SimCLR_MotionSense.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.6"
    },
    "orig_nbformat": 2,
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3.7.6 64-bit",
      "metadata": {
        "interpreter": {
          "hash": "369eacbd862d285c195556fd0399bd44fd39c7fe14d5ede675fff8f359efd9d4"
        }
      }
    },
    "colab": {
      "name": "SimCLR_MotionSense.ipynb",
      "provenance": [],
      "toc_visible": true
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KPaf3-1MP53I"
      },
      "source": [
        "# SimCLR Contrastive Training for Human Activity Recognition Tutorial"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5ko1vFAZP53N"
      },
      "source": [
        "# Author: C. I. Tang\n",
        "# Based on work of Tang et al.: https://arxiv.org/abs/2011.11542\n",
        "# Contact: cit27@cl.cam.ac.uk\n",
        "# License: GNU General Public License v3.0\n",
        "\n",
        "%load_ext autoreload\n",
        "%autoreload 2"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "G46-v0cuP53N"
      },
      "source": [
        "## Imports"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "W-YuikCiP53N"
      },
      "source": [
        "import os\n",
        "import pickle\n",
        "import scipy\n",
        "import datetime\n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "\n",
        "seed = 1\n",
        "tf.random.set_seed(seed)\n",
        "np.random.seed(seed)"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5LtrK5xXP53O"
      },
      "source": [
        "# Libraries for plotting\n",
        "import matplotlib as mpl\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import sklearn.manifold\n",
        "\n",
        "sns.set_context('poster')"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dwFHmiuFP53O"
      },
      "source": [
        "# Library scripts\n",
        "import raw_data_processing\n",
        "import data_pre_processing\n",
        "import simclr_models\n",
        "import simclr_utitlities\n",
        "import transformations"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "A8MHhQrfP53O"
      },
      "source": [
        "working_directory = 'test_run/'\n",
        "dataset_save_path = working_directory\n",
        "if not os.path.exists(working_directory):\n",
        "    os.mkdir(working_directory)\n"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jxwMwhqeP53P"
      },
      "source": [
        "## MotionSense Dataset\n",
        "\n",
        "In this section, the MotionSense dataset will be downloaded and parsed. The results will then be saved in a python pickle file.\n",
        "(Note: This section only needs to be run once)\n",
        "\n",
        "Citation:\n",
        "```\n",
        "@inproceedings{Malekzadeh:2019:MSD:3302505.3310068,\n",
        "author = {Malekzadeh, Mohammad and Clegg, Richard G. and Cavallaro, Andrea and Haddadi, Hamed},\n",
        "title = {Mobile Sensor Data Anonymization},\n",
        "booktitle = {Proceedings of the International Conference on Internet of Things Design and Implementation},\n",
        "series = {IoTDI '19},\n",
        "year = {2019},\n",
        "isbn = {978-1-4503-6283-2},\n",
        "location = {Montreal, Quebec, Canada},\n",
        "pages = {49--58},\n",
        "numpages = {10},\n",
        "url = {http://doi.acm.org/10.1145/3302505.3310068},\n",
        "doi = {10.1145/3302505.3310068},\n",
        "acmid = {3310068},\n",
        "publisher = {ACM},\n",
        "address = {New York, NY, USA},\n",
        "keywords = {adversarial training, deep learning, edge computing, sensor data privacy, time series analysis},\n",
        "} \n",
        "```"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4Pn7qJSyP53P"
      },
      "source": [
        "### Downloading & Unzipping"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QTk1khZbP53P"
      },
      "source": [
        "import requests\n",
        "import zipfile"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vOpnh48_P53P"
      },
      "source": [
        "dataset_url = 'https://github.com/mmalekzadeh/motion-sense/blob/master/data/B_Accelerometer_data.zip?raw=true'\n",
        "\n",
        "r = requests.get(dataset_url, allow_redirects=True)\n",
        "with open(working_directory + 'B_Accelerometer_data.zip', 'wb') as f:\n",
        "    f.write(r.content)"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E6bX8R7oP53Q"
      },
      "source": [
        "\n",
        "with zipfile.ZipFile(working_directory + 'B_Accelerometer_data.zip', 'r') as zip_ref:\n",
        "    zip_ref.extractall(working_directory)"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "suWBmPPdP53Q"
      },
      "source": [
        "### Data Processing"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zSJKFFsUP53Q",
        "outputId": "bf1f9d9d-be85-4775-e43d-76348e8992e6"
      },
      "source": [
        "accelerometer_data_folder_path = working_directory + 'B_Accelerometer_data/'\n",
        "user_datasets = raw_data_processing.process_motion_sense_accelerometer_files(accelerometer_data_folder_path)"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "test_run/B_Accelerometer_data/dws_1\n",
            "test_run/B_Accelerometer_data/dws_11\n",
            "test_run/B_Accelerometer_data/dws_2\n",
            "test_run/B_Accelerometer_data/jog_16\n",
            "test_run/B_Accelerometer_data/jog_9\n",
            "test_run/B_Accelerometer_data/sit_13\n",
            "test_run/B_Accelerometer_data/sit_5\n",
            "test_run/B_Accelerometer_data/std_14\n",
            "test_run/B_Accelerometer_data/std_6\n",
            "test_run/B_Accelerometer_data/ups_12\n",
            "test_run/B_Accelerometer_data/ups_3\n",
            "test_run/B_Accelerometer_data/ups_4\n",
            "test_run/B_Accelerometer_data/wlk_15\n",
            "test_run/B_Accelerometer_data/wlk_7\n",
            "test_run/B_Accelerometer_data/wlk_8\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lE_PlKn0P53R"
      },
      "source": [
        "with open(working_directory + 'motion_sense_user_split.pkl', 'wb') as f:\n",
        "    pickle.dump({\n",
        "        'user_split': user_datasets,\n",
        "    }, f)"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eMHhVkPeP53R"
      },
      "source": [
        "## Pre-processing"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V8oWxqFZP53R"
      },
      "source": [
        "# Parameters\n",
        "window_size = 400\n",
        "input_shape = (window_size, 3)\n",
        "\n",
        "# Dataset Metadata \n",
        "transformation_multiple = 1\n",
        "dataset_name = 'motion_sense.pkl'\n",
        "dataset_name_user_split = 'motion_sense_user_split.pkl'\n",
        "\n",
        "label_list = ['null', 'sit', 'std', 'wlk', 'ups', 'dws', 'jog']\n",
        "label_list_full_name = ['null', 'sitting', 'standing', 'walking', 'walking upstairs', 'walking downstairs', 'jogging']\n",
        "has_null_class = True\n",
        "\n",
        "label_map = dict([(l, i) for i, l in enumerate(label_list)])\n",
        "\n",
        "output_shape = len(label_list)\n",
        "\n",
        "model_save_name = f\"motionsense_acc\"\n",
        "\n",
        "sampling_rate = 50.0\n",
        "unit_conversion = scipy.constants.g\n",
        "\n",
        "# a fixed user-split\n",
        "test_users_fixed = [1, 14, 19, 23, 6]\n",
        "def get_fixed_split_users(har_users):\n",
        "    # test_users = har_users[0::5]\n",
        "    test_users = test_users_fixed\n",
        "    train_users = [u for u in har_users if u not in test_users]\n",
        "    return (train_users, test_users)"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pvQ1DV5vP53S"
      },
      "source": [
        "with open(dataset_save_path + dataset_name_user_split, 'rb') as f:\n",
        "    dataset_dict = pickle.load(f)\n",
        "    user_datasets = dataset_dict['user_split']\n"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yl0jGEOFP53S",
        "outputId": "b6d10ab4-3ca1-4909-e5c5-59953dcc76ff"
      },
      "source": [
        "har_users = list(user_datasets.keys())\n",
        "train_users, test_users = get_fixed_split_users(har_users)\n",
        "print(f'Testing: {test_users}, Training: {train_users}')"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Testing: [1, 14, 19, 23, 6], Training: [10, 11, 12, 13, 15, 16, 17, 18, 2, 20, 21, 22, 24, 3, 4, 5, 7, 8, 9]\n"
          ]
        }
      ]
    },
    {
      "source": [
        "Here we convert the original dataset into a windowed one, and split it into training, validation and testing sets. "
      ],
      "cell_type": "markdown",
      "metadata": {}
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DRdmCo0ZP53S",
        "outputId": "a5ed1259-c764-4d65-a696-bcc1d645153e"
      },
      "source": [
        "np_train, np_val, np_test = data_pre_processing.pre_process_dataset_composite(\n",
        "    user_datasets=user_datasets, \n",
        "    label_map=label_map, \n",
        "    output_shape=output_shape, \n",
        "    train_users=train_users, \n",
        "    test_users=test_users, \n",
        "    window_size=window_size, \n",
        "    shift=window_size//2, \n",
        "    normalise_dataset=True, \n",
        "    verbose=1\n",
        ")\n",
        "\n"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test\n(array(['dws', 'jog', 'sit', 'std', 'ups', 'wlk'], dtype='<U3'), array([112, 133, 360, 335, 148, 331]))\n(array([1, 2, 3, 4, 5, 6]), array([360, 335, 331, 148, 112, 133]))\n-----------------\nTrain\n(array(['dws', 'jog', 'sit', 'std', 'ups', 'wlk'], dtype='<U3'), array([ 449,  480, 1282, 1146,  546, 1308]))\n(array([1, 2, 3, 4, 5, 6]), array([1282, 1146, 1308,  546,  449,  480]))\n-----------------\nTraining data shape: (4168, 400, 3)\nValidation data shape: (1043, 400, 3)\nTesting data shape: (1419, 400, 3)\n"
          ]
        }
      ]
    },
    {
      "source": [
        "An alternative processing pipeline can be used to retain the user id as a feature and/or a label in the output dataset.\n",
        "\n",
        "Note: This is not used in the study, and thus is commented in this notebook."
      ],
      "cell_type": "markdown",
      "metadata": {}
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "metadata": {},
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Data shape of user 1 : (293, 400, 3)\n",
            "Data shape of user 10 : (280, 400, 3)\n",
            "Data shape of user 11 : (279, 400, 3)\n",
            "Data shape of user 12 : (248, 400, 3)\n",
            "Data shape of user 13 : (238, 400, 3)\n",
            "Data shape of user 14 : (263, 400, 3)\n",
            "Data shape of user 15 : (281, 400, 3)\n",
            "Data shape of user 16 : (311, 400, 3)\n",
            "Data shape of user 17 : (258, 400, 3)\n",
            "Data shape of user 18 : (297, 400, 3)\n",
            "Data shape of user 19 : (343, 400, 3)\n",
            "Data shape of user 2 : (292, 400, 3)\n",
            "Data shape of user 20 : (259, 400, 3)\n",
            "Data shape of user 21 : (327, 400, 3)\n",
            "Data shape of user 22 : (263, 400, 3)\n",
            "Data shape of user 23 : (252, 400, 3)\n",
            "Data shape of user 24 : (238, 400, 3)\n",
            "Data shape of user 3 : (295, 400, 3)\n",
            "Data shape of user 4 : (262, 400, 3)\n",
            "Data shape of user 5 : (244, 400, 3)\n",
            "Data shape of user 6 : (268, 400, 3)\n",
            "Data shape of user 7 : (287, 400, 3)\n",
            "Data shape of user 8 : (285, 400, 3)\n",
            "Data shape of user 9 : (267, 400, 3)\n",
            "User 1: id 1 -> 0, data shape (293, 400, 3) -> (293, 400, 4), labels shape (293, 7) -> (293, 7)\n",
            "User 10: id 10 -> 9, data shape (280, 400, 3) -> (280, 400, 4), labels shape (280, 7) -> (280, 7)\n",
            "User 11: id 11 -> 10, data shape (279, 400, 3) -> (279, 400, 4), labels shape (279, 7) -> (279, 7)\n",
            "User 12: id 12 -> 11, data shape (248, 400, 3) -> (248, 400, 4), labels shape (248, 7) -> (248, 7)\n",
            "User 13: id 13 -> 12, data shape (238, 400, 3) -> (238, 400, 4), labels shape (238, 7) -> (238, 7)\n",
            "User 14: id 14 -> 13, data shape (263, 400, 3) -> (263, 400, 4), labels shape (263, 7) -> (263, 7)\n",
            "User 15: id 15 -> 14, data shape (281, 400, 3) -> (281, 400, 4), labels shape (281, 7) -> (281, 7)\n",
            "User 16: id 16 -> 15, data shape (311, 400, 3) -> (311, 400, 4), labels shape (311, 7) -> (311, 7)\n",
            "User 17: id 17 -> 16, data shape (258, 400, 3) -> (258, 400, 4), labels shape (258, 7) -> (258, 7)\n",
            "User 18: id 18 -> 17, data shape (297, 400, 3) -> (297, 400, 4), labels shape (297, 7) -> (297, 7)\n",
            "User 19: id 19 -> 18, data shape (343, 400, 3) -> (343, 400, 4), labels shape (343, 7) -> (343, 7)\n",
            "User 2: id 2 -> 1, data shape (292, 400, 3) -> (292, 400, 4), labels shape (292, 7) -> (292, 7)\n",
            "User 20: id 20 -> 19, data shape (259, 400, 3) -> (259, 400, 4), labels shape (259, 7) -> (259, 7)\n",
            "User 21: id 21 -> 20, data shape (327, 400, 3) -> (327, 400, 4), labels shape (327, 7) -> (327, 7)\n",
            "User 22: id 22 -> 21, data shape (263, 400, 3) -> (263, 400, 4), labels shape (263, 7) -> (263, 7)\n",
            "User 23: id 23 -> 22, data shape (252, 400, 3) -> (252, 400, 4), labels shape (252, 7) -> (252, 7)\n",
            "User 24: id 24 -> 23, data shape (238, 400, 3) -> (238, 400, 4), labels shape (238, 7) -> (238, 7)\n",
            "User 3: id 3 -> 2, data shape (295, 400, 3) -> (295, 400, 4), labels shape (295, 7) -> (295, 7)\n",
            "User 4: id 4 -> 3, data shape (262, 400, 3) -> (262, 400, 4), labels shape (262, 7) -> (262, 7)\n",
            "User 5: id 5 -> 4, data shape (244, 400, 3) -> (244, 400, 4), labels shape (244, 7) -> (244, 7)\n",
            "User 6: id 6 -> 5, data shape (268, 400, 3) -> (268, 400, 4), labels shape (268, 7) -> (268, 7)\n",
            "User 7: id 7 -> 6, data shape (287, 400, 3) -> (287, 400, 4), labels shape (287, 7) -> (287, 7)\n",
            "User 8: id 8 -> 7, data shape (285, 400, 3) -> (285, 400, 4), labels shape (285, 7) -> (285, 7)\n",
            "User 9: id 9 -> 8, data shape (267, 400, 3) -> (267, 400, 4), labels shape (267, 7) -> (267, 7)\n",
            "Training data shape: (4168, 400, 4)\n",
            "Validation data shape: (1043, 400, 4)\n",
            "Testing data shape: (1419, 400, 4)\n"
          ]
        }
      ],
      "source": [
        "# user_datasets_processed = data_pre_processing.pre_process_dataset_composite_in_user_format(\n",
        "#     user_datasets=user_datasets, \n",
        "#     label_map=label_map, \n",
        "#     output_shape=output_shape, \n",
        "#     train_users=train_users, \n",
        "#     window_size=window_size, \n",
        "#     shift=window_size//2, \n",
        "#     normalise_dataset=True, \n",
        "#     verbose=1\n",
        "# )\n",
        "\n",
        "# user_datasets_with_user_ids, user_id_mapping = data_pre_processing.add_user_id_to_windowed_dataset(\n",
        "#     user_datasets_processed, \n",
        "#     encode_user_id=True, \n",
        "#     as_feature=True, \n",
        "#     as_label=False, \n",
        "#     verbose=1\n",
        "# )\n",
        "\n",
        "# train_x, train_y, test_x, test_y = data_pre_processing.combine_windowed_dataset(user_datasets_with_user_ids, train_users)\n",
        "# train_x_split, val_x_split, train_y_split, val_y_split = sklearn.model_selection.train_test_split(train_x, train_y, test_size=0.2, random_state=42)\n",
        "\n",
        "# np_train = (train_x_split, train_y_split)\n",
        "# np_val = (val_x_split, val_y_split)\n",
        "# np_test = (test_x, test_y)\n",
        "\n",
        "# print(\"Training data shape:\", train_x_split.shape)\n",
        "# print(\"Validation data shape:\", val_x_split.shape)\n",
        "# print(\"Testing data shape:\", test_x.shape)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n6ekD_oFP53S"
      },
      "source": [
        "## SimCLR Training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0nliJ-GgP53T"
      },
      "source": [
        "batch_size = 512\n",
        "decay_steps = 1000\n",
        "epochs = 200\n",
        "temperature = 0.1\n",
        "transform_funcs = [\n",
        "    # transformations.scaling_transform_vectorized, # Use Scaling trasnformation\n",
        "    transformations.rotation_transform_vectorized # Use rotation trasnformation\n",
        "]\n",
        "transformation_function = simclr_utitlities.generate_composite_transform_function_simple(transform_funcs)\n",
        "\n",
        "# trasnformation_indices = [2] # Use rotation trasnformation only\n",
        "# trasnformation_indices = [1, 2] # Use Scaling and rotation trasnformation\n",
        "\n",
        "# trasnform_funcs_vectorized = [\n",
        "#     transformations.noise_transform_vectorized, \n",
        "#     transformations.scaling_transform_vectorized, \n",
        "#     transformations.rotation_transform_vectorized, \n",
        "#     transformations.negate_transform_vectorized, \n",
        "#     transformations.time_flip_transform_vectorized, \n",
        "#     transformations.time_segment_permutation_transform_improved, \n",
        "#     transformations.time_warp_transform_low_cost, \n",
        "#     transformations.channel_shuffle_transform_vectorized\n",
        "# ]\n",
        "# transform_funcs_names = ['noised', 'scaled', 'rotated', 'negated', 'time_flipped', 'permuted', 'time_warped', 'channel_shuffled']\n",
        "\n"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0 <function rotation_transform_vectorized at 0x7fb19d504b00>\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Vxk1iPzjP53T",
        "outputId": "b9a57028-4518-4e04-a23b-16dcaa001aeb"
      },
      "source": [
        "start_time = datetime.datetime.now()\n",
        "start_time_str = start_time.strftime(\"%Y%m%d-%H%M%S\")\n",
        "tf.keras.backend.set_floatx('float32')\n",
        "\n",
        "lr_decayed_fn = tf.keras.experimental.CosineDecay(initial_learning_rate=0.1, decay_steps=decay_steps)\n",
        "optimizer = tf.keras.optimizers.SGD(lr_decayed_fn)\n",
        "# transformation_function = simclr_utitlities.generate_combined_transform_function(trasnform_funcs_vectorized, indices=trasnformation_indices)\n",
        "\n",
        "base_model = simclr_models.create_base_model(input_shape, model_name=\"base_model\")\n",
        "simclr_model = simclr_models.attach_simclr_head(base_model)\n",
        "simclr_model.summary()\n",
        "\n",
        "trained_simclr_model, epoch_losses = simclr_utitlities.simclr_train_model(simclr_model, np_train[0], optimizer, batch_size, transformation_function, temperature=temperature, epochs=epochs, is_trasnform_function_vectorized=True, verbose=1)\n",
        "\n",
        "simclr_model_save_path = f\"{working_directory}{start_time_str}_simclr.hdf5\"\n",
        "trained_simclr_model.save(simclr_model_save_path)\n",
        "\n"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<function rotation_transform_vectorized at 0x7fc7876e7510>\n",
            "Model: \"base_model_simclr\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input (InputLayer)           [(None, 400, 3)]          0         \n",
            "_________________________________________________________________\n",
            "conv1d (Conv1D)              (None, 377, 32)           2336      \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 377, 32)           0         \n",
            "_________________________________________________________________\n",
            "conv1d_1 (Conv1D)            (None, 362, 64)           32832     \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 362, 64)           0         \n",
            "_________________________________________________________________\n",
            "conv1d_2 (Conv1D)            (None, 355, 96)           49248     \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 355, 96)           0         \n",
            "_________________________________________________________________\n",
            "global_max_pooling1d (Global (None, 96)                0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 256)               24832     \n",
            "_________________________________________________________________\n",
            "activation (Activation)      (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 128)               32896     \n",
            "_________________________________________________________________\n",
            "activation_1 (Activation)    (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 50)                6450      \n",
            "=================================================================\n",
            "Total params: 148,594\n",
            "Trainable params: 148,594\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "epoch: 1 loss: 13.476\n",
            "epoch: 2 loss: 13.216\n",
            "epoch: 3 loss: 11.934\n",
            "epoch: 4 loss: 12.001\n",
            "epoch: 5 loss: 11.487\n",
            "epoch: 6 loss: 11.823\n",
            "epoch: 7 loss: 10.649\n",
            "epoch: 8 loss: 10.352\n",
            "epoch: 9 loss: 12.044\n",
            "epoch: 10 loss: 10.670\n",
            "epoch: 11 loss: 10.542\n",
            "epoch: 12 loss: 10.129\n",
            "epoch: 13 loss: 10.028\n",
            "epoch: 14 loss: 10.261\n",
            "epoch: 15 loss: 9.689\n",
            "epoch: 16 loss: 9.261\n",
            "epoch: 17 loss: 9.088\n",
            "epoch: 18 loss: 9.164\n",
            "epoch: 19 loss: 9.208\n",
            "epoch: 20 loss: 9.102\n",
            "epoch: 21 loss: 8.512\n",
            "epoch: 22 loss: 8.465\n",
            "epoch: 23 loss: 8.099\n",
            "epoch: 24 loss: 8.660\n",
            "epoch: 25 loss: 8.069\n",
            "epoch: 26 loss: 8.360\n",
            "epoch: 27 loss: 7.977\n",
            "epoch: 28 loss: 7.633\n",
            "epoch: 29 loss: 7.912\n",
            "epoch: 30 loss: 7.387\n",
            "epoch: 31 loss: 7.476\n",
            "epoch: 32 loss: 7.659\n",
            "epoch: 33 loss: 6.886\n",
            "epoch: 34 loss: 6.936\n",
            "epoch: 35 loss: 6.783\n",
            "epoch: 36 loss: 6.924\n",
            "epoch: 37 loss: 6.874\n",
            "epoch: 38 loss: 6.590\n",
            "epoch: 39 loss: 6.513\n",
            "epoch: 40 loss: 6.271\n",
            "epoch: 41 loss: 6.574\n",
            "epoch: 42 loss: 6.320\n",
            "epoch: 43 loss: 6.220\n",
            "epoch: 44 loss: 6.087\n",
            "epoch: 45 loss: 5.850\n",
            "epoch: 46 loss: 6.218\n",
            "epoch: 47 loss: 5.826\n",
            "epoch: 48 loss: 6.018\n",
            "epoch: 49 loss: 5.503\n",
            "epoch: 50 loss: 5.637\n",
            "epoch: 51 loss: 5.526\n",
            "epoch: 52 loss: 5.545\n",
            "epoch: 53 loss: 5.400\n",
            "epoch: 54 loss: 5.486\n",
            "epoch: 55 loss: 5.111\n",
            "epoch: 56 loss: 5.302\n",
            "epoch: 57 loss: 5.227\n",
            "epoch: 58 loss: 5.091\n",
            "epoch: 59 loss: 5.051\n",
            "epoch: 60 loss: 4.782\n",
            "epoch: 61 loss: 4.771\n",
            "epoch: 62 loss: 5.092\n",
            "epoch: 63 loss: 4.864\n",
            "epoch: 64 loss: 4.583\n",
            "epoch: 65 loss: 4.638\n",
            "epoch: 66 loss: 4.752\n",
            "epoch: 67 loss: 4.403\n",
            "epoch: 68 loss: 4.908\n",
            "epoch: 69 loss: 4.460\n",
            "epoch: 70 loss: 4.441\n",
            "epoch: 71 loss: 4.427\n",
            "epoch: 72 loss: 4.223\n",
            "epoch: 73 loss: 4.292\n",
            "epoch: 74 loss: 4.140\n",
            "epoch: 75 loss: 4.142\n",
            "epoch: 76 loss: 3.993\n",
            "epoch: 77 loss: 3.997\n",
            "epoch: 78 loss: 3.932\n",
            "epoch: 79 loss: 3.872\n",
            "epoch: 80 loss: 3.824\n",
            "epoch: 81 loss: 3.738\n",
            "epoch: 82 loss: 3.759\n",
            "epoch: 83 loss: 3.671\n",
            "epoch: 84 loss: 3.629\n",
            "epoch: 85 loss: 3.612\n",
            "epoch: 86 loss: 3.575\n",
            "epoch: 87 loss: 3.579\n",
            "epoch: 88 loss: 3.406\n",
            "epoch: 89 loss: 3.434\n",
            "epoch: 90 loss: 3.361\n",
            "epoch: 91 loss: 3.378\n",
            "epoch: 92 loss: 3.328\n",
            "epoch: 93 loss: 3.305\n",
            "epoch: 94 loss: 3.290\n",
            "epoch: 95 loss: 3.175\n",
            "epoch: 96 loss: 3.162\n",
            "epoch: 97 loss: 3.170\n",
            "epoch: 98 loss: 3.125\n",
            "epoch: 99 loss: 3.055\n",
            "epoch: 100 loss: 3.052\n",
            "epoch: 101 loss: 3.000\n",
            "epoch: 102 loss: 3.009\n",
            "epoch: 103 loss: 3.008\n",
            "epoch: 104 loss: 2.979\n",
            "epoch: 105 loss: 2.977\n",
            "epoch: 106 loss: 2.906\n",
            "epoch: 107 loss: 2.954\n",
            "epoch: 108 loss: 2.934\n",
            "epoch: 109 loss: 2.882\n",
            "epoch: 110 loss: 2.884\n",
            "epoch: 111 loss: 2.879\n",
            "epoch: 112 loss: 2.851\n",
            "epoch: 113 loss: 2.884\n",
            "epoch: 114 loss: 2.897\n",
            "epoch: 115 loss: 2.874\n",
            "epoch: 116 loss: 2.874\n",
            "epoch: 117 loss: 2.918\n",
            "epoch: 118 loss: 2.875\n",
            "epoch: 119 loss: 2.913\n",
            "epoch: 120 loss: 2.912\n",
            "epoch: 121 loss: 2.886\n",
            "epoch: 122 loss: 2.857\n",
            "epoch: 123 loss: 2.925\n",
            "epoch: 124 loss: 2.883\n",
            "epoch: 125 loss: 2.878\n",
            "epoch: 126 loss: 2.888\n",
            "epoch: 127 loss: 2.938\n",
            "epoch: 128 loss: 2.831\n",
            "epoch: 129 loss: 2.862\n",
            "epoch: 130 loss: 2.891\n",
            "epoch: 131 loss: 2.889\n",
            "epoch: 132 loss: 2.864\n",
            "epoch: 133 loss: 2.883\n",
            "epoch: 134 loss: 2.863\n",
            "epoch: 135 loss: 2.862\n",
            "epoch: 136 loss: 2.900\n",
            "epoch: 137 loss: 2.899\n",
            "epoch: 138 loss: 2.877\n",
            "epoch: 139 loss: 2.907\n",
            "epoch: 140 loss: 2.899\n",
            "epoch: 141 loss: 2.929\n",
            "epoch: 142 loss: 2.884\n",
            "epoch: 143 loss: 2.899\n",
            "epoch: 144 loss: 2.866\n",
            "epoch: 145 loss: 2.868\n",
            "epoch: 146 loss: 2.883\n",
            "epoch: 147 loss: 2.911\n",
            "epoch: 148 loss: 2.843\n",
            "epoch: 149 loss: 2.906\n",
            "epoch: 150 loss: 2.871\n",
            "epoch: 151 loss: 2.916\n",
            "epoch: 152 loss: 2.899\n",
            "epoch: 153 loss: 2.905\n",
            "epoch: 154 loss: 2.853\n",
            "epoch: 155 loss: 2.871\n",
            "epoch: 156 loss: 2.878\n",
            "epoch: 157 loss: 2.849\n",
            "epoch: 158 loss: 2.875\n",
            "epoch: 159 loss: 2.890\n",
            "epoch: 160 loss: 2.870\n",
            "epoch: 161 loss: 2.936\n",
            "epoch: 162 loss: 2.853\n",
            "epoch: 163 loss: 2.876\n",
            "epoch: 164 loss: 2.872\n",
            "epoch: 165 loss: 2.925\n",
            "epoch: 166 loss: 2.929\n",
            "epoch: 167 loss: 2.900\n",
            "epoch: 168 loss: 2.866\n",
            "epoch: 169 loss: 2.872\n",
            "epoch: 170 loss: 2.880\n",
            "epoch: 171 loss: 2.908\n",
            "epoch: 172 loss: 2.879\n",
            "epoch: 173 loss: 2.887\n",
            "epoch: 174 loss: 2.854\n",
            "epoch: 175 loss: 2.892\n",
            "epoch: 176 loss: 2.848\n",
            "epoch: 177 loss: 2.911\n",
            "epoch: 178 loss: 2.881\n",
            "epoch: 179 loss: 2.865\n",
            "epoch: 180 loss: 2.882\n",
            "epoch: 181 loss: 2.873\n",
            "epoch: 182 loss: 2.891\n",
            "epoch: 183 loss: 2.914\n",
            "epoch: 184 loss: 2.904\n",
            "epoch: 185 loss: 2.863\n",
            "epoch: 186 loss: 2.876\n",
            "epoch: 187 loss: 2.887\n",
            "epoch: 188 loss: 2.899\n",
            "epoch: 189 loss: 2.894\n",
            "epoch: 190 loss: 2.914\n",
            "epoch: 191 loss: 2.883\n",
            "epoch: 192 loss: 2.895\n",
            "epoch: 193 loss: 2.887\n",
            "epoch: 194 loss: 2.955\n",
            "epoch: 195 loss: 2.916\n",
            "epoch: 196 loss: 2.923\n",
            "epoch: 197 loss: 2.880\n",
            "epoch: 198 loss: 2.909\n",
            "epoch: 199 loss: 2.866\n",
            "epoch: 200 loss: 2.879\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 537
        },
        "id": "VDRIHsXLP53T",
        "outputId": "2a3d3837-fe61-43c7-cdaa-8304c1183458"
      },
      "source": [
        "plt.figure(figsize=(12,8))\n",
        "plt.plot(epoch_losses)\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.show()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HhhL7Zs9P53U"
      },
      "source": [
        "## Fine-tuning and Evaluation"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TMELGS4VP53U"
      },
      "source": [
        "### Linear Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8FVlN91PP53U",
        "outputId": "25ec0bda-977e-4840-cf9b-ca2eeb8864ea"
      },
      "source": [
        "\n",
        "total_epochs = 50\n",
        "batch_size = 200\n",
        "tag = \"linear_eval\"\n",
        "\n",
        "simclr_model = tf.keras.models.load_model(simclr_model_save_path)\n",
        "linear_evaluation_model = simclr_models.create_linear_model_from_base_model(simclr_model, output_shape, intermediate_layer=7)\n",
        "\n",
        "linear_eval_best_model_file_name = f\"{working_directory}{start_time_str}_simclr_{tag}.hdf5\"\n",
        "best_model_callback = tf.keras.callbacks.ModelCheckpoint(linear_eval_best_model_file_name,\n",
        "    monitor='val_loss', mode='min', save_best_only=True, save_weights_only=False, verbose=0\n",
        ")\n",
        "\n",
        "training_history = linear_evaluation_model.fit(\n",
        "    x = np_train[0],\n",
        "    y = np_train[1],\n",
        "    batch_size=batch_size,\n",
        "    shuffle=True,\n",
        "    epochs=total_epochs,\n",
        "    callbacks=[best_model_callback],\n",
        "    validation_data=np_val\n",
        ")\n",
        "\n",
        "linear_eval_best_model = tf.keras.models.load_model(linear_eval_best_model_file_name)\n",
        "\n",
        "print(\"Model with lowest validation Loss:\")\n",
        "print(simclr_utitlities.evaluate_model_simple(linear_eval_best_model.predict(np_test[0]), np_test[1], return_dict=True))\n",
        "print(\"Model in last epoch\")\n",
        "print(simclr_utitlities.evaluate_model_simple(linear_evaluation_model.predict(np_test[0]), np_test[1], return_dict=True))\n"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:No training configuration found in the save file, so the model was *not* compiled. Compile it manually.\n",
            "Epoch 1/50\n",
            "21/21 [==============================] - 2s 40ms/step - loss: 1.5929 - categorical_accuracy: 0.5003 - auc: 0.7834 - precision: 0.6352 - recall: 0.3418 - val_loss: 0.8596 - val_categorical_accuracy: 0.7852 - val_auc: 0.9484 - val_precision: 0.7807 - val_recall: 0.5187\n",
            "Epoch 2/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.9289 - categorical_accuracy: 0.7364 - auc: 0.9311 - precision: 0.7709 - recall: 0.4958 - val_loss: 0.6866 - val_categorical_accuracy: 0.8102 - val_auc: 0.9729 - val_precision: 0.8610 - val_recall: 0.5465\n",
            "Epoch 3/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.6992 - categorical_accuracy: 0.7940 - auc: 0.9617 - precision: 0.8519 - recall: 0.6823 - val_loss: 0.8440 - val_categorical_accuracy: 0.6711 - val_auc: 0.9446 - val_precision: 0.7531 - val_recall: 0.5264\n",
            "Epoch 4/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.6927 - categorical_accuracy: 0.7748 - auc: 0.9606 - precision: 0.8347 - recall: 0.7208 - val_loss: 0.6007 - val_categorical_accuracy: 0.8198 - val_auc: 0.9695 - val_precision: 0.8488 - val_recall: 0.7641\n",
            "Epoch 5/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.6244 - categorical_accuracy: 0.8068 - auc: 0.9675 - precision: 0.8467 - recall: 0.7521 - val_loss: 0.4938 - val_categorical_accuracy: 0.8610 - val_auc: 0.9861 - val_precision: 0.9195 - val_recall: 0.7881\n",
            "Epoch 6/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.5112 - categorical_accuracy: 0.8314 - auc: 0.9807 - precision: 0.8813 - recall: 0.7719 - val_loss: 0.5355 - val_categorical_accuracy: 0.8303 - val_auc: 0.9823 - val_precision: 0.9053 - val_recall: 0.7421\n",
            "Epoch 7/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.5238 - categorical_accuracy: 0.8164 - auc: 0.9788 - precision: 0.8721 - recall: 0.7661 - val_loss: 0.4802 - val_categorical_accuracy: 0.8523 - val_auc: 0.9830 - val_precision: 0.8840 - val_recall: 0.7967\n",
            "Epoch 8/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.5092 - categorical_accuracy: 0.8355 - auc: 0.9794 - precision: 0.8712 - recall: 0.7852 - val_loss: 0.5008 - val_categorical_accuracy: 0.8514 - val_auc: 0.9858 - val_precision: 0.9219 - val_recall: 0.7584\n",
            "Epoch 9/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.4607 - categorical_accuracy: 0.8610 - auc: 0.9834 - precision: 0.9010 - recall: 0.7985 - val_loss: 0.4476 - val_categorical_accuracy: 0.8619 - val_auc: 0.9851 - val_precision: 0.8918 - val_recall: 0.8140\n",
            "Epoch 10/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.4185 - categorical_accuracy: 0.8605 - auc: 0.9868 - precision: 0.8975 - recall: 0.8115 - val_loss: 0.4115 - val_categorical_accuracy: 0.9022 - val_auc: 0.9908 - val_precision: 0.9395 - val_recall: 0.8188\n",
            "Epoch 11/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.4129 - categorical_accuracy: 0.8872 - auc: 0.9871 - precision: 0.9222 - recall: 0.8267 - val_loss: 0.4169 - val_categorical_accuracy: 0.9060 - val_auc: 0.9903 - val_precision: 0.9440 - val_recall: 0.8245\n",
            "Epoch 12/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.4217 - categorical_accuracy: 0.8661 - auc: 0.9863 - precision: 0.9054 - recall: 0.8170 - val_loss: 0.4139 - val_categorical_accuracy: 0.8811 - val_auc: 0.9884 - val_precision: 0.9209 - val_recall: 0.8150\n",
            "Epoch 13/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.4076 - categorical_accuracy: 0.8688 - auc: 0.9874 - precision: 0.9080 - recall: 0.8108 - val_loss: 0.3981 - val_categorical_accuracy: 0.8849 - val_auc: 0.9893 - val_precision: 0.9240 - val_recall: 0.8274\n",
            "Epoch 14/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.4004 - categorical_accuracy: 0.8755 - auc: 0.9875 - precision: 0.9136 - recall: 0.8206 - val_loss: 0.4036 - val_categorical_accuracy: 0.9070 - val_auc: 0.9902 - val_precision: 0.9414 - val_recall: 0.8322\n",
            "Epoch 15/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.4071 - categorical_accuracy: 0.8684 - auc: 0.9864 - precision: 0.9033 - recall: 0.8213 - val_loss: 0.3825 - val_categorical_accuracy: 0.9089 - val_auc: 0.9913 - val_precision: 0.9402 - val_recall: 0.8437\n",
            "Epoch 16/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.4012 - categorical_accuracy: 0.8763 - auc: 0.9872 - precision: 0.9084 - recall: 0.8225 - val_loss: 0.4040 - val_categorical_accuracy: 0.8686 - val_auc: 0.9866 - val_precision: 0.8956 - val_recall: 0.8389\n",
            "Epoch 17/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.4014 - categorical_accuracy: 0.8801 - auc: 0.9857 - precision: 0.9053 - recall: 0.8372 - val_loss: 0.3704 - val_categorical_accuracy: 0.9147 - val_auc: 0.9916 - val_precision: 0.9439 - val_recall: 0.8552\n",
            "Epoch 18/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3916 - categorical_accuracy: 0.8838 - auc: 0.9870 - precision: 0.9125 - recall: 0.8355 - val_loss: 0.3929 - val_categorical_accuracy: 0.8744 - val_auc: 0.9874 - val_precision: 0.8997 - val_recall: 0.8514\n",
            "Epoch 19/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3994 - categorical_accuracy: 0.8736 - auc: 0.9865 - precision: 0.9049 - recall: 0.8333 - val_loss: 0.3693 - val_categorical_accuracy: 0.9108 - val_auc: 0.9916 - val_precision: 0.9338 - val_recall: 0.8523\n",
            "Epoch 20/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3648 - categorical_accuracy: 0.8889 - auc: 0.9895 - precision: 0.9221 - recall: 0.8451 - val_loss: 0.3742 - val_categorical_accuracy: 0.8830 - val_auc: 0.9898 - val_precision: 0.9116 - val_recall: 0.8504\n",
            "Epoch 21/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3544 - categorical_accuracy: 0.8876 - auc: 0.9902 - precision: 0.9192 - recall: 0.8445 - val_loss: 0.3558 - val_categorical_accuracy: 0.9012 - val_auc: 0.9917 - val_precision: 0.9336 - val_recall: 0.8495\n",
            "Epoch 22/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3876 - categorical_accuracy: 0.8799 - auc: 0.9875 - precision: 0.9058 - recall: 0.8412 - val_loss: 0.3530 - val_categorical_accuracy: 0.9051 - val_auc: 0.9921 - val_precision: 0.9322 - val_recall: 0.8571\n",
            "Epoch 23/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3460 - categorical_accuracy: 0.8943 - auc: 0.9907 - precision: 0.9215 - recall: 0.8477 - val_loss: 0.3505 - val_categorical_accuracy: 0.8984 - val_auc: 0.9923 - val_precision: 0.9312 - val_recall: 0.8562\n",
            "Epoch 24/50\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.3605 - categorical_accuracy: 0.8955 - auc: 0.9890 - precision: 0.9175 - recall: 0.8448 - val_loss: 0.3469 - val_categorical_accuracy: 0.9195 - val_auc: 0.9928 - val_precision: 0.9468 - val_recall: 0.8696\n",
            "Epoch 25/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3378 - categorical_accuracy: 0.9024 - auc: 0.9910 - precision: 0.9296 - recall: 0.8593 - val_loss: 0.3862 - val_categorical_accuracy: 0.8715 - val_auc: 0.9876 - val_precision: 0.8939 - val_recall: 0.8562\n",
            "Epoch 26/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3858 - categorical_accuracy: 0.8778 - auc: 0.9873 - precision: 0.9060 - recall: 0.8449 - val_loss: 0.3452 - val_categorical_accuracy: 0.9099 - val_auc: 0.9922 - val_precision: 0.9307 - val_recall: 0.8629\n",
            "Epoch 27/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3258 - categorical_accuracy: 0.9051 - auc: 0.9918 - precision: 0.9315 - recall: 0.8637 - val_loss: 0.3385 - val_categorical_accuracy: 0.9156 - val_auc: 0.9933 - val_precision: 0.9440 - val_recall: 0.8725\n",
            "Epoch 28/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3365 - categorical_accuracy: 0.9020 - auc: 0.9910 - precision: 0.9281 - recall: 0.8558 - val_loss: 0.3378 - val_categorical_accuracy: 0.9137 - val_auc: 0.9928 - val_precision: 0.9360 - val_recall: 0.8696\n",
            "Epoch 29/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3271 - categorical_accuracy: 0.9047 - auc: 0.9915 - precision: 0.9253 - recall: 0.8604 - val_loss: 0.3360 - val_categorical_accuracy: 0.9156 - val_auc: 0.9934 - val_precision: 0.9428 - val_recall: 0.8686\n",
            "Epoch 30/50\n",
            "21/21 [==============================] - 0s 20ms/step - loss: 0.3367 - categorical_accuracy: 0.8932 - auc: 0.9909 - precision: 0.9208 - recall: 0.8623 - val_loss: 0.3361 - val_categorical_accuracy: 0.9041 - val_auc: 0.9926 - val_precision: 0.9320 - val_recall: 0.8677\n",
            "Epoch 31/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3397 - categorical_accuracy: 0.8954 - auc: 0.9900 - precision: 0.9256 - recall: 0.8613 - val_loss: 0.3251 - val_categorical_accuracy: 0.9185 - val_auc: 0.9938 - val_precision: 0.9478 - val_recall: 0.8706\n",
            "Epoch 32/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3406 - categorical_accuracy: 0.8939 - auc: 0.9902 - precision: 0.9209 - recall: 0.8588 - val_loss: 0.3340 - val_categorical_accuracy: 0.9089 - val_auc: 0.9932 - val_precision: 0.9389 - val_recall: 0.8686\n",
            "Epoch 33/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3168 - categorical_accuracy: 0.9115 - auc: 0.9916 - precision: 0.9315 - recall: 0.8736 - val_loss: 0.3304 - val_categorical_accuracy: 0.9012 - val_auc: 0.9927 - val_precision: 0.9321 - val_recall: 0.8686\n",
            "Epoch 34/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3164 - categorical_accuracy: 0.9055 - auc: 0.9919 - precision: 0.9328 - recall: 0.8713 - val_loss: 0.3624 - val_categorical_accuracy: 0.8897 - val_auc: 0.9900 - val_precision: 0.9140 - val_recall: 0.8562\n",
            "Epoch 35/50\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.3541 - categorical_accuracy: 0.8928 - auc: 0.9884 - precision: 0.9183 - recall: 0.8591 - val_loss: 0.3339 - val_categorical_accuracy: 0.9041 - val_auc: 0.9922 - val_precision: 0.9277 - val_recall: 0.8734\n",
            "Epoch 36/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3338 - categorical_accuracy: 0.8973 - auc: 0.9904 - precision: 0.9194 - recall: 0.8641 - val_loss: 0.3317 - val_categorical_accuracy: 0.8955 - val_auc: 0.9922 - val_precision: 0.9234 - val_recall: 0.8667\n",
            "Epoch 37/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3021 - categorical_accuracy: 0.9045 - auc: 0.9928 - precision: 0.9288 - recall: 0.8743 - val_loss: 0.3272 - val_categorical_accuracy: 0.9156 - val_auc: 0.9932 - val_precision: 0.9356 - val_recall: 0.8773\n",
            "Epoch 38/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3253 - categorical_accuracy: 0.9052 - auc: 0.9907 - precision: 0.9273 - recall: 0.8676 - val_loss: 0.3254 - val_categorical_accuracy: 0.9118 - val_auc: 0.9934 - val_precision: 0.9337 - val_recall: 0.8782\n",
            "Epoch 39/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3147 - categorical_accuracy: 0.9059 - auc: 0.9917 - precision: 0.9312 - recall: 0.8730 - val_loss: 0.3165 - val_categorical_accuracy: 0.9195 - val_auc: 0.9942 - val_precision: 0.9472 - val_recall: 0.8763\n",
            "Epoch 40/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3031 - categorical_accuracy: 0.9105 - auc: 0.9928 - precision: 0.9359 - recall: 0.8744 - val_loss: 0.3239 - val_categorical_accuracy: 0.9214 - val_auc: 0.9943 - val_precision: 0.9470 - val_recall: 0.8734\n",
            "Epoch 41/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3174 - categorical_accuracy: 0.9045 - auc: 0.9914 - precision: 0.9314 - recall: 0.8693 - val_loss: 0.3169 - val_categorical_accuracy: 0.9185 - val_auc: 0.9938 - val_precision: 0.9415 - val_recall: 0.8802\n",
            "Epoch 42/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3119 - categorical_accuracy: 0.9061 - auc: 0.9916 - precision: 0.9330 - recall: 0.8730 - val_loss: 0.3375 - val_categorical_accuracy: 0.9003 - val_auc: 0.9918 - val_precision: 0.9204 - val_recall: 0.8763\n",
            "Epoch 43/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3019 - categorical_accuracy: 0.9112 - auc: 0.9923 - precision: 0.9340 - recall: 0.8822 - val_loss: 0.3149 - val_categorical_accuracy: 0.9128 - val_auc: 0.9938 - val_precision: 0.9387 - val_recall: 0.8802\n",
            "Epoch 44/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3162 - categorical_accuracy: 0.9092 - auc: 0.9914 - precision: 0.9312 - recall: 0.8713 - val_loss: 0.3285 - val_categorical_accuracy: 0.9041 - val_auc: 0.9925 - val_precision: 0.9273 - val_recall: 0.8811\n",
            "Epoch 45/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3045 - categorical_accuracy: 0.9060 - auc: 0.9922 - precision: 0.9292 - recall: 0.8768 - val_loss: 0.3212 - val_categorical_accuracy: 0.9214 - val_auc: 0.9941 - val_precision: 0.9455 - val_recall: 0.8811\n",
            "Epoch 46/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3050 - categorical_accuracy: 0.9089 - auc: 0.9920 - precision: 0.9310 - recall: 0.8791 - val_loss: 0.3134 - val_categorical_accuracy: 0.9156 - val_auc: 0.9940 - val_precision: 0.9398 - val_recall: 0.8830\n",
            "Epoch 47/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3067 - categorical_accuracy: 0.9097 - auc: 0.9917 - precision: 0.9322 - recall: 0.8840 - val_loss: 0.3156 - val_categorical_accuracy: 0.9022 - val_auc: 0.9932 - val_precision: 0.9318 - val_recall: 0.8782\n",
            "Epoch 48/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3144 - categorical_accuracy: 0.9057 - auc: 0.9910 - precision: 0.9249 - recall: 0.8702 - val_loss: 0.3213 - val_categorical_accuracy: 0.9128 - val_auc: 0.9931 - val_precision: 0.9323 - val_recall: 0.8840\n",
            "Epoch 49/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3150 - categorical_accuracy: 0.9051 - auc: 0.9906 - precision: 0.9269 - recall: 0.8763 - val_loss: 0.3209 - val_categorical_accuracy: 0.9156 - val_auc: 0.9937 - val_precision: 0.9403 - val_recall: 0.8754\n",
            "Epoch 50/50\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3084 - categorical_accuracy: 0.9049 - auc: 0.9923 - precision: 0.9269 - recall: 0.8704 - val_loss: 0.3088 - val_categorical_accuracy: 0.9185 - val_auc: 0.9941 - val_precision: 0.9418 - val_recall: 0.8840\n",
            "Model with lowest validation Loss:\n",
            "{'Confusion Matrix': array([[360,   0,   0,   0,   0,   0],\n",
            "       [ 75, 258,   2,   0,   0,   0],\n",
            "       [  0,   0, 273,  51,   7,   0],\n",
            "       [  0,   3,   9, 136,   0,   0],\n",
            "       [  0,   1,  57,   8,  46,   0],\n",
            "       [  0,   0,   6,   1,   1, 125]]), 'F1 Macro': 0.814866418310411, 'F1 Micro': 0.8442565186751233, 'F1 Weighted': 0.8386994437133516, 'Precision': 0.8574653250077394, 'Recall': 0.8107342493869965, 'Kappa': 0.8043250784792466}\n",
            "Model in last epoch\n",
            "{'Confusion Matrix': array([[360,   0,   0,   0,   0,   0],\n",
            "       [ 75, 258,   2,   0,   0,   0],\n",
            "       [  0,   0, 273,  51,   7,   0],\n",
            "       [  0,   3,   9, 136,   0,   0],\n",
            "       [  0,   1,  57,   8,  46,   0],\n",
            "       [  0,   0,   6,   1,   1, 125]]), 'F1 Macro': 0.814866418310411, 'F1 Micro': 0.8442565186751233, 'F1 Weighted': 0.8386994437133516, 'Precision': 0.8574653250077394, 'Recall': 0.8107342493869965, 'Kappa': 0.8043250784792466}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c54Gf4jKP53U"
      },
      "source": [
        "### Full HAR Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ty-ZQvLvP53U",
        "outputId": "59d9f85a-97a5-463d-9acc-a9b1f68183c7"
      },
      "source": [
        "\n",
        "total_epochs = 50\n",
        "batch_size = 200\n",
        "tag = \"full_eval\"\n",
        "\n",
        "simclr_model = tf.keras.models.load_model(simclr_model_save_path)\n",
        "full_evaluation_model = simclr_models.create_full_classification_model_from_base_model(simclr_model, output_shape, model_name=\"TPN\", intermediate_layer=7, last_freeze_layer=4)\n",
        "\n",
        "full_eval_best_model_file_name = f\"{working_directory}{start_time_str}_simclr_{tag}.hdf5\"\n",
        "best_model_callback = tf.keras.callbacks.ModelCheckpoint(full_eval_best_model_file_name,\n",
        "    monitor='val_loss', mode='min', save_best_only=True, save_weights_only=False, verbose=0\n",
        ")\n",
        "\n",
        "training_history = full_evaluation_model.fit(\n",
        "    x = np_train[0],\n",
        "    y = np_train[1],\n",
        "    batch_size=batch_size,\n",
        "    shuffle=True,\n",
        "    epochs=total_epochs,\n",
        "    callbacks=[best_model_callback],\n",
        "    validation_data=np_val\n",
        ")\n",
        "\n",
        "full_eval_best_model = tf.keras.models.load_model(full_eval_best_model_file_name)\n",
        "\n",
        "print(\"Model with lowest validation Loss:\")\n",
        "print(simclr_utitlities.evaluate_model_simple(full_eval_best_model.predict(np_test[0]), np_test[1], return_dict=True))\n",
        "print(\"Model in last epoch\")\n",
        "print(simclr_utitlities.evaluate_model_simple(full_evaluation_model.predict(np_test[0]), np_test[1], return_dict=True))\n"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:No training configuration found in the save file, so the model was *not* compiled. Compile it manually.\n",
            "Epoch 1/50\n",
            "21/21 [==============================] - 2s 41ms/step - loss: 1.2525 - categorical_accuracy: 0.5996 - auc: 0.8578 - precision: 0.7691 - recall: 0.4609 - val_loss: 0.4262 - val_categorical_accuracy: 0.8648 - val_auc: 0.9898 - val_precision: 0.9042 - val_recall: 0.8140\n",
            "Epoch 2/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.3613 - categorical_accuracy: 0.8753 - auc: 0.9895 - precision: 0.9097 - recall: 0.8363 - val_loss: 0.2751 - val_categorical_accuracy: 0.9367 - val_auc: 0.9974 - val_precision: 0.9578 - val_recall: 0.8926\n",
            "Epoch 3/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.2203 - categorical_accuracy: 0.9421 - auc: 0.9970 - precision: 0.9571 - recall: 0.9155 - val_loss: 0.1963 - val_categorical_accuracy: 0.9674 - val_auc: 0.9986 - val_precision: 0.9800 - val_recall: 0.9377\n",
            "Epoch 4/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.1621 - categorical_accuracy: 0.9547 - auc: 0.9985 - precision: 0.9630 - recall: 0.9435 - val_loss: 0.1460 - val_categorical_accuracy: 0.9722 - val_auc: 0.9992 - val_precision: 0.9776 - val_recall: 0.9636\n",
            "Epoch 5/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.1428 - categorical_accuracy: 0.9622 - auc: 0.9983 - precision: 0.9676 - recall: 0.9537 - val_loss: 0.1380 - val_categorical_accuracy: 0.9722 - val_auc: 0.9990 - val_precision: 0.9750 - val_recall: 0.9703\n",
            "Epoch 6/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.1041 - categorical_accuracy: 0.9747 - auc: 0.9993 - precision: 0.9792 - recall: 0.9687 - val_loss: 0.1101 - val_categorical_accuracy: 0.9818 - val_auc: 0.9996 - val_precision: 0.9855 - val_recall: 0.9789\n",
            "Epoch 7/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0827 - categorical_accuracy: 0.9845 - auc: 0.9997 - precision: 0.9874 - recall: 0.9790 - val_loss: 0.0969 - val_categorical_accuracy: 0.9818 - val_auc: 0.9996 - val_precision: 0.9837 - val_recall: 0.9818\n",
            "Epoch 8/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0678 - categorical_accuracy: 0.9882 - auc: 0.9998 - precision: 0.9888 - recall: 0.9865 - val_loss: 0.0906 - val_categorical_accuracy: 0.9770 - val_auc: 0.9997 - val_precision: 0.9817 - val_recall: 0.9760\n",
            "Epoch 9/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0710 - categorical_accuracy: 0.9845 - auc: 0.9998 - precision: 0.9851 - recall: 0.9814 - val_loss: 0.1075 - val_categorical_accuracy: 0.9808 - val_auc: 0.9992 - val_precision: 0.9836 - val_recall: 0.9799\n",
            "Epoch 10/50\n",
            "21/21 [==============================] - 0s 20ms/step - loss: 0.0616 - categorical_accuracy: 0.9882 - auc: 0.9998 - precision: 0.9888 - recall: 0.9857 - val_loss: 0.0756 - val_categorical_accuracy: 0.9885 - val_auc: 0.9998 - val_precision: 0.9885 - val_recall: 0.9885\n",
            "Epoch 11/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0546 - categorical_accuracy: 0.9906 - auc: 0.9999 - precision: 0.9909 - recall: 0.9897 - val_loss: 0.0702 - val_categorical_accuracy: 0.9885 - val_auc: 0.9999 - val_precision: 0.9904 - val_recall: 0.9866\n",
            "Epoch 12/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0587 - categorical_accuracy: 0.9852 - auc: 0.9999 - precision: 0.9877 - recall: 0.9838 - val_loss: 0.0901 - val_categorical_accuracy: 0.9799 - val_auc: 0.9996 - val_precision: 0.9808 - val_recall: 0.9779\n",
            "Epoch 13/50\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.0560 - categorical_accuracy: 0.9887 - auc: 0.9998 - precision: 0.9899 - recall: 0.9866 - val_loss: 0.0622 - val_categorical_accuracy: 0.9866 - val_auc: 0.9999 - val_precision: 0.9904 - val_recall: 0.9856\n",
            "Epoch 14/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0408 - categorical_accuracy: 0.9932 - auc: 1.0000 - precision: 0.9946 - recall: 0.9930 - val_loss: 0.0770 - val_categorical_accuracy: 0.9895 - val_auc: 0.9998 - val_precision: 0.9894 - val_recall: 0.9885\n",
            "Epoch 15/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0430 - categorical_accuracy: 0.9927 - auc: 1.0000 - precision: 0.9928 - recall: 0.9927 - val_loss: 0.0608 - val_categorical_accuracy: 0.9885 - val_auc: 0.9999 - val_precision: 0.9885 - val_recall: 0.9856\n",
            "Epoch 16/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0421 - categorical_accuracy: 0.9922 - auc: 0.9999 - precision: 0.9922 - recall: 0.9920 - val_loss: 0.0578 - val_categorical_accuracy: 0.9875 - val_auc: 1.0000 - val_precision: 0.9923 - val_recall: 0.9875\n",
            "Epoch 17/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0343 - categorical_accuracy: 0.9958 - auc: 1.0000 - precision: 0.9965 - recall: 0.9958 - val_loss: 0.0527 - val_categorical_accuracy: 0.9914 - val_auc: 0.9999 - val_precision: 0.9923 - val_recall: 0.9914\n",
            "Epoch 18/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0301 - categorical_accuracy: 0.9975 - auc: 1.0000 - precision: 0.9975 - recall: 0.9975 - val_loss: 0.0485 - val_categorical_accuracy: 0.9904 - val_auc: 1.0000 - val_precision: 0.9914 - val_recall: 0.9895\n",
            "Epoch 19/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0322 - categorical_accuracy: 0.9974 - auc: 1.0000 - precision: 0.9974 - recall: 0.9967 - val_loss: 0.0730 - val_categorical_accuracy: 0.9875 - val_auc: 0.9998 - val_precision: 0.9885 - val_recall: 0.9847\n",
            "Epoch 20/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0430 - categorical_accuracy: 0.9929 - auc: 0.9999 - precision: 0.9932 - recall: 0.9928 - val_loss: 0.0565 - val_categorical_accuracy: 0.9895 - val_auc: 0.9999 - val_precision: 0.9895 - val_recall: 0.9895\n",
            "Epoch 21/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0358 - categorical_accuracy: 0.9946 - auc: 1.0000 - precision: 0.9946 - recall: 0.9940 - val_loss: 0.0608 - val_categorical_accuracy: 0.9923 - val_auc: 0.9999 - val_precision: 0.9952 - val_recall: 0.9923\n",
            "Epoch 22/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0329 - categorical_accuracy: 0.9959 - auc: 1.0000 - precision: 0.9960 - recall: 0.9959 - val_loss: 0.0776 - val_categorical_accuracy: 0.9904 - val_auc: 0.9997 - val_precision: 0.9914 - val_recall: 0.9904\n",
            "Epoch 23/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0394 - categorical_accuracy: 0.9955 - auc: 0.9996 - precision: 0.9957 - recall: 0.9954 - val_loss: 0.0604 - val_categorical_accuracy: 0.9895 - val_auc: 0.9999 - val_precision: 0.9904 - val_recall: 0.9885\n",
            "Epoch 24/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0429 - categorical_accuracy: 0.9908 - auc: 0.9999 - precision: 0.9917 - recall: 0.9908 - val_loss: 0.0658 - val_categorical_accuracy: 0.9895 - val_auc: 0.9998 - val_precision: 0.9894 - val_recall: 0.9885\n",
            "Epoch 25/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0456 - categorical_accuracy: 0.9933 - auc: 0.9999 - precision: 0.9933 - recall: 0.9933 - val_loss: 0.0676 - val_categorical_accuracy: 0.9895 - val_auc: 0.9998 - val_precision: 0.9904 - val_recall: 0.9885\n",
            "Epoch 26/50\n",
            "21/21 [==============================] - 0s 20ms/step - loss: 0.0292 - categorical_accuracy: 0.9969 - auc: 1.0000 - precision: 0.9972 - recall: 0.9964 - val_loss: 0.0596 - val_categorical_accuracy: 0.9933 - val_auc: 0.9999 - val_precision: 0.9933 - val_recall: 0.9914\n",
            "Epoch 27/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0240 - categorical_accuracy: 0.9994 - auc: 1.0000 - precision: 0.9994 - recall: 0.9994 - val_loss: 0.0475 - val_categorical_accuracy: 0.9952 - val_auc: 1.0000 - val_precision: 0.9952 - val_recall: 0.9933\n",
            "Epoch 28/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0233 - categorical_accuracy: 0.9999 - auc: 1.0000 - precision: 0.9999 - recall: 0.9995 - val_loss: 0.0517 - val_categorical_accuracy: 0.9933 - val_auc: 0.9999 - val_precision: 0.9933 - val_recall: 0.9914\n",
            "Epoch 29/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0245 - categorical_accuracy: 0.9989 - auc: 1.0000 - precision: 0.9989 - recall: 0.9989 - val_loss: 0.0558 - val_categorical_accuracy: 0.9923 - val_auc: 0.9999 - val_precision: 0.9933 - val_recall: 0.9923\n",
            "Epoch 30/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0229 - categorical_accuracy: 0.9996 - auc: 1.0000 - precision: 0.9996 - recall: 0.9994 - val_loss: 0.0516 - val_categorical_accuracy: 0.9942 - val_auc: 0.9999 - val_precision: 0.9942 - val_recall: 0.9933\n",
            "Epoch 31/50\n",
            "21/21 [==============================] - 0s 20ms/step - loss: 0.0239 - categorical_accuracy: 0.9987 - auc: 1.0000 - precision: 0.9987 - recall: 0.9985 - val_loss: 0.0536 - val_categorical_accuracy: 0.9933 - val_auc: 0.9994 - val_precision: 0.9942 - val_recall: 0.9933\n",
            "Epoch 32/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0237 - categorical_accuracy: 0.9990 - auc: 1.0000 - precision: 0.9990 - recall: 0.9990 - val_loss: 0.0484 - val_categorical_accuracy: 0.9933 - val_auc: 0.9999 - val_precision: 0.9942 - val_recall: 0.9933\n",
            "Epoch 33/50\n",
            "21/21 [==============================] - 0s 20ms/step - loss: 0.0224 - categorical_accuracy: 0.9996 - auc: 1.0000 - precision: 0.9996 - recall: 0.9994 - val_loss: 0.0626 - val_categorical_accuracy: 0.9933 - val_auc: 0.9998 - val_precision: 0.9942 - val_recall: 0.9933\n",
            "Epoch 34/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0225 - categorical_accuracy: 0.9998 - auc: 1.0000 - precision: 0.9998 - recall: 0.9998 - val_loss: 0.0504 - val_categorical_accuracy: 0.9952 - val_auc: 0.9999 - val_precision: 0.9952 - val_recall: 0.9942\n",
            "Epoch 35/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0212 - categorical_accuracy: 0.9998 - auc: 1.0000 - precision: 0.9998 - recall: 0.9998 - val_loss: 0.0492 - val_categorical_accuracy: 0.9952 - val_auc: 0.9994 - val_precision: 0.9962 - val_recall: 0.9942\n",
            "Epoch 36/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0209 - categorical_accuracy: 0.9996 - auc: 1.0000 - precision: 0.9996 - recall: 0.9996 - val_loss: 0.0446 - val_categorical_accuracy: 0.9962 - val_auc: 0.9999 - val_precision: 0.9971 - val_recall: 0.9962\n",
            "Epoch 37/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0211 - categorical_accuracy: 0.9994 - auc: 1.0000 - precision: 0.9994 - recall: 0.9994 - val_loss: 0.0499 - val_categorical_accuracy: 0.9942 - val_auc: 0.9993 - val_precision: 0.9942 - val_recall: 0.9942\n",
            "Epoch 38/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0207 - categorical_accuracy: 0.9998 - auc: 1.0000 - precision: 0.9998 - recall: 0.9998 - val_loss: 0.0490 - val_categorical_accuracy: 0.9952 - val_auc: 0.9999 - val_precision: 0.9952 - val_recall: 0.9952\n",
            "Epoch 39/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0217 - categorical_accuracy: 0.9999 - auc: 1.0000 - precision: 0.9999 - recall: 0.9993 - val_loss: 0.0542 - val_categorical_accuracy: 0.9952 - val_auc: 0.9998 - val_precision: 0.9952 - val_recall: 0.9942\n",
            "Epoch 40/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0215 - categorical_accuracy: 0.9992 - auc: 1.0000 - precision: 0.9992 - recall: 0.9992 - val_loss: 0.0473 - val_categorical_accuracy: 0.9942 - val_auc: 0.9994 - val_precision: 0.9952 - val_recall: 0.9942\n",
            "Epoch 41/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0199 - categorical_accuracy: 0.9999 - auc: 1.0000 - precision: 0.9999 - recall: 0.9999 - val_loss: 0.0457 - val_categorical_accuracy: 0.9933 - val_auc: 0.9994 - val_precision: 0.9933 - val_recall: 0.9923\n",
            "Epoch 42/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0239 - categorical_accuracy: 0.9982 - auc: 1.0000 - precision: 0.9982 - recall: 0.9982 - val_loss: 0.0526 - val_categorical_accuracy: 0.9952 - val_auc: 0.9999 - val_precision: 0.9952 - val_recall: 0.9952\n",
            "Epoch 43/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0212 - categorical_accuracy: 0.9990 - auc: 1.0000 - precision: 0.9990 - recall: 0.9990 - val_loss: 0.0465 - val_categorical_accuracy: 0.9952 - val_auc: 0.9995 - val_precision: 0.9952 - val_recall: 0.9952\n",
            "Epoch 44/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0207 - categorical_accuracy: 0.9992 - auc: 1.0000 - precision: 0.9992 - recall: 0.9992 - val_loss: 0.0466 - val_categorical_accuracy: 0.9952 - val_auc: 0.9994 - val_precision: 0.9962 - val_recall: 0.9952\n",
            "Epoch 45/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0210 - categorical_accuracy: 0.9994 - auc: 1.0000 - precision: 0.9994 - recall: 0.9994 - val_loss: 0.0473 - val_categorical_accuracy: 0.9942 - val_auc: 0.9994 - val_precision: 0.9942 - val_recall: 0.9933\n",
            "Epoch 46/50\n",
            "21/21 [==============================] - 0s 20ms/step - loss: 0.0221 - categorical_accuracy: 0.9982 - auc: 1.0000 - precision: 0.9982 - recall: 0.9982 - val_loss: 0.0462 - val_categorical_accuracy: 0.9952 - val_auc: 0.9999 - val_precision: 0.9952 - val_recall: 0.9952\n",
            "Epoch 47/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0213 - categorical_accuracy: 0.9991 - auc: 1.0000 - precision: 0.9991 - recall: 0.9991 - val_loss: 0.0550 - val_categorical_accuracy: 0.9923 - val_auc: 0.9993 - val_precision: 0.9923 - val_recall: 0.9914\n",
            "Epoch 48/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0480 - categorical_accuracy: 0.9896 - auc: 0.9997 - precision: 0.9904 - recall: 0.9893 - val_loss: 0.0901 - val_categorical_accuracy: 0.9885 - val_auc: 0.9995 - val_precision: 0.9885 - val_recall: 0.9875\n",
            "Epoch 49/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0651 - categorical_accuracy: 0.9866 - auc: 0.9991 - precision: 0.9869 - recall: 0.9866 - val_loss: 0.0677 - val_categorical_accuracy: 0.9885 - val_auc: 0.9998 - val_precision: 0.9885 - val_recall: 0.9885\n",
            "Epoch 50/50\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 0.0297 - categorical_accuracy: 0.9970 - auc: 1.0000 - precision: 0.9971 - recall: 0.9970 - val_loss: 0.0468 - val_categorical_accuracy: 0.9952 - val_auc: 0.9999 - val_precision: 0.9952 - val_recall: 0.9942\n",
            "Model with lowest validation Loss:\n",
            "{'Confusion Matrix': array([[360,   0,   0,   0,   0,   0],\n",
            "       [  1, 332,   1,   0,   1,   0],\n",
            "       [  0,   0, 288,   6,  37,   0],\n",
            "       [  0,   0,   2, 135,  11,   0],\n",
            "       [  0,   0,   9,  10,  93,   0],\n",
            "       [  0,   0,   0,   2,   2, 129]]), 'F1 Macro': 0.9192095574274819, 'F1 Micro': 0.9422128259337562, 'F1 Weighted': 0.9444995471991302, 'Precision': 0.9142360319012185, 'Recall': 0.9289299212683119, 'Kappa': 0.9280475799430109}\n",
            "Model in last epoch\n",
            "{'Confusion Matrix': array([[360,   0,   0,   0,   0,   0],\n",
            "       [  2, 331,   0,   2,   0,   0],\n",
            "       [  0,   0, 264,  19,  48,   0],\n",
            "       [  0,   0,   0, 143,   5,   0],\n",
            "       [  1,   0,   6,  15,  90,   0],\n",
            "       [  0,   0,   1,   2,   1, 129]]), 'F1 Macro': 0.9040132941070583, 'F1 Micro': 0.9281183932346723, 'F1 Weighted': 0.9303611733731109, 'Precision': 0.8968267545843807, 'Recall': 0.9208925399802089, 'Kappa': 0.9107779329832785}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ohMg5fypP53V"
      },
      "source": [
        "## Extra: t-SNE Plots"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Gk5Lg-8tP53V"
      },
      "source": [
        "### Parameters"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OnIMMgf5P53V"
      },
      "source": [
        "# Select a model from which the intermediate representations are extracted\n",
        "target_model = simclr_model \n",
        "perplexity = 30.0\n"
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oRm0pPJRP53V"
      },
      "source": [
        "### t-SNE Representations"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EJP3ZM9OP53V",
        "outputId": "3807ffa8-982c-47e5-a7dd-2e4cd1492edf"
      },
      "source": [
        "intermediate_model = simclr_models.extract_intermediate_model_from_base_model(target_model, intermediate_layer=7)\n",
        "intermediate_model.summary()\n",
        "\n",
        "embeddings = intermediate_model.predict(np_test[0], batch_size=600)\n",
        "tsne_model = sklearn.manifold.TSNE(perplexity=perplexity, verbose=1, random_state=42)\n",
        "tsne_projections = tsne_model.fit_transform(embeddings)\n",
        "\n"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"base_model_simclr_layer_7\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input (InputLayer)           [(None, 400, 3)]          0         \n",
            "_________________________________________________________________\n",
            "conv1d (Conv1D)              (None, 377, 32)           2336      \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 377, 32)           0         \n",
            "_________________________________________________________________\n",
            "conv1d_1 (Conv1D)            (None, 362, 64)           32832     \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 362, 64)           0         \n",
            "_________________________________________________________________\n",
            "conv1d_2 (Conv1D)            (None, 355, 96)           49248     \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 355, 96)           0         \n",
            "_________________________________________________________________\n",
            "global_max_pooling1d (Global (None, 96)                0         \n",
            "=================================================================\n",
            "Total params: 84,416\n",
            "Trainable params: 49,248\n",
            "Non-trainable params: 35,168\n",
            "_________________________________________________________________\n",
            "[t-SNE] Computing 91 nearest neighbors...\n",
            "[t-SNE] Indexed 1419 samples in 0.012s...\n",
            "[t-SNE] Computed neighbors for 1419 samples in 0.139s...\n",
            "[t-SNE] Computed conditional probabilities for sample 1000 / 1419\n",
            "[t-SNE] Computed conditional probabilities for sample 1419 / 1419\n",
            "[t-SNE] Mean sigma: 0.296585\n",
            "[t-SNE] KL divergence after 250 iterations with early exaggeration: 51.449738\n",
            "[t-SNE] KL divergence after 1000 iterations: 0.224750\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lioCbwhKP53W"
      },
      "source": [
        "### Plotting"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 585
        },
        "id": "M6R-sUCYP53W",
        "outputId": "31147479-0489-4e5e-e182-d9446076b632"
      },
      "source": [
        "labels_argmax = np.argmax(np_test[1], axis=1)\n",
        "unique_labels = np.unique(labels_argmax)\n",
        "\n",
        "plt.figure(figsize=(16,8))\n",
        "graph = sns.scatterplot(\n",
        "    x=tsne_projections[:,0], y=tsne_projections[:,1],\n",
        "    hue=labels_argmax,\n",
        "    palette=sns.color_palette(\"hsv\", len(unique_labels)),\n",
        "    s=50,\n",
        "    alpha=1.0,\n",
        "    rasterized=True\n",
        ")\n",
        "plt.xticks([], [])\n",
        "plt.yticks([], [])\n",
        "\n",
        "\n",
        "plt.legend(loc='lower left', bbox_to_anchor=(0.25, -0.3), ncol=2)\n",
        "legend = graph.legend_\n",
        "for j, label in enumerate(unique_labels):\n",
        "    legend.get_texts()[j].set_text(label_list_full_name[label]) "
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SAiOv0b8P53W"
      },
      "source": [
        "### Custom Color maps (Optional)\n",
        "\n",
        "This section can be run to produce plots where semantically similar classes share similar colors. This requires the definition of a custom mapping of classes to colors."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c-_R4GFAP53W"
      },
      "source": [
        "# This is used to select colors for labels which are close to each other\n",
        "# Each pair corresponds to one label class\n",
        "# i.e. ['null', 'sitting', 'standing', 'walking', 'walking upstairs', 'walking downstairs', 'jogging']\n",
        "# The first number determines the color map, and the second determines its value along the color map\n",
        "# So 'sitting', 'standing' will share similar colors, and 'walking', 'walking upstairs', 'walking downstairs' will share another set of similar colors\n",
        "label_color_spectrum = [(0, 0), (1, 0), (1, 1), (2, 0), (2, 1), (2, 2), (3, 0)] \n",
        "\n",
        "# This step generates a list of colors for different categories of activities\n",
        "# Here we assume 5 categories, and 5 different intesities within each category\n",
        "major_colors = ['cool', 'Blues', 'Greens', 'Oranges', 'Purples']\n",
        "color_map_base = dict (\n",
        "    [((i, j), color) for i, major_color in enumerate(major_colors) for j, color in enumerate(reversed(sns.color_palette(major_color, 5))) ]\n",
        ")\n",
        "color_palette = np.array([color_map_base[color_index] for color_index in label_color_spectrum])"
      ],
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 585
        },
        "id": "ifVihJbFP53X",
        "outputId": "ea759bf7-d627-443e-9abd-bd1f9cb97203"
      },
      "source": [
        "# This selects the appropriate number of colors to be used in the plot\n",
        "labels_argmax = np.argmax(np_test[1], axis=1)\n",
        "unique_labels = np.unique(labels_argmax)\n",
        "\n",
        "plt.figure(figsize=(16,8))\n",
        "graph = sns.scatterplot(\n",
        "    x=tsne_projections[:,0], y=tsne_projections[:,1],\n",
        "    hue=labels_argmax,\n",
        "    palette=list(color_palette[unique_labels]),\n",
        "    s=50,\n",
        "    alpha=1.0,\n",
        "    rasterized=True\n",
        ")\n",
        "plt.xticks([], [])\n",
        "plt.yticks([], [])\n",
        "\n",
        "\n",
        "plt.legend(loc='lower left', bbox_to_anchor=(0.25, -0.3), ncol=2)\n",
        "legend = graph.legend_\n",
        "for j, label in enumerate(unique_labels):\n",
        "    legend.get_texts()[j].set_text(label_list_full_name[label]) \n"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rvGEU5UAP53X"
      },
      "source": [],
      "execution_count": null,
      "outputs": []
    }
  ]
}