpy as np\n",
        "from sklearn.utils import shuffle\n",
        "\n",
        "class Siamese_Loader:\n",
        "    \"\"\"For loading batches and testing tasks to a siamese net\"\"\"\n",
        "    def __init__(self, path, data_subsets = [\"train\", \"val\"]):\n",
        "        self.data = {}\n",
        "        self.categories = {}\n",
        "        self.info = {}\n",
        "        \n",
        "        for name in data_subsets:\n",
        "            file_path = os.path.join(path, name + \".pickle\")\n",
        "            print(\"loading data from {}\".format(file_path))\n",
        "            with open(file_path,\"rb\") as f:\n",
        "                (X,c) = pickle.load(f)\n",
        "                self.data[name] = X\n",
        "                self.categories[name] = c\n",
        "\n",
        "    def get_batch(self,batch_size,s=\"train\"):\n",
        "        \"\"\"Create batch of n pairs, half same class, half different class\"\"\"\n",
        "        X=self.data[s]\n",
        "        n_classes, n_examples, w, h = X.shape\n",
        "\n",
        "        #randomly sample several classes to use in the batch\n",
        "        categories = rng.choice(n_classes,size=(batch_size,),replace=False)\n",
        "        #initialize 2 empty arrays for the input image batch\n",
        "        pairs=[np.zeros((batch_size, h, w,1)) for i in range(2)]\n",
        "        #initialize vector for the targets, and make one half of it '1's, so 2nd half of batch has same class\n",
        "        targets=np.zeros((batch_size,))\n",
        "        targets[batch_size//2:] = 1\n",
        "        for i in range(batch_size):\n",
        "            category = categories[i]\n",
        "            idx_1 = rng.randint(0, n_examples)\n",
        "            pairs[0][i,:,:,:] = X[category, idx_1].reshape(w, h, 1)\n",
        "            idx_2 = rng.randint(0, n_examples)\n",
        "            #pick images of same class for 1st half, different for 2nd\n",
        "            if i >= batch_size // 2:\n",
        "                category_2 = category  \n",
        "            else: \n",
        "                #add a random number to the category modulo n classes to ensure 2nd image has\n",
        "                # ..different category\n",
        "                category_2 = (category + rng.randint(1,n_classes)) % n_classes\n",
        "            pairs[1][i,:,:,:] = X[category_2,idx_2].reshape(w, h,1)\n",
        "        return pairs, targets\n",
        "    \n",
        "    def generate(self, batch_size, s=\"train\"):\n",
        "        \"\"\"a generator for batches, so model.fit_generator can be used. \"\"\"\n",
        "        while True:\n",
        "            pairs, targets = self.get_batch(batch_size,s)\n",
        "            yield (pairs, targets)    \n",
        "\n",
        "    def make_oneshot_task(self,N,s=\"val\",language=None):\n",
        "        \"\"\"Create pairs of test image, support set for testing N way one-shot learning. \"\"\"\n",
        "        X=self.data[s]\n",
        "        n_classes, n_examples, w, h = X.shape\n",
        "        indices = rng.randint(0,n_examples,size=(N,))\n",
        "        if language is not None:\n",
        "            low, high = self.categories[s][language]\n",
        "            if N > high - low:\n",
        "                raise ValueError(\"This language ({}) has less than {} letters\".format(language, N))\n",
        "            categories = rng.choice(range(low,high),size=(N,),replace=False)\n",
        "            \n",
        "        else:#if no language specified just pick a bunch of random letters\n",
        "            categories = rng.choice(range(n_classes),size=(N,),replace=False)            \n",
        "        true_category = categories[0]\n",
        "        ex1, ex2 = rng.choice(n_examples,replace=False,size=(2,))\n",
        "        test_image = np.asarray([X[true_category,ex1,:,:]]*N).reshape(N, w, h,1)\n",
        "        support_set = X[categories,indices,:,:]\n",
        "        support_set[0,:,:] = X[true_category,ex2]\n",
        "        support_set = support_set.reshape(N, w, h,1)\n",
        "        targets = np.zeros((N,))\n",
        "        targets[0] = 1\n",
        "        targets, test_image, support_set = shuffle(targets, test_image, support_set)\n",
        "        pairs = [test_image,support_set]\n",
        "\n",
        "        return pairs, targets\n",
        "    \n",
        "    def test_oneshot(self,model,N,k,s=\"val\",verbose=0):\n",
        "        \"\"\"Test average N way oneshot learning accuracy of a siamese neural net over k one-shot tasks\"\"\"\n",
        "        n_correct = 0\n",
        "        if verbose:\n",
        "            print(\"Evaluating model on {} random {} way one-shot learning tasks ... \\n\".format(k,N))\n",
        "        for i in range(k):\n",
        "            inputs, targets = self.make_oneshot_task(N,s)\n",
        "            probs = model.predict(inputs)\n",
        "            if np.argmax(probs) == np.argmax(targets):\n",
        "                n_correct+=1\n",
        "        percent_correct = (100.0*n_correct / k)\n",
        "        if verbose:\n",
        "            print(\"Got an average of {}% {} way one-shot learning accuracy \\n\".format(percent_correct,N))\n",
        "        return percent_correct\n",
        "    \n",
        "    def train(self, model, epochs, verbosity):\n",
        "        model.fit_generator(self.generate(batch_size))\n",
        "    \n",
        "\n",
        "loader = Siamese_Loader(project_path)"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "loading data from /content/drive/My Drive/Colab Notebooks/databases/datasets-master/Omniglot Dataset/train.pickle\n",
            "loading data from /content/drive/My Drive/Colab Notebooks/databases/datasets-master/Omniglot Dataset/val.pickle\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "4YMUFwK6BGg9"
      },
      "source": [
        "### Concatenate a bunch of images into a big matrix for plotting purposes"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "WBwmTWbrbec0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 255
        },
        "outputId": "fcb1e4a6-c170-4bf4-a94a-9a4adc8b98d0"
      },
      "source": [
        "def concat_images(X):\n",
        "    \"\"\"Concatenates a bunch of images into a big matrix for plotting purposes.\"\"\"\n",
        "    nc, h , w, _ = X.shape\n",
        "    X = X.reshape(nc, h, w)\n",
        "    n = np.ceil(np.sqrt(nc)).astype(\"int8\")\n",
        "    img = np.zeros((n*w,n*h))\n",
        "    x = 0\n",
        "    y = 0\n",
        "    for example in range(nc):\n",
        "        img[x*w:(x+1)*w,y*h:(y+1)*h] = X[example]\n",
        "        y += 1\n",
        "        if y >= n:\n",
        "            y = 0\n",
        "            x += 1\n",
        "    return img\n",
        "\n",
        "\n",
        "def plot_oneshot_task(pairs):\n",
        "    fig,(ax1,ax2) = plt.subplots(2)\n",
        "    ax1.matshow(pairs[0][0].reshape(105,105), cmap='gray')\n",
        "    img = concat_images(pairs[1])\n",
        "    ax1.get_yaxis().set_visible(False)\n",
        "    ax1.get_xaxis().set_visible(False)\n",
        "    ax2.matshow(img,cmap='gray')\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "    plt.show()\n",
        "\n",
        "pairs, targets = loader.make_oneshot_task(20,\"train\",\"Japanese_(katakana)\")\n",
        "plot_oneshot_task(pairs)"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "io5n-b2rlBiZ",
        "colab_type": "text"
      },
      "source": [
        ""
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "1Hsj4xBJbec1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "0aaa5777-1781-482a-e8bf-cf462a3dbdb8"
      },
      "source": [
        "import time\n",
        "\n",
        "evaluate_every = 10 # interval for evaluating on one-shot tasks\n",
        "loss_every = 20 # interval for printing loss (iterations)\n",
        "batch_size = 32\n",
        "n_iter = 150\n",
        "N_way = 20 # how many classes for testing one-shot tasks>\n",
        "n_val = 250 # how many one-shot tasks to validate on?\n",
        "best = -1\n",
        "print(\"Starting training process!\")\n",
        "print(\"-------------------------------------\")\n",
        "t_start = time.time()\n",
        "for i in range(1, n_iter):\n",
        "    (inputs,targets)=loader.get_batch(batch_size)\n",
        "    loss=model.train_on_batch(inputs,targets)\n",
        "    print(\"\\n ------------- \\n\")\n",
        "    print(\"Loss: {0}\".format(loss)) \n",
        "    if i % evaluate_every == 0:\n",
        "        print(\"Time for {0} iterations: {1}\".format(i, time.time()-t_start))\n",
        "        val_acc = loader.test_oneshot(model,N_way,n_val,verbose=True)\n",
        "        if val_acc >= best:\n",
        "            print(\"Current best: {0}, previous best: {1}\".format(val_acc, best))\n",
        "            print(\"Saving weights to: {0} \\n\".format(project_path))\n",
        "            model.save_weights(project_path)\n",
        "            best=val_acc\n",
        "    \n",
        "    if i % loss_every == 0:\n",
        "        print(\"iteration {}, training loss: {:.2f},\".format(i,loss))\n",
        "\n",
        "        \n",
        "# weights_path = os.path.join(project_path, \"model_weights.h5\")\n",
        "# model.load_weights(weights_path)"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Starting training process!\n",
            "-------------------------------------\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6813125610351562\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.845028281211853\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 1.039741039276123\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6924778819084167\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6808640956878662\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6966125965118408\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.714089035987854\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.704186737537384\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6803456544876099\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6870616674423218\n",
            "Time for 10 iterations: 7.490972995758057\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 8.8% 20 way one-shot learning accuracy \n",
            "\n",
            "Current best: 8.8, previous best: -1\n",
            "Saving weights to: /content/drive/My Drive/Colab Notebooks/databases/datasets-master/Omniglot Dataset/ \n",
            "\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6836696267127991\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6832743883132935\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6982652544975281\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6914240121841431\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6883374452590942\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6903301477432251\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6869367957115173\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.690883994102478\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6805320978164673\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6842007040977478\n",
            "Time for 20 iterations: 19.920472145080566\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 10.0% 20 way one-shot learning accuracy \n",
            "\n",
            "Current best: 10.0, previous best: 8.8\n",
            "Saving weights to: /content/drive/My Drive/Colab Notebooks/databases/datasets-master/Omniglot Dataset/ \n",
            "\n",
            "iteration 20, training loss: 0.68,\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6845278739929199\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6913983821868896\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6855804920196533\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6871395111083984\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6908627152442932\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6865600943565369\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6857274770736694\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6904840469360352\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6888808608055115\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6899422407150269\n",
            "Time for 30 iterations: 32.148271799087524\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 12.4% 20 way one-shot learning accuracy \n",
            "\n",
            "Current best: 12.4, previous best: 10.0\n",
            "Saving weights to: /content/drive/My Drive/Colab Notebooks/databases/datasets-master/Omniglot Dataset/ \n",
            "\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6873795986175537\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.691596269607544\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.690192699432373\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6946794986724854\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6927345395088196\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6910198926925659\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6930452585220337\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6953619122505188\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6933744549751282\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6904515624046326\n",
            "Time for 40 iterations: 44.69210648536682\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 12.0% 20 way one-shot learning accuracy \n",
            "\n",
            "iteration 40, training loss: 0.69,\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6922237873077393\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6912879347801208\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6919132471084595\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6926447153091431\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.690656840801239\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6945619583129883\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6919093728065491\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6935319900512695\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931690573692322\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.692872166633606\n",
            "Time for 50 iterations: 55.43156695365906\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 8.0% 20 way one-shot learning accuracy \n",
            "\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6929929256439209\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6928512454032898\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6927276849746704\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.692625880241394\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6937401294708252\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6929575204849243\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6914821863174438\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6900153756141663\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6930431723594666\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6923107504844666\n",
            "Time for 60 iterations: 66.03817105293274\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 2.8% 20 way one-shot learning accuracy \n",
            "\n",
            "iteration 60, training loss: 0.69,\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6952341198921204\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6929632425308228\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6932367086410522\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931371688842773\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931599974632263\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6933426856994629\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6930892467498779\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.693356990814209\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6937954425811768\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.693640410900116\n",
            "Time for 70 iterations: 76.04377865791321\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 4.8% 20 way one-shot learning accuracy \n",
            "\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6932100057601929\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931746602058411\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931567192077637\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931564807891846\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931776404380798\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931781768798828\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6930454969406128\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6929432153701782\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6932012438774109\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6925629377365112\n",
            "Time for 80 iterations: 86.18309092521667\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 10.0% 20 way one-shot learning accuracy \n",
            "\n",
            "iteration 80, training loss: 0.69,\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6923662424087524\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6933177709579468\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6943390965461731\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931987404823303\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6927189826965332\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6927268505096436\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6929636001586914\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6925085186958313\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6911836862564087\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6933072805404663\n",
            "Time for 90 iterations: 96.14080142974854\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 7.6% 20 way one-shot learning accuracy \n",
            "\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6908032894134521\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6941705346107483\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6928815841674805\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931484937667847\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.693118155002594\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931110620498657\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.69288170337677\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.692977786064148\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.693130612373352\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931499242782593\n",
            "Time for 100 iterations: 106.22343373298645\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 2.8% 20 way one-shot learning accuracy \n",
            "\n",
            "iteration 100, training loss: 0.69,\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931476593017578\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931475400924683\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931577920913696\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.693136990070343\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6934211254119873\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.693150520324707\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931520104408264\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931395530700684\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931860446929932\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6933358907699585\n",
            "Time for 110 iterations: 116.154376745224\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 3.6% 20 way one-shot learning accuracy \n",
            "\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6930860280990601\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.693158745765686\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6930116415023804\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6932806968688965\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6929519176483154\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6920723915100098\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6930491924285889\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6927310228347778\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6933292150497437\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6933472156524658\n",
            "Time for 120 iterations: 126.23153066635132\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 4.4% 20 way one-shot learning accuracy \n",
            "\n",
            "iteration 120, training loss: 0.69,\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6929473876953125\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6932454109191895\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6930946111679077\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931132078170776\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6932016611099243\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6930518746376038\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6933580040931702\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6935399770736694\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6933647394180298\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931117177009583\n",
            "Time for 130 iterations: 136.1901228427887\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 5.6% 20 way one-shot learning accuracy \n",
            "\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931556463241577\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6933766603469849\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6933222413063049\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931817531585693\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.693210780620575\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931357383728027\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931436061859131\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931148767471313\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6927042007446289\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6924809217453003\n",
            "Time for 140 iterations: 146.34068369865417\n",
            "Evaluating model on 250 random 20 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 4.4% 20 way one-shot learning accuracy \n",
            "\n",
            "iteration 140, training loss: 0.69,\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931125521659851\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6930420398712158\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6933534741401672\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931416988372803\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931711435317993\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931565999984741\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931418180465698\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931848526000977\n",
            "\n",
            " ------------- \n",
            "\n",
            "Loss: 0.6931449770927429\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "CN0FksuMbec3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "cf5b81f5-49fe-4b38-fe47-86f4333d4e4a"
      },
      "source": [
        "def nearest_neighbour_correct(pairs,targets):\n",
        "    \"\"\"returns 1 if nearest neighbour gets the correct answer for a one-shot task\n",
        "        given by (pairs, targets)\"\"\"\n",
        "    L2_distances = np.zeros_like(targets)\n",
        "    for i in range(len(targets)):\n",
        "        L2_distances[i] = np.sum(np.sqrt(pairs[0][i]**2 - pairs[1][i]**2))\n",
        "    if np.argmin(L2_distances) == np.argmax(targets):\n",
        "        return 1\n",
        "    return 0\n",
        "\n",
        "\n",
        "def test_nn_accuracy(N_ways,n_trials,loader):\n",
        "    \"\"\"Returns accuracy of one shot \"\"\"\n",
        "    print(\"Evaluating nearest neighbour on {} unique {} way one-shot learning tasks ...\".format(n_trials,N_ways))\n",
        "\n",
        "    n_right = 0\n",
        "    \n",
        "    for i in range(n_trials):\n",
        "        pairs,targets = loader.make_oneshot_task(N_ways,\"val\")\n",
        "        correct = nearest_neighbour_correct(pairs,targets)\n",
        "        n_right += correct\n",
        "    return 100.0 * n_right / n_trials\n",
        "\n",
        "\n",
        "ways = np.arange(1, 20, 2)\n",
        "resume =  False\n",
        "val_accs, train_accs,nn_accs = [], [], []\n",
        "trials = 450\n",
        "for N in ways:\n",
        "    val_accs.append(loader.test_oneshot(model, N,trials, \"val\", verbose=True))\n",
        "    train_accs.append(loader.test_oneshot(model, N,trials, \"train\", verbose=True))\n",
        "    nn_accs.append(test_nn_accuracy(N,trials, loader))\n",
        "    \n",
        "#plot the accuracy vs num categories for each\n",
        "plt.plot(ways, val_accs, \"m\")\n",
        "plt.plot(ways, train_accs, \"y\")\n",
        "plt.plot(ways, nn_accs, \"c\")\n",
        "\n",
        "plt.plot(ways,100.0/ways,\"r\")\n",
        "plt.show()"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Evaluating model on 450 random 1 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 100.0% 1 way one-shot learning accuracy \n",
            "\n",
            "Evaluating model on 450 random 1 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 100.0% 1 way one-shot learning accuracy \n",
            "\n",
            "Evaluating nearest neighbour on 450 unique 1 way one-shot learning tasks ...\n",
            "Evaluating model on 450 random 3 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 31.555555555555557% 3 way one-shot learning accuracy \n",
            "\n",
            "Evaluating model on 450 random 3 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 39.77777777777778% 3 way one-shot learning accuracy \n",
            "\n",
            "Evaluating nearest neighbour on 450 unique 3 way one-shot learning tasks ...\n",
            "Evaluating model on 450 random 5 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 24.22222222222222% 5 way one-shot learning accuracy \n",
            "\n",
            "Evaluating model on 450 random 5 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 27.333333333333332% 5 way one-shot learning accuracy \n",
            "\n",
            "Evaluating nearest neighbour on 450 unique 5 way one-shot learning tasks ...\n",
            "Evaluating model on 450 random 7 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 18.88888888888889% 7 way one-shot learning accuracy \n",
            "\n",
            "Evaluating model on 450 random 7 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 18.22222222222222% 7 way one-shot learning accuracy \n",
            "\n",
            "Evaluating nearest neighbour on 450 unique 7 way one-shot learning tasks ...\n",
            "Evaluating model on 450 random 9 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 12.444444444444445% 9 way one-shot learning accuracy \n",
            "\n",
            "Evaluating model on 450 random 9 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 16.22222222222222% 9 way one-shot learning accuracy \n",
            "\n",
            "Evaluating nearest neighbour on 450 unique 9 way one-shot learning tasks ...\n",
            "Evaluating model on 450 random 11 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 11.777777777777779% 11 way one-shot learning accuracy \n",
            "\n",
            "Evaluating model on 450 random 11 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 12.444444444444445% 11 way one-shot learning accuracy \n",
            "\n",
            "Evaluating nearest neighbour on 450 unique 11 way one-shot learning tasks ...\n",
            "Evaluating model on 450 random 13 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 7.777777777777778% 13 way one-shot learning accuracy \n",
            "\n",
            "Evaluating model on 450 random 13 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 11.333333333333334% 13 way one-shot learning accuracy \n",
            "\n",
            "Evaluating nearest neighbour on 450 unique 13 way one-shot learning tasks ...\n",
            "Evaluating model on 450 random 15 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 9.333333333333334% 15 way one-shot learning accuracy \n",
            "\n",
            "Evaluating model on 450 random 15 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 8.0% 15 way one-shot learning accuracy \n",
            "\n",
            "Evaluating nearest neighbour on 450 unique 15 way one-shot learning tasks ...\n",
            "Evaluating model on 450 random 17 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 8.222222222222221% 17 way one-shot learning accuracy \n",
            "\n",
            "Evaluating model on 450 random 17 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 8.222222222222221% 17 way one-shot learning accuracy \n",
            "\n",
            "Evaluating nearest neighbour on 450 unique 17 way one-shot learning tasks ...\n",
            "Evaluating model on 450 random 19 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 6.0% 19 way one-shot learning accuracy \n",
            "\n",
            "Evaluating model on 450 random 19 way one-shot learning tasks ... \n",
            "\n",
            "Got an average of 6.666666666666667% 19 way one-shot learning accuracy \n",
            "\n",
            "Evaluating nearest neighbour on 450 unique 19 way one-shot learning tasks ...\n"
          ],
          "name": "stdout"
        },
        {
          "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": "code",
      "metadata": {
        "colab_type": "code",
        "id": "zDSw9a9Fbec5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 533
        },
        "outputId": "c29e7105-9c9f-4a49-ce6b-463c15ea8923"
      },
      "source": [
        "fig,ax = plt.subplots(1)\n",
        "ax.plot(ways, val_accs, \"m\", label=\"Siamese(val set)\")\n",
        "ax.plot(ways, train_accs, \"y\", label=\"Siamese(train set)\")\n",
        "plt.plot(ways, nn_accs, label=\"Nearest neighbour\")\n",
        "\n",
        "ax.plot(ways, 100.0/ways, \"g\", label=\"Random guessing\")\n",
        "plt.xlabel(\"Number of possible classes in one-shot tasks\")\n",
        "plt.ylabel(\"% Accuracy\")\n",
        "plt.title(\"Omiglot One-Shot Learning Performance of a Siamese Network\")\n",
        "box = ax.get_position()\n",
        "ax.set_position([box.x0, box.y0, box.width * 0.8, box.height])\n",
        "ax.legend(loc='center left', bbox_to_anchor=(1, 0.5))\n",
        "inputs,targets = loader.make_oneshot_task(20, \"val\")\n",
        "plt.show()\n",
        "\n",
        "plot_oneshot_task(inputs)"
      ],
      "execution_count": 48,
      "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"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "16yDrbxeqohf",
        "colab_type": "text"
      },
      "source": [
        "https://drive.google.com/file/d/1VRPWuHVcp2_7z3tZfkaq7hHm6DA5M_6H/view?usp=sharing"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FlCSBBeAqpSE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}