{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Image Classification - Rock, Paper, Scissors\n", "\n", "This tutorial describes how to use the MLTK to develop a image classification machine learning model to detect the hand gestures: \n", "- __Rock__\n", "- __Paper__\n", "- __Scissors__\n", "- __Unknown__" ] }, { "attachments": {}, "cell_type": "markdown", "metadata": {}, "source": [ "## Demo Video\n", "\n", "The following is a video of the demo described in this tutorial:\n", "\n", "" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Quick Links\n", "\n", "- [GitHub Source](https://github.com/SiliconLabs/mltk/blob/master/mltk/tutorials/image_classification.ipynb) - View this tutorial on Github\n", "- [Run on Colab](https://colab.research.google.com/github/siliconlabs/mltk/blob/master/mltk/tutorials/image_classification.ipynb) - Run this tutorial on Google Colab\n", "- [Train in the \"Cloud\"](../../mltk/tutorials/cloud_training_with_vast_ai.md) - _Vastly_ improve training times by training this model in the \"cloud\"\n", "- [C++ Example Application](../../docs/cpp_development/examples/image_classifier.md) - View this tutorial's associated C++ example application\n", "- [Machine Learning Model](../../docs/python_api/models/siliconlabs/rock_paper_scissors.md) - View this tutorial's associated machine learning model" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Overview\n", "\n", "### Objectives\n", "\n", "After completing this tutorial, you will have:\n", "1. A better understanding of how image classification machine learning models work\n", "2. A better understanding of how labeled datasets are created\n", "3. All of the tools needed to develop your own image classification model\n", "4. A working demo to detect the hand gestures: \"Rock\", \"Paper\", \"Scissors\"\n", "\n", "### Content\n", "\n", "This tutorial is divided into the following sections:\n", "1. [Overview of classification machine learning models](#classification-machine-learning-models-overview)\n", "2. [Creating a labeled dataset](#creating-a-labeled-dataset)\n", "3. [Creating the model specification](#creating-the-model-specification)\n", "4. [Note about model parameters](#model-parameters)\n", "5. [Summarizing the model](#model-visualization)\n", "6. [Visualizing the model graph](#model-visualization)\n", "7. [Profiling the model](#model-profiler)\n", "8. [Training the model](#model-training)\n", "9. [Evaluating the model](#model-evaluation)\n", "10. [Testing the model](#model-testing)\n", "11. [Deploying the model to an embedded device](#deploying-the-model)\n", "\n", "### Running this tutorial from a notebook\n", "\n", "For documentation purposes, this tutorial was designed to run within a [Jupyter Notebook](https://jupyter.org). \n", "The notebook can either run locally on your PC _or_ on a remote server like [Google Colab](https://colab.research.google.com/notebooks/welcome.ipynb). \n", "\n", "- Refer to the [Notebook Examples Guide](../../docs/guides/notebook_examples_guide.md) for more details\n", "- Click here: [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/siliconlabs/mltk/blob/master/mltk/tutorials/image_classification.ipynb) to run this tutorial interactively in your browser\n", "\n", "__NOTE:__ Some of the following sections require this tutorial to be running locally with a supported embedded platform connected.\n", "\n", "### Running this tutorial from the command-line\n", "\n", "While this tutorial uses a [Jupyter Notebook](https://jupyter.org), \n", "the recommended approach is to use your favorite text editor and standard command terminal, no Jupyter Notebook required. \n", "\n", "See the [Standard Python Package Installation](https://siliconlabs.github.io/mltk/docs/installation.html#standard-python-package) guide for more details on how to enable the `mltk` command in your local terminal.\n", "\n", "In this mode, when you encounter a `!mltk` command in this tutorial, the command should actually run in your local terminal (excluding the `!`)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Required Hardware\n", "\n", "Some parts of the tutorial requires a supported development board and the [ArduCAM](https://www.arducam.com/product/arducam-2mp-spi-camera-b0067-arduino) camera module.\n", "\n", "See the [Hardware Setup](https://siliconlabs.github.io/mltk/docs/cpp_development/examples/image_classifier.html#hardware-setup) section of the Image Classification C++ example application for details on how to connect the camera to the development board. \n", "\n", "__NOTE:__ Only the camera needs to be connected to the development board. You do _not_ need to build the C++ application from source for this tutorial." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Install MLTK Python Package\n", "\n", "Before using the MLTK, it must first be installed. \n", "See the [Installation Guide](../../docs/installation.md) for more details." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "!pip install --upgrade silabs-mltk" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "All MLTK modeling operations are accessible via the `mltk` command. \n", "Run the command `mltk --help` to ensure it is working. \n", "__NOTE:__ The exclamation point `!` tells the Notebook to run a shell command, it is not required in a [standard terminal](https://siliconlabs.github.io/mltk/docs/installation.html#standard-python-package)" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Usage: mltk [OPTIONS] COMMAND [ARGS]...\n", "\n", " Silicon Labs Machine Learning Toolkit\n", "\n", " This is a Python package with command-line utilities and scripts to aid the\n", " development of machine learning models for Silicon Lab's embedded platforms.\n", "\n", "Options:\n", " --version Display the version of this mltk package and exit\n", " --gpu / --no-gpu Disable usage of the GPU. \n", " This does the same as defining the environment variable: CUDA_VISIBLE_DEVICES=-1\n", " Example:\n", " mltk --no-gpu train image_example1\n", " --help Show this message and exit.\n", "\n", "Commands:\n", " build MLTK build commands\n", " classify_audio Classify keywords/events detected in a microphone's...\n", " classify_image Classify images detected by a camera connected to...\n", " commander Silab's Commander Utility\n", " compile Compile a model for the specified accelerator\n", " custom Custom Model Operations\n", " evaluate Evaluate a trained ML model\n", " fingerprint_reader View/save fingerprints captured by the fingerprint...\n", " profile Profile a model\n", " quantize Quantize a model into a .tflite file\n", " summarize Generate a summary of a model\n", " train Train an ML model\n", " update_params Update the parameters of a previously trained model\n", " utest Run the all unit tests\n", " view View an interactive graph of the given model in a...\n", " view_audio View the spectrograms generated by the...\n" ] } ], "source": [ "!mltk --help" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Classification Machine Learning Models Overview\n", "\n", "Before continuing with this tutorial, it is recommended to review the [MLTK Overview](../../docs/overview.md), which provides an overview of the core concepts used by the this tutorial.\n", "\n", "Image classification is one of the most important applications of deep learning and Artificial Intelligence. Image classification refers to assigning labels to images based on certain characteristics or features present in them. The algorithm identifies these features and uses them to differentiate between different images and assign labels to them [[1]](https://www.simplilearn.com/tutorials/deep-learning-tutorial/guide-to-building-powerful-keras-image-classification-models).\n", "\n", "\n", "### Class IDs\n", "\n", "In this tutorial, we have a dataset with four different image types, a.k.a. __classes__: \n", "- __rock__ - Images of a person's hand making a \"rock\" gesture\n", "- __paper__ - Images of a person's hand making a \"paper\" gesture\n", "- __scissors__ - Images of a persons's hand making a \"scissors\" gesture\n", "- __unknown__ - Random images not containing any of the above\n", "\n", "We assign an ID, a.k.a. __label__, 0-3, to each of these classes. \n", "We then \"train\" a machine learning model so that when we input an image from one of the classes is given to the model, the model's output is the corresponding class ID. In this way, at runtime on the embedded device when the camera captures an image of a person's hand, the ML model predicts its corresponding class ID which the firmware application uses accordingly. i.e.\n", "\n", "![](../../docs/img/rock_paper_scissors_overview.png)\n", "\n", "\n", "### Convolution Neural Networks\n", "\n", "The type of machine learning model used in this tutorial is Convolution Neural Network (CNN).\n", "\n", "A Convolutional Neural Network (ConvNet/CNN) is a Deep Learning algorithm which can take in an input image, assign importance (learnable weights and biases) to various aspects/objects in the image and be able to differentiate one from the other [[2]](https://towardsdatascience.com/a-comprehensive-guide-to-convolutional-neural-networks-the-eli5-way-3bd2b1164a53). \n", "\n", "A typical CNN can be visualized as follows:\n", "\n", "![](https://miro.medium.com/max/1400/1*vkQ0hXDaQv57sALXAJquxA.jpeg) \n", "[Typical CNN Diagram](https://towardsdatascience.com/a-comprehensive-guide-to-convolutional-neural-networks-the-eli5-way-3bd2b1164a53)\n", "\n", "A typical CNN is comprised of multiple __layers__. A given layer is basically a mathematical operation that operates on multi-dimensional arrays (a.k.a tensors).\n", "The layers of a CNN can be split into two core phases: \n", "- __Feature Learning__ - This uses Convolutional layers to extract \"features\" from the input image\n", "- __Classification__ - This takes the flatten \"feature vector\" from the feature learning layers and uses \"fully connected\" layer(s) to make a prediction on which class the input image belongs" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Creating a Labeled Dataset\n", "\n", "The most important part of a machine learning model is the dataset that was used to train the model.\n", "For a machine learning model to work well in the field, it must be trained with a dataset this is __representative__ of what would be seen in the field.\n", "Put another way, a machine learning model can only make accurate predictions on samples that are similar to what it has previously seen (i.e. trained with).\n", "As such, approximately 80% of the effort of creating a robust machine learning model is generating the dataset.\n", "\n", "Typically, a good dataset should have the following characteristics: \n", "- __Numerous samples per class__ - 1k+ -> ok, 10k+ -> good, 100k+ -> great\n", "- __Mostly \"balanced\"__ - The sample count for each class should be mostly the same\n", "- __Representative__ - There should be samples for all the possible orientations, lighting, backgrounds, etc. that could be seen in the field (the model can only make accurate predictions on stuff it has seen during training)\n", "- __Non-redundant__ - Each of the samples should be relatively unique, duplicate samples usually doesn't make the model more robust\n", "- __Correctly labeled__ - The samples in the dataset should be correctly labeled. A few mislabled samples is typically ok, but too many can degrade the model's accuracy\n", "- __Uses same sensor as the one in the field__ - While not a hard requirement, it is usually best if the training dataset samples are generated using the same sensor as the one that will be used in the field. This way, the samples \"look\" the same during training as they do in the field\n", "\n", "\n", "### Rock, Paper, Scissors Dataset Overview\n", "\n", "This tutorial uses the [Rock, Paper, Scissors](https://siliconlabs.github.io/mltk/docs/python_api/datasets/index.html#rock-paper-scissors-v2) dataset.\n", "\n", "You can import this dataset into a Python script using:" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "Extracting: C:/Users/reed/.mltk/downloads/rock_paper_scissors_v2.7z\n", "to: C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2\n", "(This may take awhile, please be patient ...)\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "patool: Extracting C:/Users/reed/.mltk/downloads/rock_paper_scissors_v2.7z ...\n", "patool: running \"C:\\Program Files\\7-Zip\\7z.EXE\" x -y -oE:/reed/mltk/tmp_archives/rock_paper_scissors_v2 -- C:/Users/reed/.mltk/downloads/rock_paper_scissors_v2.7z\n", "patool: ... C:/Users/reed/.mltk/downloads/rock_paper_scissors_v2.7z extracted to `E:/reed/mltk/tmp_archives/rock_paper_scissors_v2'.\n", "Rock, Paper, Scissors dataset directory path: C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2\n" ] } ], "source": [ "# Import the Rock, Paper, Scissors v2 dataset\n", "from mltk.datasets.image import rock_paper_scissors_v2\n", "\n", "# Then download and extract the archive\n", "dataset_dir = rock_paper_scissors_v2.load_data()\n", "\n", "print(f'Rock, Paper, Scissors dataset directory path: {dataset_dir}')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "This dataset has the following subdirectories:" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "paper\n", "rock\n", "scissor\n", "_unknown_\n" ] } ], "source": [ "import os \n", "\n", "# The dataset has the following sub-directories:\n", "for sub_dir in os.listdir(dataset_dir):\n", " if os.path.isdir(f'{dataset_dir}/{sub_dir}'):\n", " print(sub_dir)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Each subdirectory represents a __class__.\n", "So the \"paper\" subdirectory contains images of someone's hand making the \"paper\" gesture, and similar for the other subdirectories.\n", "\n", "Each image file (a.k.a \"sample\") is a 96x96 grayscale JPEG image.\n", "\n", "The following shows some of the samples in the dataset:" ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Class: paper, path: C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/paper\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAr8AAACCCAYAAABPYqC0AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAACm0ElEQVR4nO39eXCc55Xfi3/fXoFegMYOcF9EkSIpEiIpUgslS2NLXsbbbM5olmvP5qQqqUlSk1RSN/EvqUnlzs2tWzd/JL+amkx+1zM143EysWPHnhlvY1v2SBqKEmlKoqhlJHHH3migu7E00N3v7w/wc/B0E6QoCQQa0HuqWAQaQPf7vud5znPO93zPOZ7v+wokkEACCSSQQAIJJJD3g4RW+wICCSSQQAIJJJBAAglkpSRwfgMJJJBAAgkkkEACed9I4PwGEkgggQQSSCCBBPK+kcD5DSSQQAIJJJBAAgnkfSOB8xtIIIEEEkgggQQSyPtGAuc3kEACCSSQQAIJJJD3jbwn59fzvI94nvea53lveJ73L5frogJZeQl0uX4k0OX6kUCX60MCPa4fCXS5PsR7t31+Pc8LS3pd0mOSrkh6TtITvu+fW77LC2QlJNDl+pFAl+tHAl2uDwn0uH4k0OX6kfeC/B6V9Ibv+2/5vj8n6b9J+tTyXFYgKyyBLtePBLpcPxLocn1IoMf1I4Eu14lE3sPfbpR02fn+iqRjN/sDz/P8ZDKpqamp9/CxgbwHGfN9v2uJ19+VLpubmzUzM7Oc1xfILYrv+94NfvSudJlIJDQ9PV3/uoIJkLdflkuXnuf5TU1Nmp2dDXS3OrKs9hVdBrIqEuhy/ciSurztBW+e533e87znPc97XpKmpqbkeQu2PhJ5L753IO9CLr6XP67X5d69e92fvcdLC2QlpV6Xd911l+kwGo1KUuA8rQGp1+OhQ4ck1erO87xgf66MLKt9rVQq1/1OcGaumAS6XD+ypC7fi/N7VdJm5/tN116rEd/3/4vv+0d83z8iSQ8//LAZ5nK5LEkKh8Pv4TICWQZ5V7o8deqUHaq+7wcHbGPIu9JlqVSyfTk/P2+GORQKGsKsorytLuv1mM/nl9RZEMisqryrPbl7926zqQQw5XI5sLOrK4Eu14m8l5PtOUm7PM/b7nleTNIvSvrG2/3RxMSEwuGwPM8zp7darb6HywhkGeRd6bK/v98O1VAoFBywjSHvSpexWMy+jkajFpgGOl1Vece6PHv27HWobyCrLu9qT549e1ae5ykSicj3fdNrsCdXVQJdrhN517i77/tlz/P+kaTvSApL+n9933/57f7u3LlzqlQq8jzP/gc1DBbC6si71eWZM2dMb24AE+hy9eTd6vInP/mJQqGQqtWq5ufnJS2k5XCCA1l5eTe6RIeBNI682z3peV6gywaTQJfrSIhCVuKfJP8DH/iAL8n+eZ7ne55X81rw77b9e345dXn48OEaPS71dfDv9vxb7n2JLj3P8yORiC/JD4VCq36f74d/y6nHY8eO2f4L9uGK/1tW+3rs2DF770gkYvsx0Ova0+XRo0drdBns0dXX5YoT+orFoiKRiNEenAUSyBqTgPO7fqRUKkla0GO5XA5oLGtUXO42EhS8rU159tlnJcn4odVq1dLmgawtOXnypH1dLpfl+36gy1WWFXd+T506pUqlonA4bP9LATdtrYq7efk62NBrT86ePWuFUqTOfd8PCt7WmJw5c0aSauhkUrAn16q452JQJLW2pT4IDXS5urIqvTZAl6TFFiCBcQ4kkNUVOGkuNy3gqa1NCQLR9SEBuLB+ZCm9BbpcPQlgnUACCSSQQAIJJJBA3jcSOL+BBBJIIIEEEkgggbxvJHB+AwkkkEACCSSQQAJ530jg/AYSSCCBBBJIIIEE8r6RwPkNJJBAAgkkkEACCeR9I4HzG0gggQQSSCCBBBLI+0YC5zeQQAIJJJBAAgkkkPeNrJrzy3CLQNa2eJ5XMxwhkLUt6DASWZUW4IEss9BEP2imH8hyi7u2sBvBOru5uGdkIz2rpc7u9T4ZcsW9FR4mwy3C4bCNOw5kbQlTpJCgYffaFQ4wjCBDaIJ9ubYkcEICWSlhAqTv+zYMJxhzv7SwL3lOnuc1FMCALl3d+b6/rs/0FXd+6x9wpVKxWdeBrC0JJg6tH+EAIygNxuKuTXEn8rnBaaDHQJZbPM+z9YYjF4lEgrV2C+L7vubn51f7MiRJTU1NZv/RHX7aes7mrvidVSoVe8Au9SGIFteeRKNRSbVO8HpPlaxX4fCq12Uga0tcPbpOSKDLQJZT6p2icrmsUCikcrkcrLUlxEV8G436MDs7K0k1mT8c4fU83n5V3fpqtRpwf9ewzM/PL7l5g8h/7QmHVyMY40DeveB8uHqspycFEsh7FRcldF8LkN+lxc2kNSqw4Dq7riO8XmVV7s5FJ1yuUCBrT+rR3kCPa1dcdEIKdLnWxNVbkIEJZCXEdZBAfgO5XnzfXzIr0yg2Fu42yHS1Wl3XqK90C86v53mbPc/7oed55zzPe9nzvH987fV/63neVc/zzlz797FbeC9JC9SHemQiMNQrItHl0mVzc7Oi0aihSu4mXu8RYyPIcu5LdCldj0wE+/L2y3LpsampSZFIxA6vYE+uuCybfW10wS584Qtf0L/+1/9av/ALv6BqtbqezvJl12WlUrmO8tAoz4oMfD1VqlGu73bIrZQbliX9ju/7pz3PS0s65Xne96797D/6vv9/3+qHNTU1qVKpaG5u7rqqwkBWTJZFl57n1RD2Xa7Qeo8YG0SWbV/W6zLYmysuy6JHF3kjKF0qOA3ktsqy6LLRxfd9/af/9J/00Y9+VJVKRcPDw7rnnnv03HPP6Wtf+9pqX95yybL5PeVy+To+dCPtyZ//+Z/Xrl27NDg4qJMnT+qFF15oqOu7HfK2zq/v+4OSBq99XfA87xVJG9/Nh3mep7m5OYXDYasqD2RFZd73/dPSe9clwuHaiA4vqZwbFf7Q3gXEYi1t9uXel3Xv3VBUFlc3kUjEHDx0t9Zlufdk3XuvmB6XOtij0agFVvV7rN5Br/9/qb251J6uf79bff02yLLZ10Qioenp6SV/1ii26v7771e1WlW5XNbu3bvV09OjD37wg/rEJz6h//E//oe+973v1dAg6vWJjXHpVq7+68Xd7/W/69oF9+v3IMt6VuL4NoLe6uWBBx7Qr//6r6unp0ee52l6elovv/yyXnrpJf3t3/6tnn322euesXsv9fZ5//79ampqUk9PjxKJhJqamiRJY2NjOn/+vM6ePXvd37nfu3q+nTb+HTWa8zxvm6R7JD0r6UFJ/8jzvP9N0vNaiJJyb/P3krQuDqy1Lsuly0aV+g3EgblUkUaj38vbyXLost5xaSQj7fu+otGooSeSFI/HVSqV1o0DLC3vnnT1t1K6rKdZhEIhc3xxguv34lIOLk7NUkFpva5vdM/1TrLrhK+ELJcuG9VheuSRR1SpVMxBv3z5skqlkjKZjA4ePKgtW7boV37lV3T27FmdOHFCP/rRj26qT9eZRdw142YU2fuuM1ZvF5ZT3qsuyYrW67FRdPtzP/dz2r17t8rlskqlknzf1759+7Rz5049/vjjGhsb09DQkF577TW98cYbmp6e1vT0tBKJhHp6ehSJRNTV1aW2tjYlk0lt27ZNiURCiUTCwE5JGhkZ0fDwsE6dOqXXXntNr776qsrlsumRZ1GtVhWLxSywul3i3erD9zwvJelHkv697/v/0/O8HkljknxJ/05Sn+/7v77E331e0uclKR6PH17uhRnIO5JTvu8fWS5dzs3NSVp9ZHApWSqK5PVYLGbG0035ryVHyvd9T7p9+7IR03NLHRaNcoC8F/F931sOPTY1NR3m8FoNebssyoc+9CGlUil9/etfl+d5CofD1x1u9U7rUlmJG0k9kshrK/g83jf29dOf/rT+8T/+x4rFYkomk5qdnVWpVFIikVChUFB7e7sSiYRKpZIGBgb00ksv6cKFC/b3s7Ozmp+fN+colUqpXC6rqalJpVJJ8Xhc+Xxely5d0sjIiP7u7/6uJnhZKqOwzLpeNl3GYjHTZd3vNIRuv/GNb2jLli2KxWKSFpzUVCqlYrGoWCymdDotSSqVSspms8rlcpqdnVUmk1E6nVY4HFZzc7OkhfqR1tZWlctlzc3NKRKJaGZmRtPT05qZmVGhUNDMzIwuXbqkJ598Un/1V38lSdfpb5l1esr3/SP1L96S8+t5XlTSX0j6ju/7/88SP98m6S98399/s/dpaWnxC4XCjT6jIRbCOpdTku7XMumyWCw2tM5cZzYSidyUj3yjNE6jyjWHaVn2ZTqd9ovFIn/TcPe+lB5DoVBNz/A1LjEtgx5bW1v9YrG4as7fUs4viP3Bgwf17//9v9fMzIzeeust/eQnP9Err7yiM2fOSFoasbuZM12/X6XrnUTXkV6mVPjbybLZV3dPNqIcPnxY//bf/lvNzMwomUwqkUhIWtBjoVBQZ2en0um00um0SqWSPM9TqVSqoTuC+s3PzysSiWh6elrNzc2qVCoqlUrmLL311lt66qmn9D//5/+sQYKlWieY15cJ5V82XaZSKX96erphA/evf/3r2rFjhyYmJizLJi0U6FUqFW3YsEHhcNj2U6VSUTgcVjKZVLFYlOctTKtLp9OqVquampqStFBENz4+rkqlokKhoGg0qlwup1KppHw+r6mpKb355pt64YUX9OSTT5rObsNeXdL5fVvag7dgWf5/kl5xF4DneX3+Au9Qkn5G0tlbuYpGUfj7WJZFl2sBhatv3eVSHNjk9caUDd7ostz7su69G+oZcC04SJ7nrSfHV1omPfpLcO9X8hktxc8lnfm5z31Ou3fvVqFQ0NatW7Vjxw5duHBBd999t770pS9dR2Fxv8Z5vZHOb2SHfN83dHkFW3Atm31tZDl16pTOnj2r5uZmQ/twbEOhkDZt2qRYLKaJiQmVy2WFw2F1dHTYfTU3NyubzSoej6tcLqulpUW+72tyclK5XK5mbPKOHTvU3t6u/v5+feELX6jRJY4zwIabyVsGuS32tdHkjTfe0OzsrBKJhKLRqNzsEdnRarVqFJdMJqNoNKpYLKaWlhbrLDM3N6dKpaJMJqNqtapcLqdCoaC2tjaVy2Xlcjk1NTUpHA4rHo8rFovp0KFD6uvrU2dnp86dO6cXXnhhxfjRt8L5fVDSr0p6yfO8M9de+98lPeF5Xr8W4P8Lkv7+bbi+QJZXUroNunQPm0YT+FZuMQTGsR4xuFmRSQPKbdmXpJoaifbANZVKJUWjUVUqFWuov076it6WPSmtrO7q0Vic16amJj388MOKxWLKZDLmDOE4UWV++fJlXbhwQefOnav529nZ2RvqGW4xQ1rqA1z+boU4v7fVvtZ/vZrieZ5ef/117dq1y1Ba0N/m5ma5Gd5UKmX8z3Q6rbm5OWWzWeMIz83N1Tg8pN/RaalUUrlc1oYNG/R//B//h/72b//WqDP1vYWXorq9S7ktunSlEfQoSS+//LLm5uYUi8XU3d2t+fl5JZNJzc/PKxaLKZvNKpFIqLm5WdVqVYVCQZ7nKZFImI5CoZAFqNVqVfPz84rH49q9e7empqasycHc3JxmZmYUDocVjUY1Ozurzs5OPfDAAxbkwA+/3XIr3R6ekrRUGPpXy3khjbIQ1rkU4YrWybLpspGMs0ugl2QFOKFQSL/927+tnTt3am5uTn/1V3+l73//++b4rgXu73LuS4xXIw+cgae9e/duqxZeJ46vlmtPErQQiLoBzErsy6XePxKJ6MCBA4rH48pms2ptbdXExISq1ara2trU3Nxs15bNZq2w5tKlSxobG1MoFFI0GrU2mZVKRZFIRNFoVOFwWE1NTfJ9X4VCQZOTkzpz5kzNPuZZrFCx27LZ10ZHfj3P0+nTpyVJvb29SqVSNf3CJycnNT09rXQ6rYmJCbW0tGh6elrd3d2anJxUa2urwuGw5ufnDTWMx+MKh8MWOLnOFRzgnp4ePfzww0omk/rTP/1TC4LJCiyj7b7tZ2WjyBe/+EV94Qtf0Pj4uObm5pTJZDQ9Pa2WlhaFw2Hj7La2tmp+fl7T09NqamrS/Py8CoWCWlpaNDc3p5aWFknSwMCAMpmM4vG47fVCoWCBCrzfUChko5WbmprU29urxx57TB0dHXr11Vf16quv3tbWqe+o28PtkEZxlgJ5Z3Ij49xousQYghJGIhF96lOf0qOPPqo77rhDkrRv3z7ddddd+s//+T+vCcd3JaUR9qfnebr33nv1m7/5m2pra9PTTz+tJ598Us8999yqX1sjiev8SqvT8QGE1eXZptNpjY2NaX5+3gYc8XqhUNCmTZsUCoXU09OjjRs36q677tKVK1cUi8U0NzenZDKpubk5NTc3W/CKk0TafHR0VPl8Xvfff7+Gh4d1+vRpnTt3zp7LWl4njXjt1WpVL7/8siKRiFXze56nnp4exeNx5XI5TU9Pq6urS5JULBYVDoftXpqbmxUOh835IagBPa5Wq5qcnJS0kKELh8OKxWIKhUJqa2vTgw8+qPb2dj377LN69tlnb9oCb7Wlfl824np87bXX1NfXZwh+OBw2J3dsbEyzs7OanJy0QtWpqSlls1nTSyQS0eTkpOLxuJLJpHzf18jIiPU4DofDymazmp2dVaVSsSLJeDyu2dlZhcNhJRIJJZNJHT9+3DpGnDx58rbd84o6vzw4orRGXAS3ItyDm15zX3OrmDkA1uq9vp3UO4sr3VLoRsL64tkzXadcLuvw4cPauHGjRao7d+7UZz/7Wd1xxx36J//knyz5fuhxqSKct9OtW7nKtTWaUETm6rJRAgHf9/XJT35SH/3oR1UsFrVlyxbt27dP3/72t/XlL3/5ptd9s6KnpQqz6nW6lvbtjTodrJQe3dZm7D1JmpmZ0dDQkCRpYmJC3d3dVg1OCnz79u2qVqtqbm5WOp3Wxo0bzRHC+SFDQWFcc3OzZmZmlM1m1dbWpvHxcaXTad15551KJBJqaWnRiRMnap6Buwcb2Ta7vGnWYf3UvtWWSqWin/zkJ+ro6LCiJgqifN/XzMyMKpWKOjs7NTQ0pPb2dhUKBevokEqlFI1GFY/HFY/H5fu+KKKemZmRtFhAl0qlJEmFQsEQ4rvuuku9vb1Kp9P667/+a0m1e/pG636l9Y3u+FzXP2iEted5nv78z/9cP/dzP6fe3l6NjY2pq6vLnnM0GtX4+LiBSKVSSU1NTUZjgeICuORSlYrFotFWQO9TqZSdyfUFiu7U2E996lPauXOnXnvtNcsyEEjcqj272fNdcee3vuk1JPm1MPgCw8l18lB/4zd+QyMjI/rmN78paWERgFq4KIxreBuJU/lupP6+WJCN4PgirpFxUQHW2vj4uKV24vG4Dh48qC9+8Yv6wQ9+oC996UtmsFyecD014J04SY2q53raA/fSKMM/HnjgAR0/frwGWTh48KA2bNigw4cP6+tf/7pOnDhxXbFLvZF0g5d6g1u/bpdCTxtd0Bfr3qX9rIQD7HbkINisVqt65pln9KlPfUobNmxQLBbTxYsXFQqFrIVSJpNRPp9XLBZTLBbTzMyM2traDCmMx+OSFp2ZZDKparWqfD5vB7EkQ4ir1aruvfdedXR0aOPGjfrqV7+65DNoVMdXqp2YebPXVlOwqd/97nf1kY98RNu3b9fAwIDx8yORiHzfV1NTk5qamjQ9PW2FVclkUlNTU9bdYW5uzpxnKC2zs7PGD3VRxKmpKTU1NSkWiykcDuvhhx9WIpHQN77xjZoRy/WUCGl1AnrOB/Zlo/k5PPOvfvWr+vjHP6729nbF43G1trZqbGzMAppQKKTW1lYVi0VNTk4qk8kYn5dMDnSIRCKhTCajYrGo6elpzc3NGZJfLBaN+w1vOx6Pa35+XqVSyVrfpVIpHTlyxIopX3zxRUmL9uxWbPTNfraizq8bebuOL1FAo4v7IHGg/uAP/kD33XefJiYmdOjQIX3ta1/Tiy++KPr6kc5xD6P692IBrIVnUC+uQ9koRhlxHV9p0ambnp5WNptVV1eXcrmc0um04vG4Ojs7jfR/9OhR/ehHP9JXvvIVS7VK11e01//vGt76KLVRkd9yuVyjO3e9NsK17tixQ5VKRblczlCDqakpJZNJ3X///erp6dGHPvQhPfPMM/rud797nXF0Dzz+dxHK+sJH/tYNetaCcMgvNQFvpfamG0jguJZKJZ05c0aVSkXxeFzt7e2KRCIaHx83LjfcwubmZrOFOE0MU+jo6DBEkECbwBUnp1wua2Zmxgp3uru7Va1W9b/+1/+6bo3DJ7xZJ4nVEleH7MN60GQ1pR68+fa3v60Pf/jD2rVrl4aGhtTW1mZc3omJCUkLzzwSiahYLFr1fygUUiKR0IYNG5TJZIznOzIyYsGPa5/Yl1NTU6pUKurq6tLU1JSOHz8u3/cNgLpRy7PVOKNA7OtBs/qvV0t4Tp7n6S/+4i907Ngxbdu2zZxU6CbxeFyTk5O2X2dnZ01/5XJZiURCvb29ZjcZUuFm+rEHc3NzNQXM09PTlpklUOC9d+7caQ72yy+/LKlWj+82gF1R59c1MPUR7Vpx/HjQlUpF9913n/bu3auNGzequblZH/nIR7R582b9t//23ywNAwLs/j3iOk5r5f6RerSzkdI4S12H+9yZNJPL5YznBME/EokolUrp8ccf13333aef+qmf0p/+6Z/q5MmT1xVYLZVWdlPspHWkpUdzNorcCL1vlOusVCqamprS7Oyspd0wpPPz89qxY4f27NmjRx99VD//8z+vL3/5y/rhD39oOq9HJN1AzUUE623TWnJ8paULF1c64OJzeNZQFL785S8rmUyqt7fXiqJIq46OjtqQg1KppPb2dnN+QPdCoZChSuiMNks006dlFlmdrq4uzc7O6v7771d3d7deeOEFnTx58rrJUY1YONmogTJS74iHQiF95zvf0czMjDZu3KjZ2Vm1t7erWq2afY3H4+ru7lY4HNbly5dVqVS0ZcsWbd++3dqgTU9Pa2JiwtptAVbQGcDtLAAFBuT/Ax/4gDo7O/Xcc8/p7NmzNft3FXo+m9woYGkUWtn8/LzRviTpueeeU7lc1rZt2+R5C0Oh6PwwPj6uyclJK3ILh8NqbW3Vxo0b1dvba1x8+jaD3vO7FKNOT09bVw944zjE8XjcHOKmpiZFo1G1t7fr6NGjkqRz586ZM/xenuGKOr/hcLiG3uA6ko1y0L6duEjn/fffr87OThWLRbW2tqq5uVnlcll//+//fe3atUu///u/f53jW3+Pa+W+68VFBW8087sRpN6x8TxP4+PjGh0dVSgU0uDgoHp6egx5ymQyllZNpVJ67LHHdP/99+uNN97QiRMndO7cOX3rW9+y94IeUM/NI4JFGpk/6jqJjViYMTo6qtOnT2v37t3K5XLq6uqyQxGbkkwmFY/Hdfz4cR04cEDnz5/X8PCwxsbGNDY2pgsXLujChQt6/fXXa9aqdP19S9dzgNeCuJXR9XtyJcSltbljS/n/v/7X/6pPfOIT2rlzp5qbm7V161bNzc1ZcdTc3JxCoZClR5uamqwIKpPJGOrr3g8dIKCaFQoFFYtFRaNRTU5Oyvd9tbe3KxqNWkcCWinVT3hsNGmU/XcjqQ90JOmpp55StVrV8ePHrdgN9BCEsaenR8ViUffee6+OHFmYPTA8PGzrJJ/PKxQKKZvNqlKpaHZ2VtVq1ZxfnCLWG9eRTCa1b98+9fb2KhaL6fTp0/ae9XSXlZQbUR8bRbf1eyAUCunUqVM6deqUjh07ppaWFnV2dqpUKqlYLBpnOxwOa/fu3dqwYYO2bt2qWCymqakpe97FYtFoSEzzGx8ftzMSG+t5nmZnZ5VKpRSJRDQyMmLXQQu2LVu2KBqNGs/4xIkT7zl4WFHnF0Po8sGQRlkIbycuRaG9vd04L/Pz80qlUspkMorFYnr88ce1efNmPffcc/rLv/zLGv5vI26AdyNLGedGuZ+lMgw899dee03t7e3atGmTPM+zRuwgSL7vG38wHo8rnU5r+/bt2rlzpzzP07/5N/9Gk5OTOnv2rN544w1dvnxZExMTeuqpp65zcusLIN1raxRxgwO3ILVRrvP73/++7rnnHr3xxhs2MpV0dqFQsHQ4xjedTqu/v1/Swr01NTVpampKk5OTmpqa0tTUlDzPUz6fNyTi8uXLunTpkl599VWdOnXK0JBGdYyWEldfq4Eo1WdaXEeY1958801FIhErVGpvb68piGltbdXg4KC1xoITWi6XlUwmVS6XrT8waBLdBmh95rZQikQWjjhQR7jAX/nKV2r2ZKP1jHYzgY14XrgUofqsSSgU0tNPP21Zsb1792rTpk3q6elRX1+fNm3apIcffthoEDhAIH8AEThFvCdpcToE0AGkUqloZmZGpVLJxvE+/vjjikajNZ0gpNXpSV+/FxutnsClh0mLXPhwOGwddXzf1549e9TZ2alMJqOdO3dqz549di5yboRCIc3MzNi4a9ZufZYYOgTBDWgv15HP52vGK09MTKi5uVmHDh1Sc3Ozent7denSJU1PT+u1116ruY9blRV1ft3IfS0VurkC0kfrjsHBQY2Njam1tVWFQsHI3pFIREeOHNG+fft05MgRPf/88/rmN7/ZcNyydytusYiLcjbKvS0VFXJtL774ojZt2iRJ2rp1q0qlkjo7O5XP5+0wpSqVTcfoxkgkoo6ODtuADz/8sHGjBgYGVCqVNDo6qjfffFPPPvusnnnmmes6fzRKugtx+XH1umwEp8DzPF25csUcXzhlBCmswVgsZsY0lUrZ9yDD6XS6xjHCHpVKJfX39+vixYu655579PDDD+vHP/6xTp061fDomysUF+EwuK+v9JpzC0Fdfb3++uuKxWKWqga9r1arisVixh+EJ8ohWSqVNDk5qWg0alxA3oMWasVi0QrnxsfHrYPA9PS0enp6VC6XtXnzZsXjcX3605/W17/+9euG4DSK+L5vfGieYyPZjaW45G4mTFqwwXfccYd27NihDRs2aMeOHTp06JB2795tQy9KpZJxQEEHy+WycUCZGgblpVQqKZ1Om22SZD1jcbzoHnHgwAGdOnWqIXQbi8WuQ6ClxqA+cA313VDqg8NNmzZp8+bN2rVrl/bu3autW7eqo6NDkkyXs7OztqfRB4WMbv9tabHWBDsNysswDd9f6M89OTlpGaBwOKxDhw6pVCrpwoULyufz2rx5s30256sbFD/zzDNL3veq9PldiuO71g4ZSTalZmRkROPj49q4caPK5bIpgdnmDz/8sPr7+3XkyBG9+uqreu655/TWW28ZJWIt3TviIiuknpaK8lZLljLKLq/1ypUr6urq0sWLF9XR0aFMJqO2tjZJC+2YYrGYfN83FAKeIgiF53mKRqMqFotqbm5WPB63NG53d7d6e3t1+PBhPfTQQzp58qR+8IMfmDFZbWNXL+hyKV5yIwSm1WpVb731lpqamtTV1WVV36StQRwKhYJaW1vNQaJ/ZDgcNlQIYwuPkMr0QqGgRCKhvr4+tbe3K5VKKZ1O68knn1zt279lwZlH3EKuldiP9Rx4vqeYhvV05swZlctl7du3T/Pz82ptbVW5XNb27dvt8CQglaTx8XFJ0pYtW9TW1qZEImGcbyrPfd83fvHIyIhCoZA5jwzYiEQitkaOHz+utrY2vfTSS3r++edv+7N5p1K/J10dNoLDtJRN5fsHH3xQ9957r7Zs2aK77rpL6XRavb29ymQyxtV1i6JmZmYsNU6gwzkKHYbCKGnBPsMR9TxPnZ2dtkbgkdIx5LHHHtP3vve9GkrQSp9N6JJn5e6R1dajtHRxoKvfz3zmM9q/f7/uuusudXV1adu2bTbemMw3fYDJ0szPzxsFgmCGYjk6PczPz9sagKuPjY/H45Z5zefzNlwjnU4rmUwaza1cLqu7u1v5fF4TExOamZlRJBJRc3OzcYkbyvl1D9SlKiDXglQqFf3Jn/yJjhw5ong8rqtXr9pBA+mbeeUgHR/96Ed177336ld/9Vdthvnly5c1MDCgl19+Wc8+++x1qcMb8S9BS/i81XJSlkrnN5ou3evi6zNnziiTyaivr0/SgtHs7Oy0zUZ6NZlMKpPJ2IQa0rHxeFwzMzPq6OgwXlOlUtH8/LxmZ2dtM+/YsUMdHR2qVqv64Q9/KOnGvSiX6iG8krLU4dAIuvQ8T88++6zK5bI2btyonp4e42m3tbUpm82aoc1ms9qwYYOlSKGxgDChN2mhTywII10D0G1fX58ef/xxxWIxffe7310yMGgEVNwVOiPgbC431cZFhdwK8XobVV+Zv9QzOnv2rLUw6+7uto4rOD/RaNQ4+LFYTB0dHerr61NnZ6d9RrFYNP0xSYrRrFNTUwYuuCnZzs5OTU9Pq1KpaNeuXXaIPv3003YPbocWl7eMs+Kuhdu9BpZyjm7VYXJ1s9R11mfrXIfH/Z2lagEkqb+/X9u3b1dvb6/uvPNObdu2TT09PcpkMuro6LAABCCIgqdisahEIqF8Pm9dG7LZrKQFBHFqasoKplpaWjQ2NqbJyUlzlCqVihKJhBKJhAU4kqytFgjh3Nyc9u7dq1AopL/6q7+ye61vGXcje7zccrtoSfXX/3bB7lLF4PX+wwMPPKADBw7owIEDOnTokNnbarWq7u5u+b6vXC6nfD6vpqYmTU5Omu5GR0ctE4cz6vu+ZWIKhYL1+GYyXCQSsVobnOD29nYVi0WjR1D8Ckc5kUhofn5eO3futEJJCvNmZ2fV2tqqZDJ5w+ew6hPe1pqwcIDU33zzTXOa4KjkcjmrdAUpZJO2trZqdnZWmUxG3d3d2r17t6anp/Xxj39ck5OTam5u1uTkpMbGxjQyMqKLFy/qm9/8Zg1vrr4/ciNEj2tNIpGInnzySX30ox9VJpMx48sM88nJSSuyyOVyamlpUWtrq0ZHR41z1tzcrOnpaVWrC835ESbgxGIxxeNxeZ6nn/7pn1Ymk9HXvva1GqegEdNgjSbsudOnT6u7u9v2FI6ROwY3kUhoZGREMzMz2rBhg+bm5jQ9Pa1kMmmOFbQkDHQ+n1cikTAuMLSmdDqtBx54QNFoVH/5l39Z4xzUpwUbQZZCmKRbOxBvRdy/p/0fzqCkJR3CpQo/EWgloIOTk5Pm0LS3t6u1tVWpVEpbtmzRtm3b1NfXp2g0qomJCUuHE8yAEtLqjM8C1SIFOzc3Z4VTtFyTFsCMEydOGD+fe6jnLbsIGQEWXzfSvq0PpOt7Gi/lGLuv+b5vXN2NGzdqw4YN2rlzpzZs2GCtyZLJpDkj0IxA80hxSwvrAjvJOsTxnZycNGBhdnbWqA2kynO5nLU15H0YbLJlyxYbetHV1aVSqWSTxzo6OiwQOnDggCqVisbGxvT8889f12mK+3WDnrUqN2sB5vKd63V9//33m77vvvtubdu2zWygO3mP81Ba3BP5fN44utgFulzNz8/b4JJisWjX4p6BFNFJMt4/wXt9ezyADLI/BLqVSsUoGEyPY4DKjSRwft+hsHDgEj799NM6cuSIQqGQtm3bps7OTvs5CEM0GrVFRFGGmzJobm6W7/vavHmzcV84VHK5nD760Y/q3LlzOnHihE6cOLFkf+S1xp1ebWHjfutb39JnPvMZtbS0mE5CoZA6OjqMUwq/G31lMhlJsgMY+kM0GjUOKgHS5OSkTSd68MEHlUwm9Wd/9mfXcX/Xg+G9XeJWBn/rW9/Spz/9aSuKqFQqNvYWqgJtlEjBkT5PJpNqbm6uQRAoaISzlkwmlc/njebS2tqq++67T5lMRl/60pfseuod4UaS+gLL5V5XS2Ur3J/V2yH3+bgIKtd67tw5e59HH31Uhw4dUkdHh/ELt23bZkMxpqena/puu/xQ9/1dx5TrArAAGYIWQxoXe877ugg/4qaGXRS40fZuPfrONbpdTNzeyKFQSMePH9ehQ4e0bds27dixQzt37lQkEjG6ELUP0sJeIptSrVbta0nX9W31fd8K1JgENz8/X0PTwVHCuaEfLNPDCoWC8Z+npqaUSqWsXRb30tfXp0qlYigk7dakhTZooIRzc3O6fPmyzp49awGjtDanOkq1nP6lAt36Ir9oNKr7779fd955p+644w7t2bPHsi/pdFqtra3GvYYWls/nbR3hzJZKJY2Pj1ug464FvofzD/1pqQxEIpEwKow7Qa69vd3WD34TCDHZJ7qI8LNyuaympibl83lJsj2/lATO7zuQ+sr9SqWiF154wZze8+fPWxcI+ISgwcyuZqoNhH/e023WDl+mVCqppaVFpVJJhw4dsm4S3/72t6/rHxw4vu9M6vm/W7du1ezsrHp6eixdQxo1nU5rdnZW+XzeHF+QiEQioXQ6bekdDgsasadSKZVKJWWzWUWjUe3du1e/8Au/oK9+9as1VbUu2hRIrdS34rl06ZLa2tqUTCY1MTFhfYAJFguFQk3RE4YcI0xnCILRqakpOzg46OEcet5Cn8tdu3bp53/+5/XVr371hlSkRpH6FmfLdaBzsOL8u6hvPeWB30fqHVIXWX3ggQe0efNm7dixQz09Peru7tamTZu0detWxeNxS6GDCsIBnZmZ0dTUlIrFYk2rQYqfcGwAE1pbW1UqlYyaxJqClnbnnXfqb//2b6+jNrjornvtvFatVs1ZaxRxrx86Hs9QWtDXwYMHdc899+jgwYPauHGjtmzZoq6uLqXTaRs9XCwWVSgUzFEGnGFsLRxbUtmhUMhQOzIo7DN0ND09bfxv1mShUKjp2oFup6am7NmS0qb3K6lwRupyvobDYU1MTCiTySgUWhjOEIvFLHUfCoW0ZcsWbd68WVNTU+a8tba2mjP99a9/fYU19u6lPhhdCsWPxWL62Mc+pkOHDumOO+5QT0+P7rjjDnmeZ1nLSqWiZDKp0dFR0w/Fa1NTUwYI4K+UywvT9lhTTONz9wNBD3vHPeuq1YV2olCP6NSCPSZDl0wmrfOHO1aZ9QedgoAM+hfB0o0kcH7fgZA+4kDl+69+9av65Cc/qa1bt2p8fNwWzcjIiG3Qjo4Obdq0ySa+FQoFG4LAwkilUjVtQnBuSfd0dXXpkUceUSqV0le+8hVzgAPU990JBuOZZ55RtVrVli1bNDExoUgkYpwhDLAkC044VN2CDzb9+Pi4IpGI9SSkyTfRai6X0759+5TP5/Wtb33L1pCLmDVaCrVRBATu9OnTampqst6S0iIiNzw8bAghxYjRaNTS1b6/0BEhk8nY4UvngEgkYu3u6C4QiUQUj8eVyWR06NAhxeNxfelLX7oOXW00qQ/Ul0Pqq8FZp4lEQtPT09elz+sd4QMHDqilpUVbt25Va2ur0um0+vr6tHnzZrW2tiqTyRg3tLm5WZ2dnYbQDw0N1aBQ8/PzhgpywLooHu2vwuGwOVb0EXUdJdYAvUR/9Vd/VX/8x398HV+5nsrhdkdhvTSSuA68qyv090d/9Ec6cuSIjXfH9uAwhsNh41SPjY0ZfWtsbEylUknNzc01zgjnGc4utpH9KckKT+HmwuvFUaaHL8GntFhgzH1ks9nrMjbxeNwQ7UwmY+cz6XH3d/v6+hQKhazAmdoOzmJoN2vJ+a2nG7lUjuPHj+vjH/+4Hn30UfX19dUEIzilFAEXCgVVq1XLqEEDg8rS1NQk3/fV2tpqaH6pVLLiYdYRyD4/K5fLNW0IWWu0OAMcLBaLNjTDzexhy7q7u62zRzabVSaTUVNTk/H9ydLy2bOzszf1iwLn9x0KBrheLl26pGQyaQ2gN23aZMaXSOXq1auSpLa2NqMspFIptbS0GG8GjgrpXNqoSQtVz1NTUzpw4IDS6bS++MUvSmq87gFrQeoLO5599lkVi0Ur2MhkMlY8QcRLP9nm5mbjrOFExePxmgrTbDZryEdzc7PGx8ctnVcsFrVnzx5Vq1V95zvfuW4gQaDPWuHwdg37M888o3w+r02bNml2dlbpdFpjY2N2iILuMi63UChoZmbGkPqWlhZrnTY7O6tisahisaiZmRlLnYH4g0TMzs6qq6tLn/zkJ/WNb3xDUuOhv67Dczt6S9fz1SORiKWv0c/999+vlpYW67Xc2dmp5uZmtbe328CDnp4es4/d3d2qVCrKZDKGQvm+bwUxtK/DyZ2cnLRD1G2RxT3jPLnToijKYVSrJEMv2btTU1Pq7OzU5z//eb311lv667/+a3ue9dxlNxOxVN/61RbXpuCoHzx4UL/6q7+qD33oQ9q8ebPK5bIGBgaUTCbV1tZmzi5gC9QE0FocIgL7cnlhIAW8XDInnGk40fDu4de7GctkMmkUCdoYAvyA2vKsKWTD3hJ40A4L2hnZNrqGJBIJQ69xtDo7O5VKpSxrQNbO930bsrKWxM0E7927V7t379YnPvEJPfjgg9qyZYtisZgFHQADBIY8F7jXAD7YS/YMAQafQ6DS3t5ueiFQIfPN70gyhJb3kmQUi2g0qtbWVs3Pz9uQk6amJtu/0WhUvu8rlUrVcMBnZmbU0tKiYrFozjpZdDLoN5LA+X0H4ha7EeXy/5kzZ9TS0qIdO3ZoenraDluKN6BA0M6F9AuLhHZZExMTamlpUSaT0djYmLZu3aqpqSkVCgVDsEC9nnjiCf35n/95gPq+S6l3CM6ePauzZ8+qv79fHR0dikQidmDTfxQ0A7SINKEkQ5dIFcE942AmcMJo33nnnYrFYjaPPqA9LC1uit1N7509e1Yvv/yyjh07pmQyqWw2q6amJg0PD1vhIgY4kUho+/bt2rFjhyGAoPigvOgIB4C2TAw8Ab04ePCgwuGwFS82ssC5W65hHS51IR6P66GHHlI6ndaePXvU09NjlK7e3l55nqf29nZVKhW1tLSoXC5rw4YN5sRu2rRJ0WjUqvulBYcUXbvFUzhEbiYFJFGS8fWr1aqhfm63BxAlAlICBBBL6GpkAii2I7X75JNPXkcf6e/vVzgc1qlTpxpuHdQ7vl/4whf0qU99ygrWBgYGNDU1pbGxMUUiEQ0NDRmFoFQqWRBIGnl2dtbGDre0tCiZTNp+ZPQ4VBO64rgFTxQ1UaWfyWRqsp7oju4QqVTKioqhXxDEEJzOzMzUDLsBkWS90T9WWszaUVhFRgcHi7XEe641mZub02OPPab77rtP+/fv1+7du9Xb26tEImE9sEdGRlQqlTQ2NmY0gnw+b+cXwYxb8zA1NWVcanwaAg1oYRTGAdxR+CgtZBIIYNnTTU1NCofDxtllMq6bmcD2EmzBOXf5xy0tLZJkWQQkl8vZGQvneym5JefX87wLkgqSKpLKvu8f8TyvXdJ/l7RN0gVJn/F9P3drqlrb4nKp3O//5m/+xow/kQvplNHRUc3MzNjgBAxssVjUhQsXLFLu7Ow0flVnZ6fGxsbsczHiGI39+/erqalJX/ziF28ZgQp0eb3UP7szZ87I8zzdf//9CoVCNWm3trY2mw63efNmtbe3W5oFRGpmZsYKSUj/ulw7l9S/adMmHT9+XCdOnKhBfwNdLgrPor44kMP3xIkTkqRDhw6pra3NDjg4X3v27NH+/fu1ZcsWM/TudDc4iAS2BCFQlCQZZx9HbP/+/cpkMpZ9ea+yXLp0OzywZpdzSh0j3Xfv3q1du3Zp8+bNVijDIZVOp1WpVNTb22soDcggleNQgxgsQ3W/JEOi6AQAlYwgpFAoGMWBQlOQYQpu0B8V5nB+cYIpOnb7stOSi0O8s7PT2nBt27atht+YSqXMzh8+fFiFQkFf/vKXG2pPkrL+vd/7PX3oQx9SOBzWyMiIpqamNDIyovn5eXNChoaGrDCX5yrJUFlQ22QyWYPOzc3N2WARAqNqtapsNmv0kHw+r8nJSXmep6mpKbW3t2vLli2WeYnFYmbzJicnrS+zS9sBuYSyQPEq6xx7S0bV7VDg8rJJj6N7bDG2Guf82vNrGF2+nfzH//gfdfToUbW1tVlRL8Hf5OSknUl0TaB/cjabtcwYzw8EtVwuGzKPY1qpVDQ4OFjTFtItFASVJ3NDhoZ/4XDYwCMcXYIcfBv0TiYJu+Bm0uEU00ELukYymbTAhozRjcR7B4fsEd/3x5zX/i9J477v/5+e5/1LSW2+7/+Lt3mfVQ2PXb4a3y91/25z+JtVgNa/H39LXzy3RyUjOeEhwndj43vewtCEnp4eq2QEvcDgX7161dK0VFJygP/e7/2epKW7PjjXeUpSpxpEl0tVqN4sfbhUZfmNvl+qrc+NqtRvVr0uSY888oi2bt1qvSwPHTqkTZs2Gd9oYmLCjMHg4KAikYgmJiZUKBSMt0QqjXWFwwzX6W/+5m/04osvLtmKZqnr8n3fa7R9udSzW+rZu6/faF+5jpC7F6VFGgT8LncdVatVPfDAA4rH43r00Ud17Ngxa1eGsYQz6jpL2WzW2viQAnTbaWGIQcQoqBgeHjYH+N1Wiy+XLt+rHuvpNwQZDz/8sA4ePKitW7cqlUppw4YN6ujoUGdnpzZs2GCHDdQeJurh8OKcjo2N1RSvQCFilG0kElFra6sh9RMTE9YY36U5FAoFs5dk0wgwC4WCobk44aVSyQbZgPZFIhHLzNCmyfd9Q/1IgUciEaXTaQ0PD1sLPTer4Pu+rY9/9s/+2Yra15vtN+SjH/2ofvd3f9fSxefPn9fQ0JAkWZupTCajfD6vLVu21IwJZv3jCFFclEwmDSmWZIhpR0eHPM+zYUHszVKpZHUwZAIYJFMsFk33tKvDGZMWA9OWlhblcjkbioLDQ0aGAQgglm4xHmsSXioFyqD84XBYV69etXR9tVrVZz/72YY5K5fyM5BHHnlEf/iHf6j29nabNUDXGhxcaDsMAolGozbqHc7z7OysYrGYJiYmrPUjASJ0P2lhzVHL1NPTY72YcUbds5xrwRmGMtHR0WEZFRfQo+aCf2QioBJi90dGRoyqxLVLsswN9NJ0Oq2pqSn99m//9inf94/UP7v3Qnv4lKRHrn39x5KelHTTRbDaUr94XOdIqh27Jy22yah3hJd6P35WLpd17tw5nTx5Uo8++qjS6bQZa1oxEdkMDQ2ZgY5Go7rjjjssNcRB5LbqSCQSKhaLGhoaUk9Pj0ZHR+0Q+NznPqc/+ZM/saiXxXeLxXArrkuX2+o6wPXGHJ3Up715jd9x/9Z1QqTrWz/xt6B69W3H+vv7bSPu2LFDu3bt0p49e3TPPfeop6dHqVTK0nIgDkTHGGLoLZKs6JHrIgpOJBKamZlRa2urHnroISUSCT3zzDP2e26D/foK3pvIqu1L95pdvbn/u8bRvR/39917rOetuk4ZOvypn/opdXV1qbe3V319fUomk9qxY4f27t1bk0kBeSQtLsla4vC5FL9Isv2DnjlQcc4SiYR27dqlX/mVX9Gf/umfXscjXwYe6Irqsr646+DBg3rkkUd0+PBhtbe329CXpqYmS5+TVp+cnDReKI6OG+Chq0KhoJGREUtj4kyCUs3NzRldjD2N00tqmlQrHXHcbAtOL/cDSgSaR+BJFTnoIetidnZWyWTS0sXQ0zjgyQzw/jhYbl/bJWTZ9VgfsLOn6gPmgwcPqlAo6NKlS9Yjt1gs6uLFi4rFYtbaz/M8DQ8P23p3C5wI0Dds2GDFbuiKQJGBA7QdbGpqsnS2JGv5SMobQAD9YQuamppUKBTU0tJiBYXSwl6kQA0+J8g1aC72F7qD23GCbjw4XbxOESx9Y31/cVLgSuny7cS1Ky4Y8C/+xb/Qb/zGb9igH4ZLzM3NaWRkxJDY6elpy8AUCgUbDAPI1t3drampKZuS5j539qG0eGbiw7S0tNQMuyHT4LZfxX+Zm5uz/cbzlxZRX77mHEgkEnaegNK73ZH4PdaAtDjhkrXwdhTCW3V+fUnfvRbB/IHv+/9FUo/v+4PXfj4kqWepP/Q87/OSPn+Ln3NbxT1sXd4XGxwnxaUzuAjU2z1M3oOUERO9HnvsMaVSKY2OjpoRh9MSjUa1efNm7dy503rWTU1NmZFmXj2IhiTjYOH4VqsLU1d+6Zd+Sf/9v//3mrHJrgN47b4aQpdLPdd6p5XDD2e+XC7ryJEjam9vV0dHh5LJpH0N6hAKhQxZBw2Cp4uhJkAYHx833hG6TyaTSqfTZly7uroMkUqn08pkMha4EJG6DhUoB4gvhVKgFBzQFApRLZ1IJHTw4EHFYjH96Ec/shSi60ySenIOuIbQpbR4GNdTgepbQLmv79y5U6FQSJs3b7ZKfwpm3MMI/ujs7KwikYh6e3uVTqcVCoWUSCTU1dWlVCpl1do4JThVlUrF2p5NTEzYAUqqHWpRpVIxw0/LM0mW1mW/eZ5nXVmampq0efNm/eIv/qK+8pWvvBfO9rvS5XLq0aV7fO5zn9Phw4e1b98+dXV1KRqNanR01CqsoQpgR7PZrDmu2Bo4o6TM+/r6NDIyYj1boR/A9cTxIDU7OTlpRcAUKuGwsI84BCXZkAxJtlZoX8UByz6UZMMvsA+ZTMaui2fBOnQLrThYKeKiQ8h70eM71eXNMlU4JFA9BgcHjeYzNTWl4eFhXbhwQW1tbTXnUbVa1cTEhKHB09PT142xHx4eVi6Xs2fAVC/GjpMZ4XyDZwsViefl+76NkcdpBkGPx+MaHR21UbaSLPBwRxuTzWE9Mp0zkUhY0RtnSaFQqOlIIi1QVzhncfThnK+kLt+JcO1/8Ad/oI985COanZ3Vc889Zz2R6byQz+fV3NxseopEIrp69aoVuNE2bnp62rjtcLOZVQAlCLoJa811PiORiNlDQCaycvgf+Cqcxe7aBaHm3kB66/0tt48069W71gkE6oVrC2iNdzOw6Fad3+O+71/1PK9b0vc8z3vV/aG/kLdb8lOuLZj/Iq0+7cF9EPUoYf2hXS/11b717+cijy7aGolE9L3vfU/33Xef0um0FUpFIhHdc8892rdvn43spBLWbRYNxwzDT5TKWEAQD9/3tXXrVv3yL/+y/uiP/siuzY2cr0lD6JJnWY/+us+Tzfbxj39chw4d0l133aXe3l6b3JNIJAyRyWQy5rDiyEqytBtIAmk8HNfJycma6BQ0CsNB6i4ejxvCxMYeHx+3Ig/6zc7NzWl8fNzWg1sV77ZtYeIfCBlrZf/+/RofH9eLL75oesNgLHHQNYQuXT1eez9bfxis++67T319fUqn0zp48KC2b9+utrY2KwglFS3VNiancpeMCAclB697yFLJTDqVdDwTiXK5XA0vm4Ic0CeMPf23CVpw6gmS3CpxDHZLS4s+/OEP10yCu4HObiTvSpfLqUc4mv/u3/073X333brzzjvt+WezWZVKJV24cEGtra2GsGKv3HXMoYqDyuGay+Usdc1hK8nei8rsoaEhc6Tm5+etcId9OzExoZmZmZoUN5OhsHMcjCD0cIOxMfBYcZC5f/TFPuf6pIXiHQ56AtFisWgO5HvR47vRpetIuIigmzkkwC8UCpqcnDSEHroI633Dhg0aGxszB5lgETuYyWR0+fJljY2N1QQC7CFaoEUiEXV0dBiaDLJOhhO7MDc3Z52R2L8gdxSWzszMqLOz02h+OGBkUrEXBEjYaoKSmZmZmuCb65NkY64BVoaHh62Y3OkQ0BD2td6G/OEf/qHuueceXbp0SSMjI7pw4YLdIyg8EzDJSPLs6IDAvhgbG9PExIRGR0eNvnL16lXL8NBHGy48jiZrgP1Ehofn6YKKgHkgubQrw7/BmQapZggNgJfraHOeUj+FPt22bePj4xb83gyMuCXn1/f9q9f+H/E872uSjkoa9jyvz/f9Qc/z+iSNvGOtroIQIWDQiDxBHDnM3I3qosRLvZ+kmpQT6TJelxZaae3bt087duzQ/v37dezYMe3fv9+iXdJtoBI4tXCwaDtD2xg2O04xB01HR4dRIFz6BvfbSLrkmWGgXGTz+PHjeuCBB/TQQw/pzjvvVFtbm0Khxd5/pNVAEkKhkB2yFCWhH9/3aygKtD3KZDK2yUAdMdgclqAoVMVOTExYai8cDhtfkOIPunKgQ3TmHrretaICkGhQSO7n/vvv10svvWTjk6XasarO82sIXRLhu3SGUCikD3zgA/r0pz+te++911LmoLocnBQ/uQ4UxpWUF0EIxhZ0iQMQHjz8QLhkuVzOdENAMzU1ZfooFovWI9SlqpAmxQHi0ECv8Xhck5OTNUhxc3OzMpmM9u7dq1deecUM8a3SHhpBl6FQSP/0n/5To+BEIhHlcjmb4kRWg9GhxWLRWsFBC2lubjbnlJoGgkH0REcb3guHVpJ1ICDYZF/hTLsjw2nNhIMuyVL0BEygUW6BGgKKDJrFesKxdkeWNzU12eEsydBqDn0+fyX1WE8ZcrNm7I/z58+rqalJV65cMb24lC+Cl6GhIQ0NDSmXy1mP83Q6bd0V4Mq7lCB0GgqF1NfXZ45JsVhUMpm0gUHsF/ZES0uLZTGhS5BFoWsDxeJuQMv3Um06HmGv0YqNf7w364kR2awN3/etxz6ZwZXW5c3EPRc9z9PDDz9sBYHj4+OanJzU5cuXLWMxMDBg65q+9bQDQ5fQDQYHBw3xhW+Nc0lgiM0kgCEQxR4A6MGppmjR8zzLpFKQxnrBKWaPuWeIS1lxCynJ4rW0tNTQIdiXcJiXAiqXkrd1fj3PS0oK+b5fuPb145J+V9I3JH1W0v957f//9e7VuzLiIlJ87VIEUOKdd96p9vZ24zTxcEmT4GjWy9TUlDo6OiyChJC9bds2q26GQ9rV1WXpNgrb6K2H01soFGzBkn4gepJkKaJQKGQpP6pVn3jiCf3Zn/2Zpd6vScjzvHQj6BI94PiCypRKJX3mM5/RL//yL2v79u3mSF29elWtra0aGRlRJLIw6xtSPugd1cpsMpBYNhAVw7RAuXLliqGJOJ8cyFQHU1gxPj5uhqRQKBgSJsneA0OL4XEn+NXrD7Qok8nYFCUCoVQqpccff1zf+c53arhU0mL020j70g38QqGQPvWpT+lnf/ZndeTIkZrpQdJixW4ymdTMzIw1QXfRKBAGlzPm7jdS85VKRSMjI4bskQrFQM/NzVlxD1QT9xCnSCuXy1mASSofp6C+Ypi58276HfvQ29urvXv36ty5c9fZmZtJo+jS930dPXpU3d3dmp6e1iuvvKLZ2VnjE2LPLl68aPxP1iPPC4SRA48gx+0by8Hmrgnf942r5w5G4OfYRFKlOD8ESex1njV0Cqr+q9VqTaGcJNMzX5OidwNnSTW8XkAQ9i9O87XPXXH7Wl+n4ra1i8ViOn/+vLq6upTNZiUtOFKXLl2yYqaxsTErGurr69PGjRvV2tpqrR6hAUBHqlQqhhS6z5HMFx0C9u3bZ88MagKUPvQAiECaXpINqIFyEY1GLTiiOI0AGceLLiysJzKioMDoDGoHGRx3mBSBKij0aujyZoKj/rGPfUyvv/66KpWKLl++bAVqo6OjNRMtmYSHvSO4hBbS29urCxcumOOL/lpaWjQ5OWmI/YULF7Rp0ybzX6TFvtZupiGVStleGhsbM1oaAQuDZqCWsB7cMcT4X5wR7G+KxTkf3U4QBDKuM45PQWbpRnIryG+PpK9dMwQRSX/m+/63Pc97TtKfe573G5IuSvrMO9Tnigsb5a677tKWLVuUyWQMEejp6dG2bdskLXCBSPfANaMNDlGIy0FxCy2o+q/vJ4gBbW5uViqVUjwet555TFZhIbn/u8aN4gSugeIq3h+jQST+S7/0S/qTP/kT9/CNSHqqEXTJoUkkiMH+zd/8TX3sYx/T5s2bDZGgUfnAwIDdM+R50JyxsTF7xqTPcJzp2+qmbOBH8azhBGOoMcw43UTRjNTF8aEhOIaBCls3cwCahFPsZhcYesGQhkplodl/X1+fHn74Yf34xz82RIpNf00aal+GQiH9yq/8ij796U9r//79NqUnl8tpdHRUhULBUACX+4wzRNN8DrV8Pm/7j/3k+761uWIIBSgtzq3LeeN5gwqDnoPOs18kGZeTLgTuAemiiqypdDptqXyoM9FoVBs2bNATTzyhy5cv6+mnn35bx/eaNIQuP/nJTyqTyWhgYEATExO6evWqisWiRkdHJckKpiimIajEVoFGxeNxdXV1aXBwUBMTE+YswuNtbW01vfNMobKQhZFk1fdu1gSUUFqkBNHSrrm52dpHgvS7lCFJNXxfl/4kLTrhBNPwCZPJpK0vHF1qLXCgr73vitlXN9PiBp44hvfff7+2bNminTt32hAlSerp6VFvb6+lxeHab9q0Se3t7eru7jb0D+cFm8uzdtsEAgjwXGmTlclkrH3n2NiY7XmoNdAaOAfJHvAsoS64QxJYQ/R0xm6SCRwfHzfOP73WWUfuWnMzDpzj8XjcKCDX/q5hzkoy0tKC//Dmm2/auXb+/Hnlcjnl83l7NrT0Gx0dNURXWigEi0aj1u+3UqlYb3R8CrpmUNtCwMtURDfYdQEZMtRkmQF+2NMAOJy5LpXFLXTjfqVa+pskA/IIntEpa9DNdFMvUP8erryt8+v7/luSDi7xelbSB9/u7xtJnnjiCd1111267777bIwjCAKGk35zrpJJvVK84dIMpMXCOJQDpxODgaEksiGiZhHUUxok1SAkKJrqZtfgggzSWogNUCgU1NPToyeeeEJf/vKXeQRz/hItP1ZDl/VcHBZ8b2+vpcV4lhMTE7p8+bLC4bBtQBY2k7pwQEDq3UrvXC5n/V9bWlrMYcXw8T71TiZIOoU85XLZ2qhw6MBPIrJ1KTVwqIh+XR43LZ8qlYo5AzwHUIzu7m4dPnxYp06dMhSL59Zo+/I3f/M39fnPf16e52l0dFRDQ0NWXJHL5YymUywW1draarQBUFwOHaJ2kKqOjg5Vq1VDgmiLxfoHDSKVhmMSDoc1PDxsaDLOLo6CS3Hg4GYvFgoFK9KSFqucef+WlhYbfkJRDbSdWCymzZs3q7e3V5s2bdKFCxesD/GNpFF02d/fb5XYExMTGhkZ0cWLFzUxMWGc7HB4YWJhLperQeUJ8tGpJJuC6CJ49BaFsw/iiFC5T5ARCoXsPSTZYYmtpdNBOp02h5wgFyoG64rMEvsbgAN7Ds3Izc5Uq9Wa0dfX9GLIsu/7hjJqFewrTrBbGPvxj39cTzzxhHp6etTS0qLp6Wnt37/fgnWCvm3btpkDDC2Js2NmZsaGHZBNodK/vb3dPgskmD1AwEERGSgvNo3AiaEVjIN3i9UABnDQyAwSHDNkAeeJASoUO3Id7v52O+dQcIkew+HFYSkgz6uhyxuJmz0aGhrS6OioZmdndfnyZbOVZIHn5+c1ODhoYI/v++ro6FBbW5uSyaR6e3slLTjCrrPIfgHhBSTyrtU+AAhxlrm+CxSGqampmh7r+FFudoR9B11KWhycAwBBNhdKKnrlWWAbsAHxeNx6cEejUXPECXxvJOtywpvb4DkcDut3fud3dPDgQfX39xtfRJLB6iiHiSHSYjRDeh1EoFgsqqury/h+buqnWCyacwxnkLZIpPlwpkFsQQgpdsNxIhUO/wajDSLF75KeoJUIXCq+pt/pk9cmFDWa4PT6vq/9+/dbsUQul7PODUynmZ2dVSqVUjgcNj4aKF+lstAuLpvN1gQrbGaeH1Gh+0w9zzMnytUfyBIHdrlcNudLkiGXcExBwPxrlcwY53w+b8UDHCSgH3xeV1eXRfNTU1N2nwcOHFBnZ6e+853vvJdOAu9Z3KKL+q9939dDDz1kh1yxWNTw8LC1MoKSwtAAaCUYQIoROaAIBDCSPFMQJ5yeRCKhvr4+c6yQXC5nxRCgRBQtTk9PK5fLmXFnneAQ5HI5c4JwpsimUIXsOsEbN27U0NCQ3QNTprq7u9Xd3a3t27fr2LFjeuuttzQ6OmqO8MGDB9XZ2amTJ0+uih5dDrkbVF26dMmoC1evXrVAc2RkxGhfGzduNAcWtKxSqairq0szMzNKp9N6/fXXrXANW0mAghD4zM/PW/cWF3XE4YxEIjYinM4Nk5OTNcVe9E+nlyn3w37DEaunSdAlxO0k4DpCpI8nJibU0dFh07EIiDmMyQ6stPAMuN9Dhw7pn//zf17jnGYyGXV0dKhYLOrq1auWRWxpaVFPz0KzAhyYt956q8apL5fLFgy2tLRYRkdSTXAhyWwWdAOXnsA6YdIaxY88Y9rfQT1pbW01biqZP85MuMaVSsXafHV0dEiSZYPoH8y5C/2K7BDpfdehht98K3zRlRT33H766ad17Ngxy85MTExoenpaL7/8sp2hyWTSKCxbtmzR7t27lUqlLJDh2eKvQJUg2wHVhTWNLl2qF44nwAR7mWt1u4iQPQNpp+0dvlc2m1U8Hlc6nVYul9PGjRsN8IMG4dZmEWBxHS5NqlQqmZ2AxnEjWVfOL6gRiOgjjzyi3/qt39L+/fvV0tJiU4BQ5vz8vI1BTCaTGhgYkLS42Jh6AiooyVJqblsWKmRxXHt6emzTVioVM5gcnmxqnGkOIa4L40GEBQIhqcZBlmTEdO6jq6tLAwMDdgjF43Ht2LFD2WzWOgg0kmCgQGbGx8fV2dmpbDZrzwcqAc/B8zzlcjl1d3fXFLeQcvN935ApEBo2L4gwXFAcaTYY3CicAxr0u+nWQqFgaW82OEUWoFQciCAgIPpsVtAaqqPZzG6/ULixu3btUiwW08WLF1dNh64Bdh2UarWqT37yk+rs7NTFixfleZ7y+bxmZmYM/aWgj+bkW7ZsqZm5Pjo6as4w+w3DS0BQLpfN+XVpRKB4oG84s6QxJVkw1NLSYv9T6T0+Pm5IAgihi4RQVUywIsmoMbFYTD09PVZASc9asg+sC8ZjVyoV9ff32/WEQiGdO3dupVQoSTd1fF955RXt3r3bDtR8Pm8FhFBOQPZdNAaKEMGqtNh/E8eHvUUKFIoBzx20h8ybd60o1HWkyKDxXEGZyZq56B5gAM4M9oTaDRwkrp+1QrEb+zkWiymbzVpAxGFcb09W2vl10V5XHnroIUtZRyIRTU1NGTWHs4RuJdlstga1n5ubM0oX6Cz7iTMHx5fWdnwPj9PlV1MUCV3B9xcGENCxAVRvcnLSzmOer2tD3QwDwS+8fnoTQwckOOX+sa+cHzhUBEncF46dC2g1irj0x2q1qoGBAQMOoOt96EMfUiQSUWdnp/r7+7Vx40Z1d3fXUM04mwYHBw0c7O3tNSoZaLgblLr7C4EyWCwWVSqVLDOE/4VtwXawv91ezDi36Im/gVaIk85+dx1ft5MMnUwIanhfScvC+V0z4vaBq1Qq+q3f+i3t27dPTU1NGhwctEpODBfV+aSe4bC5jqbrtI6NjRmK19vbq/HxcTOw2WzW0CcMAAiXtMgx8/2FvqKkEqikdLkybsqHSAz0DCSYBeB5njm5bNrm5mZr3QJqtmvXroZyfl3OmrSgu7Nnz+r+++9Xa2urBgYGNDk5aYfSxMSE5ubmrEVNT0+PFUHgoEgypH52dtZQANIqbGppkU/ExnQ5aVTRug4VHDHSfoz5HBkZMcMEYokjjbEFvWTDzs/PK5PJmH6q1ao1gGcTF4vFmrnk+/fv1x133KF7771XX/nKV1ZGSdcEXfEsQJUIRLZs2WL7qFAoKJfLaXp6WpcvXzanCMelq6tL5XLZCjnh4RNcgNYSVNBpAGcUdN4tvgDBd1FcaeGgpkCNBvme5xkaSboWHUiyDACOD/dKAR3BLpxJaE3wVqHmuMU4lUqlBvXmfhjFuZJ6JCXN1xwUOKE4heVy2RxS9hY8azjpHR0dhhyx5kFcSIOXy2WNjIyoqanJeNjQzXCuU6mUFVSBQGLnOjs7TdekYNnPLpcXB4g1wv9QLtxm+i4tibXG4YrTC1+1Wq2aXskGcVBjT3C2VlrclC7reO/evdbqiaAFUAEeLPdC9wcCNgqu2Yf8IwgkK8UzdJ1ieva6Zw4cz2g0anxc0EK3cw0oIjbSLVzF6WJtwTXFtrMmWFcULsMFxsa4vgFr3a3NyOVy8n2/JmhuNDl69KiOHTumO+64QxcvXlRXV5fa2tpsDXZ2dmrXrl3auXOnjfqF+wtljHuFdkJmGhogdTisZ7IcoLuSavYBhaqcr+gf/VILwVoh+MYWEmzgxJKRI2NKv3bXJhBMNzc3298R2OK8Q2m5WbZ7XTm/0mK/w+PHj2v79u2G9rJxx8fHNTU1ZdXIpVLJuLJUgyJsVAYauKgcLa6gN8BDA5FyJ9m4Va3z8/MW6WLAcQxYTC6fGI5dKBTS8PBwTesfSUadwJC3tbWZs4bxLhaLNyV+r4ZwYLiOFV0dmpqajPPJpiRyJ/ioVCo2npRNDGrg6gGnig0JOkC6Dg42hpaNwwErLUSwVJyGw2G1t7fb+yaTSeMXgXDiAOAY0FYL54m1gYOXz+ft4Ae1IRhLp9PWO9HzPDPuKymug0gwyLOEHzo5OamRkRFNTEyYcQV1Io3FPrx8+XKNs0D2Ynp62ta6S3GgitlFQHh20C4kXef4gnZgGHF04ASyXnDKpIXAkf0FdxiH2K02J8CCepPJZGyPQZeoT7fyPy263GKulRA32HSLMT/+8Y+ro6NDd955Z00anO4lZFd27txpI9vhiXKfIPIUR/Eenueps7PTeHnYNdBfOID0HaUYxw30ocQQ9EuLRcZuuyPsNWlydwIc19Pc3Gx7lH840VwTvE+3DZ4ku1fWE39Hunylpb4OgNdef/11hcNho81Vq1Xrz8uz43mw7+od+Gw2W+M0Ms2L8xRKmVuoxu9SIOhW5HNGufugXC5rYmLCuNiSjN8Lmuei/GT+QNtdagyZCO4BcMjl9ONYYcspiMW+kqVbTXrZUhIKhfT3/t7f0y/90i8pk8kokUhoz549RsscGxtTU1OT2tra1NbWZo48/o4kQ3fD4cUWkJKMDuP6Nux5njE2E5CO/cj6cWswoLdwrvJ8QZVd3jwBDwKSSwYQSgzOLecrPg9BmSTznaTFTBRUmRvJunJ+XSSjq6vLFgAR7szMjM6fP6+xsTG1trYatI6BJerDMQEFAFkE5alUKjZaeGZmRoODg5YyxSATmUiLgxboOepyqkAeK5WKpdeIovhsFjkGGqnn2HAI53I5M/6RyEKvPjdt0QjiHljRaFSHDx+2jQa/1yXWYyRBE0AruGdQ01QqpYmJiZom5jhVOGwIGxcuoLTYh5cIkqiVjUe1OxsTJ5rrAsUipQhfFOSJAjrSjaAZGGomGoEaZrNZK86cn19o+E/6faUEA4beXH6uJL3xxhvavHmzstmspqenrc8jPGkQepwjqAZMWNu4caP9TTqd1vj4uEZGRtTa2mrcb9JpoIAYWrdQgr2AQ8OBjhF0U+igW9Bj3KyJ6wy5hzX9aX3ftzWIAwyVBSPMwe3y/d2paFLtOM+VEDerxRr/rd/6Ld19991KJBJqb2+39n6sYVDgtrY2bdy4UZlMxoYf8HzItriZEzJSdLeRFh0Ses12dHQYekobLSYu0ubRpXhRz8D1s+c4ALkGKBBuNk2StUJzA19JticJjgiKCIQJjEC0x8fHVSgU7Ewg/bySgi7RJ8Ex3OzBwUE7+BOJhC5fvmxoO9zqRCKhZDJZEwihA/Yt2U93oEU2mzVagtvFhmCf4F6SoYhQD/g7HE72IcEEgBC0wuHhYTU1Nam9vd2KoSRpfHxc7e3tGhsbsx692Be3D7iLaFPUR8bXzT5wrTcrkFotqVar+vVf//WamhU3OJVkdSZkaqi/kFSD6kKP5Jm74CD0BvYx2R8E34Nzk8+NxWIaHR214kMyqYBCZK7xcwCoCI54L/wVChjHxsasBsOlv7CvQaPpwV1PdSSgupGsK+eXAjaigrGxMRWLRb3xxhu2qGkKzYOHGA/C5vJrQRWmpqYUj8dtk8ERRYEolAUF75FIl+KqXC5Xs5hwsqXFXrEgjhghjDXX5KYfQDLa2trU0dFhBoPWNUNDQ3awcwA1ivCsfu3Xfk379++3McXFYlG5XE4XL160IpRsNqv29nZVKhV7hlu2bDH0HIPHsyTVKskc1FQqZc5HuVxWX1+fGVsI+JIMqa1Wq5ZaQ2ekeSuVinXr4MAFDYPXyGv83HWuXF4omxeDxHvi7Hd3d9v6IoW/GoWLbnqc50bq/Pnnn9fdd9+tYrFomRV6er7wwgs6fPiwITg8S2gEOBzSIucdjibGC+eFww8UmD1IcBAOh42fC3+QCmWXOwxiVKlUrGMD18ee6unpsb3D+4GAhEIhO0RJ+fMZLo3BDaRYC5LswKLjxUqJixSyhh577DGr+Ae9Ic3vtovq7OxUOp22PcF94AzhXLgtj/ie1mE4xThtPG/QeZxbnin7YWZmRuPj49Z2CQfJdbRDoZANUpAWkEsKe9AJwSvB0tzcnK0n9ASnkFQuTh2HOXrG+Q+HwxYMraSQheF/UP033nhDW7Zs0ejoqAYGBszBHRkZMXpOLpezTAWT2bA7rG+CVJ6PC8DwTBly4joxtJ7krMQpBxgCLeZswj66tCScXjItoNilUsm6enD2so9Z107bObPlINEuuk3NBqCCtJgNcSkZjSBf+MIXFI/H1dPTo3g8rtdff936JVNrAEVpYGBAW7ZssbMTf2R+fl7j4+MaGxszAIM1zr4gCCQLDchBdssFPLDT1AIwwhr+eFdXlyHPBIb4M+whBOe7o6PD1mA4HFZra6sh1tIiz11aLHIFVWYvUoR7K3Z1XTm/bAppoWp5YGDABhlw8NFFAfShublZHR0dKpVK+ru/+zsjxEsyXqa7gSjCcrszuGgCUQxV7rTegBrhtuzp6uqy1AAGQJIdJpLMALPwMPigW6QcMA6kGTFuFOM1Gu2hXC7rwx/+sB588EHt3r3bnmFzc7N6e3u1f/9+hUIhjY6OKpfLGYLnonQELZIs3UIaLpfLKZvN2oYfHh6WJEPkRkdH7dljcEG06AxAIQxRNhsLVBhkHs5uKLTY/5XUKJuzu7u7Brl0nQAMhpvGYa1RREJBjVt8uVLiFlu4jvfP/dzPqaenxzjnFGLgvKfTaR07dsx6aUsLzx+k3uWZYngnJyeVzWY1NTVlhXMEtSALZFlc6gh7zw1WWFNuqpaiDBB2qRYxxFGF0w0lCUeOQx7emcuPc6lOLpcOh4H+4VzLauxJl2r08Y9/3GwOyDtOHrYSBxT0HnuCLeJecAyhg/EcCWCoCmdfSItBAFXnBK1un1coKTxTkC/ew53c5R6oLlXKRf2kRfoCwAb7icJUskUEZwSi0Nxc+gzP5+2Ka26nuAHxc889p3w+r+HhYUNZ29raNHFtMmWhUNCLL76oY8eO2YAA0tmgwGQuC4WCFTPBBZ2dnbWx7m7hEX9LtpOA3m0/5hYM0+oOmpHL552amrJgDHQyHF4okgJlpzgvlUoZVxQhpe52MOAe3etxnWSCW4r1GkkOHDhgg2Z839fg4KAqlUoNj3t8fNz2CcgrzwCkH2eYLiAAFAAW0MbYW6wpgoimpibj0LPfqPOIRqPWpo6ACWACvi7oviQ7Bzkz4euThXWncXKdgBUEMugQoKOlpcVALAJzt7C6XtaV80uU6fu+zp49a4cpTaDhnMCBZbFHo1ENDAzURAsYa/q4MnYzlUpZY+menh47tF2Kg1vhSHEHqXL3IKZfIspkLjeLjc1LJA2XB/qFJCtAIVVHWpYDGxL/SiJMtyoPPvigdu3ape7ubnsuGN1SqWS9eTdu3Ki2tjZDTHFumBImyZAjnjk6m5yc1JUrVzQwMKBsNmuHwvDwsCKRiNETent71d3dbZyq+iI4EHy4UKRweO6S7JBAfwRRqVSqpvekG6SFw+Gaw98Zk2rOeS6XswIct9XaSkl9qlyS/sN/+A/atWuXpfanp6e1fft2cyBAlNBXd3e3pAXnd/v27XaPFPfBs8TIDw8PWzskOHrlctmaubOuZ2Zm1N7ebn0s5+fnbUAGlBWcH2kBRW5razPnBqQKfbgFdxy6oIUghjhZbjq9viAHqhQHEIgGKWIOg5UUFyGMRCLat2+fmpubdfnyZZs0iCMHnYDrBjzo6OiwAwrqiiRDAQEWQNRyuZxltAhQJZmTQWDpcjPrf5fApFKp2PWgX9YC3GEcHTewdZ1l9lkksjDlkWwfbdKgvfC5OFjT09NW1EngK8kO4ZUWwBCCaJ4bo7U5306dOqVjx45ZgHL27FlJ0okTJ+xZJZNJo3W5yDZUFJ7xxMSELl26ZJ9HsMf6hkrQ1dWlyclJ449js7BtPT098rzF1qLuXqDlGEWwZHm4Pyg22FN6/hIkMzQIOhKTx1xUm4AIJBjAizN/NTJrNxPP8zQ0NGSOLlllurKQaSPrS6cWsiPZbFapVMo6nkALoJ/+2NiYIb/sDdY11DRoflBSeHZzc3Pq6uqy+hTAITcDRiDEPsauup2t3PUhyfalSwvENyBAI3MEgi/JqKME6DdD8des84vj6jq8nufp7rvv1r59+xQOh/Xaa69pcnLSlAtSAan61KlT8jxP+/bt08aNG20zgZp2dXVpbm5OGzZs0PDwsB2kLAo2IJuptbVVksyJY/MRjRGp0JIE5wiUj03s9j+lmwGLkSiI1D1pN9cho1gH/g3VsY0kR48etRZRRKylUskOYZ4xG83lZvL7LkEeVEmSpWmam5vV19enLVu2SFro/wqNAsMxMjKikZERTU5OKplM2kGHs0TRFc+TwwE+IhxJd9IbCCNILocRKCIFHeHwwnANFz3k3qiWd504t6PHSovr+P7O7/yOPvaxj9mBCteaFmauUwkFB2Qc/TGJKZfLmSNSrVbV2dmpO++8U9LCoUnhDUV19A6+evWqPTcOb4IO+rWOj48b+u6m8V3kF2evUCgYdYjg0UWk4fzCP+OgIfCRZBXVFIK5XSHgoIGQgDStZCCDjZQWAom77rpLV65csRGn4+Pjtg6xdTj1VOvTQN8tUpJUww+GbjIwMGCIH510QBp5ThyqOB0c2HCo+V0CD/QPqt/T02MAAvu4r69Pzc3NhtiSbgVMiEQiNrWKLMTc3Jz1i8URo3csA1bIMsCPpKezpBVvdeYW4yI4pTi4OCAnT57Uhz/8YeXzee3fv1/SYkFjJBLRc889ZzolODp06FBNu8zLly9bpwBQQPa6S13iuafTaRs9HolENDk5aVlKSdqwYYP9DXSKTZs2KZ1OK5vN1nDI0QddOHD44e4DHmFvOQtaWlqMB8t0Vnf8No4elCdqABpNyF6Oj4/b2SjJnFzANrLT2WxWLS0tBtRIC+uE1nPDw8NqbW3V2NiYAUT0zSY4xB+h7oU96nJtCfIJlLBpUOPwa8iouXUj0ATdIkYyCh0dHTUdk4rFonXLkRbrMrBXgGEEOtjY2dnZ9dXnl4jXRU1Ii/7u7/6uenp6zKmMxRZ6o+Jo0Hh/fn5eXV1ddsiycIhYiYhRyvnz52sqEIl4mZkdDofV09NjCJW0kEpl1GahUDBDyQJhkc3MzBhaQr9X0u7042tvb7fJWDj97mhf+KO0NoGzR1QEf66RhMOtUlmYUc4hh1FyCw2j0ag6OzutMp9N4xY44mTiFOO8sgFIjaRSKWsCLskOREYLkzrnfQg0SGFLsoICDAM6IDCi/YvLJXQ7CEiyHoYYoLa2Niv44m84HFiPHDorjfxKi0h0pVLRsWPH7H6QQqFghnZ0dNSoA3DYOeBAx+ljPT09rVKppJGREdMRhxEdBaSFwxKqSrFY1B133GH6IpUJD81d9+HwQgcNaSEoIutDAQbPGFSD9C4Hqe/7am1ttTUFeuF2AXA5y9Vq1doMQvcgYMKBcw+lldYleqSgaXZ2VqOjo8rn8zZEBl2C6pJNGR8ft9QoNRKsZwpUQqGF9nIEO2RLeOZ8JgEqa4H1Ui6XrWiUNC0Zgc7OzpquDE1NTdavG+eGAzObzdag9NLCvgUAAdlndDPrDU4s/G0XVSK4JiAAeGjEVLm06CA//vjjeuihh2wiGACQtBDk9Pb22p6hfRQUBc4RisGxT1AIOGvJGkDxA5mjZyuFoHNzczU90rG/kqyVaCKRUCqVsimelUrFzm1AnGg0WoNAQvFgLwIEEXACKuB4gYDSM5xznAC6keT111/X/PzC5DbWq7RYJ4RdwzmNxWK6cOGC2R18hno/YXJyUoODg5ahcQM5HEo4wYABdGKiY42b3SLrRRbX5WlD34ReASgxPz+vzs5OA0PIjpEhw04CfjHJ1i0odm0sSDIFzjezr43lEd2CuJEu1ZylUkk/8zM/o0984hPGy8K56O7utsj+/PnzZrBdxA5kqKury9LncMFKpYWJIfl8Xq+88orm5uZ05coVS4OzYUjh0n6pu7vbyN6M6+TaaeHjVsqCTLlOfSy2MKebTQ4K2NvbW5PyR/kYdTeVgKPWaLSHiWsT0IaHhw3Vw1iDmlarVeXzeXV2dmp8fFwTExM1HCLS3qS+cVZwqEF1SA9yWGEgSIOl02njCuH0cvhjFEmXgfqCfJGydfmwrAc3xYMRh1dJezPS8DSUdztSkDUA+YLHtNK65PDgHpPJpN58882aTAZrDsNHGhTE+0aFKRRMgMqk02n19PRYdxR05hp6eiKTKUmn07ZvQRrg2uIs85noEf4a68ztF+ryrrkubA3i6oqKZhwl9M3BTKAEFYA0cj2H+nYLfGeCtvPnzyuRSNhQnNnZWUOEQL25/kQioc7OTlu7OHs4URx2Li2iublZQ0NDxhvlOc7NLQ4bkFQTKKEvSTWHeSaTsSb5gBdupqUeEQMgwLZKsmJZgmrXHrh90wEgsDOgm+xLN7ByufmNKh/4wAeUyWQsQ0ZmguFLGzZsUDabVTgc1p49e8yZamtrM5AB1JcMFFkLHGAcIhA/OO04sDw79E+RE04t4to/92cuBQYnGPtTrVbNCcLOuM6Pa5sASNra2tTa2loT0IAaNxpQ9MMf/lD79u3TxYsXDYDxfV/nzp2T7/vas2ePXnvtNUnSrl279MYbbxjogBw6dEjpdNraFhIA0AsaugrZGBxT6AigzrFYTLlczrIMtB6FpsD/BBrlctmGGAEYAOBgI7HL0MHQGY409pq9yhqgswXgx+joaE1tTn3wWy+NpeVbkHrEl032wQ9+UJ2dnWa8MEzMK5+dnVVHR4ehOG1tbUYrIOLlwOzs7KzZ/LRK2b9/v6anp3Xx4kUrzMHQguLNzMyoq6vLyPwcsqAgKIQ0KIcfKRicKYw+jlQ4HLb2bHCoKADDENMbkb/nedUj5Y0gZ8+e1b59+yyiJY1GkUs4HLbJNKOjo1Y9jBPBYcRBdvHiRSuwIFWSSCSsNRFIO0UukgyBhz5BCx7XgOPYXblyxVLeOGlu8YDbbxnkg7Q3ARJoB44uTjSHKylHPpfBKRhvgqGVRgvdw+mBBx6wFjTDw8OGhnLfkrRt2zYzkKxh+mpjvFxJp9OWzpYWB4/gVMbjcVsT7FVpcVoe6KI7CZBnPnFtYp/L3aa7BygzhwR1ATjE6ITnTSN/l8LBvcM3xKFy2znxe2QSXK7wSlaWuzSF+fl5/eQnP1F3d7empqY0ODhojrB7D25Lo1deeaWmVZRb/MVhODk5afq5fPmydWBwkVj3eWCL6bLgIsboFqCCYseZmRlz3qEicE3xeNz4nS7fG5oLDuD4+Li1fZIWe7rDK8T2uN0ecIQjkYiSyWQNt7wRByNw9hDIgFi7hYfcX0dHhz2PRCKhnp4e0wOditzgIBwOq7e3154fiOzk5KSdl7QKJbU+Pz+v9vZ2C6BisZiKxaKi0aiGhoaUyWTsjKSAyy1K43qh8jEdFRQQrjj8bK6L94FDjB1myA3A09TUlFGSGkleeOEFlUolnTt3zvYZEgqF9Prrr1tGJ51O68CBA/Z399xzj1pbW80W3XHHHUokEkZlYCw4jmRXV5fVIuXzeUkLdqOjo8PoDqwnEFw3wHc58ARN2EeABRxrnjPgHM42ACXUCt6X92DP4TsRAGPH6RABAHMjWVPOr3vou6mUw4cPq7e3V2+++aals9zpTa7BwqhiDNzhCKCl9LvjsK1UKrYwKpWKtm/fbpXopOb4fnR0VBcuXLB0C5ERi5ZIhjQABzCTo1xkkmKNrq4uKx6anJysqTBm1joOFAsMBwOeYyMiE2NjY0okEkbgpxCJ1KqrK3hNg4ODVr0MogPdZGJiQu3t7eY485x4P/Tr8j+bm5vNUabwhd+XZDoh8BgeHjbjjO5wSkEM0C/pRQpypMVG3qTZCd5GR0eN3sKBwnonZQ6isRoFGQRszzzzjD772c9qeHhY+XzeRoLD+YImItUW5fC8Wafu3iQIcNNkIDXu3uJ5EHhIi8gjzgufPz4+bmlXUEfWA/uRNLjrTGM0QQzdIiz3d0Ap6UBBsQw1AG1tbZYNctEHHCh0uJKBjOvMStKPf/xjHT16VCdPnjR7Kkn33XefPV/s1OzsrCYnJ23AC8gMgSh7kfWOE0Lw4XLa4QgShILWVSoVQ31B6uPxuPW2Zg0SwGYyGQMGyuWFXsSTk5MWpBB0sbYymYzy+bxxRnGW6LjC5+LIcW6MjY1JkvHK3SwbAZ3bPaRRZHZ2VocPH5a02CJzcnLSbBtOCXYKmlZHR4eh3gSdoMG00gJth3rHeUkmraOjQyMjIzVrgDoOAAuKU8fHx7VhwwbrKOIiflwb9hKdsebgl7tBOPqgxgI9cX+SrFaAAAmaHUWVjSTlctl43EgoFNIv/MIv2D5paWlRS0uL2cloNKrNmzcb4Nfd3a1qdbF9J0EkXatAVqFE8DWZLGkhmIJHTxACXQrOO+uCM8sFIlkfvB+BK1QHXiMLxffYTwrcWL/VatXqMDi3caChUPBZS8macn7dQ59DiaiVqmWMKAcUXF3SXDz8TZs26cKFC/Z6OBw2hwZniveTpM2bNxtdAUPO7GxJxkvavHmz7r33XnPCqTx1ET/uA/SSqKWeZ9Xa2mpoFk4Az4HfJbLG0YazxftIi3y8RpPvf//7On78uCRZ4OBOb6HIhurjSmWhdQooFanTjo4O4/5JMjQWo0fHBH5WH+yUy2Vt2LChBuWvVCqGJGEkpEXdED27DikOtCSrnAbNZVNySJBlgL9MsSTvi0HGeEEBWOnCGmmxLRsyMDCgqakpK5Jy22G5gYXneUZTcZExd+ykSx2hDZob8btFdegYTjHPgmJS2mx1dHQoFApZUJXL5QzRYl+CwLrZFrfIEgcLJ45D3UWzKfQDXYRbzDoiveg+I9CrSqWy4vxt1qIk48edPHlS0iK6H4/HtXHjRuP3UnmP006rK0lGFXK5llCTQIilxWCFwAKnJZfL2ZADtyjGzWbx/vCsOXRpV8g1cvAT5ENZwx5SACPJuuO4qD9ZAX6PQ9y9T5yC5uZmK3zjTHHrPRpBCC5feOEFPfzwwwbScA6FwwtFxOxRApm2tjYNDQ0pkUjYmHL2UjgctnZ9cEZxKikCjEajSiaT9rfutYDUUugM6IPtxZbSeoy2WdAnuFaCMVLkkoxuxfnMXsSZ5VpwpsnSSgv7YmxszNZKo2VJCUxxOiVp7969+qmf+ikVCgWrhaDTjrRgI3t7e9XS0mIdjGjfCTDAHjh//ry2bt1qfHfairkOrect9F7P5/NWm0Twii0h0OAMI5vqZvzcQlmyrNKiXqTFseEASwBfrhMuLexHitDdrCF2x63LWErWlPNLBOEarUqlYuP64I9Ki1A6abjR0VF1dXXVRIKQ6nGS6eXJFBw2ErxL+KOgq6Ta2GAQ9SmO4T3D4bAuX75sxpuCFw5EF0FhgYC4RCIRjYyM2GcQ2UoyBI0FnUwmLdWFkXa7YTSiPPXUU5IWkcVQKKRDhw7p1KlTNdd+/PhxO7xA/OmokcvlavopwwOCv8WoXVBkon03VV0oFIzCIi3y+jgwEomEVcWTBuKZgvSy6SRZcROHOzol3UjxItcIpQLnCOMtLY7/hR+70kLqHgf+xz/+sTo6OjQwMGAIKKlGdMh6ZYgMzxXuLNkNnFxauXEo0uvXLUSj6EFacKhclJG96hYRYlChitCFgvvh2XNwwMNFRzh7pGPpZ+kW22B0cdApjsSZI22P00HqGX5yfRrzdgu8X/dzeS0SiejYsWO68847NTo6aigPwVpvb68F8pK0ceNGS7W69CHWeUdHhx2AHR0dVmCEDQONJwDlgGTvwdnl2nCCWFsuDx3niDOCteA2zafgkr3IexHgplIp0xHZPtcmk00qFAr2t9Ii/abRhOdKQTFZFtY+NSEg5VNTU3aGkNGhqw2Bm9upg73C3pBkax5dMp3NPbfg4rpdGUhtQ/HCYcJBJfAnuwM6y/XA2ecaQZHre7MzaId9DJ0llUrZOmg0XWJ3XRt69913W7EggTgcaqgjZLimp6fV2dlpdUqAcugM5J59RO/vUqlkjieFkjzHcDis9vZ2NTU1WY9nAkCcYXw0rpssjUupwmbyO6wRt0AR+ii+EwOrQKcBJNwaKIpQbxbIrCnn103FYeR47cKFC7py5YoVrQCVDw0NmREDgSFlAjpDBSQV7NPT02pra7MocH5+3sYZ8ztEz0SrFAbg4NJ9APSKtJG0OE5TWhzTCm+M6CeVSpkBondhsVi068Qpo0Kb7gIYuFQqpUgkYjzG1egQcDNxHXI2tud52rt3r20QV5555hnt379fuVzO2hGBuvEeGOmOjg61t7drcHDQAgNJZoDRHQMt2NxElwQQsdhCf2YObAIUis/QAQc3Bsf9fSrbpcViDdYE6fZyuWxoh4uYuUVVtLGhsGqlxUUennzySVtTBw4c0AsvvCBJOnz4sGUgXn75Zd1zzz3yfV8XL140VA29kprkWYCc4+y47QSbmpqsgwIBD+uZteO2giODQEoNvh+OmevsSQsoCaluPpt9mEwmjc6AfkG5SSVyHfDR3Q4k/E/dAQc/qNlK67Kej+ciMZVKRQcPHlRPT48SiYQ5e/F4XJOTk2Z/UqmUfN/Xzp07FQ6HLT1Kn226e0BFkBbahlFDwaG0efNmAyukxaIjqDJjY2M1KW/eF0dMkiGTlUrF9Oi2A2QNNDU1GYhAQINzhUNFUOMOIcHOwItk/3L2AKQ0GufXRfi/8pWv6NFHHzXqw8zMjF544QWVy2UdO3bMqF9tbW0qFArWy5f2jXRuQMh0ga7zPcBALBYzvVOMPTw8rM7OTqP3AUrhjJOBhVfsBiOlUsm6/hDAkK3hvJdkgQxZJElGKcN5IgtBEBMKhcwpczmsjSTYXfYs1C2GXDDuGIoJNAhAgfn5efudmZkZK/LnjHKzMryGT0MmzC2aB5Rra2szSiagklRLdyOocukOLvDneZ4mJiaswQCf4WZ+uD7OURc0yGazNdRC7K9bUH0jeVvn1/O83ZL+u/PSDkn/H0kZSb8lid4b/7vv+391K8p8L+IiB2yWEydO6N577zWKAVW9HKogg6QzU6mUOcSk4CiswXm8dOmS2tvbje8JOogDAh+NCITmy3D+QBBCoZCR9100i+iYzhQ4RlSzE0W5n+uikvWV4txDW1ubVcLivPMekuKe551xHueq6hLhPg4fPqzDhw+rtbVVO3bssPsGWZifn9eZM4uXH4lEdPz4cevHCScoHA7rypUrhjwyHx7+J8/O931rp4XjhTEgFQtSBAeVIigKJrq7u+0A5HBxOaEgk6RlWU9EtqSBWRd0AigUCurs7DQDzzWRilqNfelG8Pz/wgsvmFE+ffq0OVS+7+snP/lJTYDz4IMPqlqtqq2trSaVWi6XrQF7S0uLDaZhj+AowgMnyCR4SCaTNh4Vagy8bJ55NBpVLpezkZmkt33fN6pLW1ub6VVapCVJsjZ1BLnYClAHDn3XMMNbxMbAWcdBdIpFzjiPeUX2JHpxEWDf95VKpZROpzU8PGzFZ+wD7A7DZ2KxmLZv326HHi0Fqae4cOFCDcUA1Lu9vd26C9BairQ1dBQ6Q1AtDm+bHtk4wBzoOM3ZbFa9vb1mTwuFgpqbm604DR1NTEzUBJqZTMYQx3rOJ3UhgAqAG6wr7I4a0L6i3x/+8IdL/hyn/dlnn9VHPvIRjY+PG00I+pILGNAWsrOzU5IMXMBpgboyPz9fM5ly8+bNhqxCWSsWi+ru7jYdeJ5nQXY4HLasjctJpZ4Hvq6bTZFkzrS0mJFrbW21TC2AF3QN52yUtJBRarSzst7uStLU1JS1AATtJtD/y7/8Sz3wwAPq6uoynyCXy5kuANJ471KppKGhIe3YscPohvRonpmZsemPtGnFfuOQkt2mnSu2kK5ZXD9Z9/HxcTszcZo55wAz8K2gOwKIwfMl+MS5xlZDVyM7fLOg9G2dX9/3X5PUL0me54UlXZX0NUm/Juk/+r7/f79zdb57cbkd3Pzc3Jx+//d/Xw8++KDK5bJOnDih/v5+qwolZTU6OqqOjg7Nz89bsRVRSiaTsX68VLhOTEyoWq3aRh8bGzPHFueXhQfXBA4wThgpYX7H5f1BWUCYJkSq0E2TkxIicqdqWlrsVcuzIYUO31mywpqS7/tHrn2/qrrESZIWN/fWrVu1a9cu65FMe7NyeWHsaVNTk/r6+ozqcurUqZqKXVrVkaok6KCDhMsF5rnhYEH0d0dDupQXHDq3mwAbFvSSrAAbn96hLS0tloan2wB8UQrA3KIbDgeK66DgQNO4di0Nsy9939fP/MzPqKury/pqRyIRjY+P6+zZs0aZeOqppxSPx7Vz504LGuBtEb0PDQ2ZMwq6i15wOkulklELaLXk2gWaubu0JZw2z1scic3eTSaTikQimpqasmI7DCfBMqiTKy6PjYOWz2Lvx+NxZbNZc+BBGjnQrwVI/dLq61GS0UYkaWRkRKVSSa2trRoaGpLv++ru7tbY2Jhx0QuFgtrb261/p4vOgvAQKKA3Anv+5/AkiCIlT79dtyXl2NiYOS84syBGbrqa19nzblEwvweyidNPIMqhjM2dm1uYFAri77ZKw9m/ZlMaxr7eikQiEcvaSNK3v/1tHTlyxEAcSTV9zEOhkDmP0ArIlJLBxBmSFp01lyfMXmVNgCqTAscmM60R4ADbKsn6OycSCaMkwUvFRlIPAsUHJ0qScbZp6Uf3AECOa9ffsLqMRqN6+eWX1dzcbENnQHjn5uZ03333GSeeoU3nz5+3gBWuN/aOgOTSpUvmjJbLZQ0ODppdzufzisfjamtr08jIiHUGcYN9zl1Jlq0kS4/9wwbQXYOzG3oDDi68Y4AO9hg2W5L5dG1tbbpy5YoNZHHpPe/J+a2TD0p60/f9i6udRueBu10ZfvSjH9nPT506JWlhLvaLL74oacEBPHbsmIrForLZrDKZjKGjw8PDkmTV4m6hDaijWxWK0cTwgiJwUGNAJRkyAb+XllvA9BhnEMHh4WEz0qCUOE4uGkU6F94c02+SyaQh0PzNErKqunQ5TKR1uru7ayrycVKIBCuVirq7u5XL5dTX16ejR4+ac0oEPzc3Z/3+QHukxVRLOp02xwl0g64AoDotLS3WAzMWi5kDzgZ004Au94/0JwgSDg5UgEqlYhNnQFHoR03aMJVKWWDmjvZkjdygz++q6vJDH/qQjhw5Ys4ohS07d+7U/v37ba0PDg4aV1uSpaMHBgas3VUulzP+NYdZpbIw0x3qB7x6Ch1wWkCcXF2xhkAZcbKy2aykxS4EIISM42Vfg8pDKwKZJ9UHf9ItgsM+IKx1UEUc5CVScquqx2g0qjfffNNaU3GAeJ5nBZ84sdindDqtq1evms2UZClW2lGVywuDLNh3HII4UwSO0mI3FLIi8N1ZQ1xnKBSybh10iOG54pjxDPkZQAlcb5xb0ES6OQBSYO9xzEqlxcmAVLTTTQKKiyMNc1beSFz7Ky04q88//7z93K2teeCBBzQ9Pa0zZ87o6NGjlpkqFArG9ecZZbNZs98AGTi8BO8AU2S72JfsO847giEXqcTBRTeSbC9iF7ANIJbQqlzuPo4z1B8+092716QhdMk5OT8/r1deecVGxlPMTSaOKX5NTU26dOmSZV/gP7vPiqziyMiItffDyWQqH1kf9j6oe3Nzs1pbW22NcEajQxxVF7mmZR7fk8mh44ub0ZEWMhOnTp3Snj171N7ebg4t4OTLL78sSTpy5Ih9Lq3O4vF4zXqul3fq/P6ipC873/8jz/P+N0nPS/od3/evmw3oed7nJX3+HX7ODQXFcIjycN1WNSzQhx9+WJLU399vEe6JEyfsfe655x4rNkokEmpvbzdI3uUPwT0kjeq2dMJBcY22izjgDEMIJ3WKIyMtTqvhMHT5wC4yiLMMGgH310WV+Wz+nm4JS7RvWXVd4txAN6DwAuI+DqbLJ2pubtbOnTs1OzurTZs2WZspECcO0mKxaG198vm8FTtOTU3ZGEXaq1HMSPrObWCPvlOplKXgJNnzpGsElAccJAzz6OiopXAl2QAA9AphH11ykLqHPo4Vkf0STtOq6nLz5s02J765udnuFQPY29trjn5bW1sNdxZuL7xR6A+sdWlxH0B94P5xpphdDycfB5VCRQoyMO50hBgeHq4pjHQzMjMzM+Y4cx1kXaTFKUik3blO1hX2AEQDOwL6QmueOl2uqh7pChONRjUyMmLFolRU89xAchKJhLLZbA2PE1oQlA5sz9WrV7Vhw4aajjogOnRyoACLwBDUCttOmzT2Fg4qfXwpzuFw5QB1qR2k1MmiYTPJQrDWQMPYg3AfoT/BGSaj04j29Wbigg+SbN339/dbEMnZyH5samrSAw88UGOPmpqajIvJVDXOZfY3SKK0ONzI7YGN/SZwhO4GrQHQiDWF3tzAk30cj8c1NTVltEeQXWmh/gdqBb383WwNHOYlpGF06RaI/c3f/E2NI4mcPXt2ySJ3XuM99u/fb+v+9OnTOnz4sEv7MEAP1Jy9z/7hPAX0cdcEfpBLmaL+iGcNNcnlfpOJffXVVyVJBw8elOd59v3+/ftr2r5xT88//7z27NmjeDyuF154QXv27NGZM2dqOmTUi3czQnDdg4tJGpC0z/f9Yc/zeiSNSfIl/TtJfb7v//rbvMeytBzAwLkE+fr7ePTRR/WzP/uzunDhggYGBgy5YJJYtVrVyZMn7W/7+/vV1tZmPVvT6bRB+0SnTHpzK5Cbm5vNMLS3txufmPfl/9nZWUOb4LuxsYHr4Yqy0TE6FK+xcDEMoLpEZFu2bDFHmhSCWwX5r/7Vvzrl+/6RRtKluzg/97nPaceOHdZmiedM4ECk3t3drR07dqhardpoTnhJBCBu6zNS4XAup6amdP78eXuGnZ2dSqVS1iLL7cELxQUEA8eY5wqnm44jzLYnICG1A2KEoZZkiHM6nda2bdvU2tqq4eFh01t9BSuHwB/90R9paGjIu6aHVdflP/gH/0AbN2609Kckq9YGUWtubjZa0d13322tlGhEDjLKnHlSr6C8cHiZukc6HQQHgw5qQ1U3TgpZlXJ5YawvASz8UZBLhphIslZlHDDsO64LJAIUk+wO9AwcNRxhDDv2JxQK6S/+4i80MjLiNYIekUceecTSki462traqkQiodHRUW3fvl2ZTMaa37tAAfzmS5cuqbm5WXv27NGVK1e0c+dOQ2nh44IWMmWRIieCyHQ6bbzG8fFx66aRSCQ0MTFhE6PoKoLDS+cenj/cQOyoW+hHy0HaDhJA0cWC9Dtt3+CEkyEANfzGN77RcPb1bT6jxgn+h//wH9Z0LJGkrq4ujY6OamBgQL29vVZsDHqLozw9PV3Tc75cLmtoaMgczJaWFrONZGukBQfLHUpAMDk2NmbtzAC0XN24wQaBGudhtVpVa2urOjs7lcvlbNAVPfc54136C+gyXWuef/75htGli8DjT7h9uV19un4Q+pVq2xwiR44csTqWaHRhciyf1d7ebv5Ib29vTTEy15LNZpVMJq1GA9oK3GL8Flq8wsUFDAIAIXNN5tb3F6bX7du3zxxdaIc3uneXrrjEczsFhaXmub4DHXxU0mnf94evPcxh5yH/oaS/eAfv9a7EdSilxe4PIDOgtaVSSUePHlVvb69aW1uVyWQMERwcHLTNB2eN1hmjo6Pm4NAOBGdofHy8pqq8UqnYqEyiV/gvpGBBeTHA8HhJ6bptkohA3fvint0RhGxYFM3hFIvF7PrdVmAur9WRhtGlGyAMDg7q2vXo6tWrhhri3IMWzczMaGBgQDt37rTCinQ6bWg8AQ6Tmly+2bZt21QoFNTT02MoI1wy0nXS4ihUgiwQDJf/5h7EbqseOkmwTt0Nye9QtAG6zUEOkgXiDMoBMlL/fmoAXVLxjWMK6jo/P6/x8XH19PRobm6hlzHUEknWN9fl7VGIhvMK+gZlBeQeBJigNBQKmTMMT40sCc+bSXEcoHw26B372qUKETSRXSKTA997ZmbGuiCgSxxcggFStPS1dLtNOJXlq65H1uuTTz5p37sBN46S53lqbW21dmjYQFLZ0FQI3i5evKhisaiRkRErYHN7WVNM1tfXJ0lqbW21nqVMdwOBps0V9hbqhMvpZJ8AIGCXyRxw8ML9ZL3St5fXsLHQyNi7bncBt9jIkVXX5a0I+oxEIrrvvvt0xx13aHx8XE1NTWpra7MhFNu3b9fGjRvl+761zHJbT4Hczc7O2p6mJV6lsjg6mswJhYmxWEw9PT2mN87oQqFgvFS3rzlFxJLsTJybm9Orr76qBx54wJxYgAa37/jo6Khl5BjTnclkLJChOHVyclKnT592H9Oq69LNguHAus6fe5a6LSldh/fgwYNqampSZ2entmzZotnZWet2wbyCSCSirq4uTU5O2rnntiKkoBUgMZlMWsDg9mfm2tDXyMiIJJkdZZrr5OSkXnvtNW3fvt1qLQA1du/ebffN3nUpLvWILtxenGSe283knTi/T8iB/j3P6/N9f/Datz8j6eySf7WMAreXm+LwcB27Uqmkw4cP68477zS+Lunq6elp9fT0aGxsTFu3blVLS4uGhoa0bdu2mgcJ/1eqTbmC3Lrp1fn5hZGN0mJKm3ZXdArA2GIgMNRuGx+4RvCZoE24Fe7MZMepJgKHzwjqxjWSWuBQd2TVdYlzwgaNx+P67ne/q8cff9xQOJwWUqxwfnEE4fZh4NyWTRQv4IDQsYMiM/TEAQYXCdSPQ5wgwu2cgTF3J9tIMkORyWQM3Uc36FlabIHl9lScmZmxAQEUj2Dw4P+6RsWRVdflpUuXrIDEDUgxWC6tgGfqtp0DNcWZpC0S6fO+vj7j7FUqFW3YsMEQddLQ0E5cp9Ll3rtce0nGWXQRIX7GqHNGsUqLPSpxgF27wHqCYgSPTlpAj8fGxtTW1lazr+scX6kB9OgiQ+7hwnXyc9/39cwzz9j3hw8flud5ev755y2DBmInyVLooHlDQ0PmLM/OzlqhKocqdAff9zU8PFzDJSTNXi6X1dXVZSAF45EZ9ONeN88cx8vt5woqjE0HUJB0XUs91ip/4/bLrUPVVl2XtyJcc7Va1V133WXtsaampoz+4q5/qAXt7e2WuSiXy2Y76QhBISJIIPseB5Mes5zd/C77iqDWBZL4e9Lks7OzOnPmjA4dOqSDBw/qmWeeUX9/vyGZjNXmnuCOHjp0yDILONixWEwDAwMWvNal1lddl/WOrLsm3doBFxF1wa9f+7Vfs3XPdD6ofPPz82prazP+MJQsQAx6rgMosNcAEUFruQ7+kQFwC8DJDgEiJpNJ7dy502hpBDYvvfSS3Yd7ry746RbZ4exLtTz2G6HEyC05v57nJSU9JunvOy//X57n9WsB/r9Q97PbIi6flw1y7frqr1dzc3O6dOmSFRrR6giHhUrjzZs3q6urSxs2bLA+dzihOCTj4+Nqb283xwsk2L9G4MYgchBS8MHmJh3HgUmfTBaa21yclCwLVFpMlbsjein4kFQT6WE8cNLoHes8m4bQpYsiSYtdPL7//e+bXg8cOGBIhBsFptNpK0LhfeCJRaNRRSIRzc7Oqru7W8Vi0Vq/4SzxmaC1vu9bgETAkc/nzRiCUOVyOePAgUK6jdXdFnY41rTZ6+vrM2MkLfY0he/o8kah2VBQksvl7Jm4KH6j6PIHP/iBIWcgphxmoG/QNuBZg6RVqwvT9+CCIRSPudxRdywr+0mSBSbsD4JPSRYUEXAy9Y2/k2ROF9dOZTOV7bwPRXTcG3scBxvEmmuGL0ealeAY9JfBDI2iR9eJY29ykNXz5tzg9dSpUzpw4ID6+/utiDSZTGpiYsICfGhe7NtUKlXTEonn5vsLHT+Y2sZnkeYulRYH1HC4ki3AbuK0uT2NydgAWGCbOXBdChn7y6WskP3BoWPtQdkBeWoUXb6dEHyxx0ZHR3X16lU7I9lrOBujo6NWYwH6zpnHIIzR0VHL9NCqMJ/Pm50lm8V7z8zMmE0lBU4GSJJ1WcEh59nTsrK/v99Q2mPHjqlSqdi0OvQzPT2tl19+2dY1/Y2PHz+u9vZ22/vPPvuspOv6zzeELuudXjcIdR1C1wlkf3jeQi/2iYkJq4khsMQZlWR+iQugMUlzdHRUyWRS+XzedE8nJT6LjhPomvXBZ2EjCWLq1x8tTF1b437tOtUEXvhb7jPhfZdCyOvllpxf3/enJHXUvfart/K3yy31vBX3NRbCqVOn9NM//dPyPK/GEcVwDg8Pa9OmTTaSlEgVygNVoaTMqZKk/QoLJhQKaWxsTMViUblcToODg+rq6jL+E+k4xjnSjw4+IP9w1t22I6RjMO6kiUA/MCRwRsfHx9XW1ma98kKhkPFVMUDXnlVD6NLdwK64Ac3Zs2ftED548KCkBYM4PDxsHRno85jP523DwbkuFArq6OhQuVy2wzGVShmX0W1mj1MCX9f3FztzEAmDLkmyfqQc7DTypgch66qlpcUOfqJdd/IY+uLeQahYP+Fw2Hh2GH7HIDSELiXpe9/7nvbt2ydJOnnypI4cOaJTp05p37591qM1Foupr6/PxmsmEgnl83nlcjnbg/F4XMlkUul0uiZNim7q033S4pAQqp4prJAWnDiQItAJDKZ70BJ8VCoVG88JqgxC4vuLE/agWRAo48y72QG3rRPtmwh6CK6uGfWG0GP9XnSfoSscWI899pgKhYI2bdokaXG0KLYI6gPt6KBD4NzSPB9EyS2AI1Pm7l26B7AHyd5ks1mNjIzYaGx4pRzupNKhz1DvcejQITv06bTi+4u9QUGnXWQSTjk6ZrCAk5ZuCF2+naBTHP2hoSENDAxoYmKiJhjlPj1vYRhBPp+3jCqZTLJbfE/GsVAoWLssggxoMvF4XJlMRp7nWbCC/cvlckqn00ZXgRYIvYYAVJLpsFwu69SpU0vW/7jC+fL0008b2njHHXdI0nWOVaPp8ma+T71DLC2sX+gDgCbULQFGQONxqZf8HkEdtRHSYu9kF2hDLwSy7J+WlhbzuVyaJ9dHE4I9e/YslQlbslCN125UxFafpbqZrKkJb7cq/rUCqWw2a5w83/ctjQkJGwOL0wt3kP6eQO0UcmBse3p6zJgzIrm3t1ebN2/W5ORkDf8L5BIHKJVKKZPJqLW1VdLi4exWK+NwYfClxd6vkgyNopDL8xYn1NEfj8p7Ws7U0R4aXurTPGfOnLENcuzYMQ0ODloxE023cQxxaKPRqAYHB5VKpdTX12cOTjQaVTabNcoIxXWg9FBV0LubkoOfyMZ3+YcERRwYONQYVNYMVf84gLw/kTHOhcublWQ89fp+s6st6OncuXP2DJ5//nkLYAhcYrGFKWdMayNtyj4DCWX9UyRB4QodPXBe29vbzRkulxemiYEKFYtFK8ahCwP6dikYINNkhAh6OJRBHEGO3Wbv2BKXI+zuMyqfabvFuqrvgblWhD1RLpf12c9+tmYMKqPd6dYCt5LgE849wUQ4HLYMDs9yZGSkpkiUQ5eDDuQYXj6fHQ4vTJkbGxtTKBQyZ4ueyugHR5pgSpI509hX1+mGikMxDYEp2RqKanHE1pJUq9UafuiJEycsO0EmkjVNAEr2zKVDuPZqdHTUvqdDEn1hJRkP3p3C5lLPABToykIROPuXFluclS6ijyN1K06PK+VyWefPn7dn4qbU14pgP7lulxv72muvaXx8XKOjo8bzhcpApyMyo9DDCPIJJEZHF+Z5EByyj8ieuxk6164D0NGO7ODBg4pGo9Z+jPP8jTfeqMnqu/clvXOd3qqsvV17EyHdX6lU9Pzzz6unp6dmmolbCc4kuM2bN1v0jxGDw8Q/eE1UJhO1SqoZhRsKhazLQrFYVD6fN2MpqWb6m6SaNBvOFxsQqgKRWSgUMh4TaQUcPBepoL0SjnBvb6+lGdaS4ODhPEqLtJeTJ0/K933t3btXbW1thgS5QQGOJ2g6KDCoI+kuggscU5f+MDU1pdbWVuvb66beSZmT5sXJwoHmOiiqISCqVquWpsVBw7hTkMOkHA4a0kOkXeuNxGpLva7cFm2lUqlmCtzx48ct2ODZozecpFKppEuXLhk9YmxsTPl83pwqCg3HxsbskJ6bm1N7e7s9H97H7YsMB8wtHCTIIIXmpukx3nwvyQIS9A3q5Xme7UPXdhCckhYELYTqspYOWg6+UCikAwcO1HDs6XpDgVhnZ6c6OzsVjUatmw3dUwgYefYEJmRF4Pgy9hTnk6ADh5r2k77vG9IuLe5lwAIcXoIbUqxw66FXVKtVDQ4OGjoPXefy5cvauHGjZQqgQJ0+fbqGHrKWhDS5tOhcfOc737Gf792712pW2EtucSl7gjVOZgoqCmdQNptVtVq1UfQUl7M/yHq6+iKoHB4ersmWsPfL5YVOHGNjY0aJc1HDm7W3Qlx9uQMy1qK4QJEkO6sI5N96662aaXA8K8CE0dHRmrohajDYT9Vq1QaCuPUKb731Vg1NAif48OHDlm12bb87UAUHmeuQFmu63Pu6nbK2PKK3ESIgSXrqqafU399vzgZIwKlTp3TfffeZYi9duqS+vj4zgGxeIiCU7fu+cWKIhFAO6ANpcmlxlF8ksjDRy03ZSDJEIZ/P10whcdOk8IFBjDg04AvjbLAAXUNOxwA3lbOWxDVeLq+JZx4KhXTu3Lma1/r7+61inAb7HMDt7e3G38Ywg1hhYOFBUfXqRsakqqVFJKS+abs7ytFFcll/tFvyrlXLg3IS3MA/dh04UsDT09PG8W40XdYjDgQp7qGCg/nUU09JWjCUd955pyGGnudZxS/IL63g4BkSkNCODCcWCkI4HFZPT49Rh6A8LVVgxiHgtlkjmOTnbsGja+Td8cSggYyvxlZIsnS5S23h/qhqbrRA5lbknnvusSIVOqpwuLInQPlAzmkpBw2LKvDu7m5zXnGCKKJhz5GlcbMBpGqxvW6LLT5fWjxk0cvp06d14MABvfLKK7bn8/m8FT3Ozs7q7NmzNQ5Uf3+/RkZGavjABMHS9S2m1oLU80jJkkiLtvWmPVJrubH2P79/9OhR5fN5vfrqqzp69KgVSNFVgGsgyMS5LpVKeu211+y9H3nkEUPiXUfq4MGD1q7yRvf2duIW5bpFZWtNn+75KC1ycKWFe3z11VdVqVSM+xwKhbR3714NDQ3p9ddfN9vmPo+l7p8gx10T/I2bRbjRYIn6mgKAHWx0fQ3X7dbFunJ+QSZ4YGfOnLHUCEbr6NGjhr68+OKLuu+++9Td3W0TjUBxcUhBiyjOCIfD1goEJIeUnluUxGdgmInCIpGI/YyiKZdvyAHvOtS0iJJkiBGpWhchpLgGg0KFNc7TWhI2U70Bdgnt9Y7DmTNn9MILL+jw4cNmLB966CFDb5kKQ8UwhHz4v7Q0ouOCtNh+B4TJ8xa6OYBkUeXsdhpg/dBtwF0jOFdcuztXnsCI9UBBJI4DtJlGQwvr01OuEUVcZwTU79y5czV/f+jQoZriCXihFy9eNAempaVFg4ODhhpTPAXCzj5hT7htxdAXeiHDAh/YnQbndnKQFvcm2RXoEtgH+pW6fbpdVB/dYp/gPa415xc+Zz6f18DAgLV5Yw9QuDQ/P6+tW7equbnZqGAEGoyQHRkZsWAPTj3oFGijS+ni8wlc4RKGw2HT99TUlPHrQZBBlF544QXT15133qnTp09r//79evXVV219uo4Zcvbs2ZpzpX69N9p+vBVx7wHHl2JEt3AKcVtI1VPS+N9FHE+ePGmI/smTJ+097rnnHsXjcf34xz/WkSNHjAZBYRo2gPccGRlRKBSy7gs4Si6KCOLItd2Kw+Q66vWI41pzfhF0SdAuLazNesS1Wq0uOQijvnaqXuo7MPD+9VkEt4sT7+W2auNsdylrSD2F43bKunJ+2RgsApwkF/k5efKk/d7Ro0fV3NysK1euGKdXWuTX4nxIC6lZUmL5fN4cqsnJSUNZc7mc9cIjFeMuFJC9yclJUz7/KNCiwpWitmq1apPK4MrhoONEkM4jPQi5nT57OMVrSdgYbhqd4OaDH/ygUT/QD2gQ6ejZ2VkdPHjQDtqZmRlNTk4akuoWOnGAcljSdJ/CMwqpKOqApybJWvWA8rnBCK2SQHfpNUtBpSQrfpufn7eUO9OtQMBI8+G8NRp/+0aFUvWvYxSPHTtW8/N8Pm8oDu3CqtWq9V2en5+3YA4907yetkytra2GJLjBA3rhs3Gi3BZB0FBwgikshY7B1wTEBKbYGSby8bkYbrejBcU7PAe36GQtie/7Onv2rF5//XUNDQ1ZIAnPHUe1o6ND0WjUxoRTmEo/6HA4rIGBAc3Ozqq9vd1ex6kl8CBLwyRECtZYE3D863s6I+x/Sbr33nvttddee82cABftcrvKSLouY8AzkGpbS91Kqr2RZKm9Wa1Wa5z8+lT6UlLvqPAs66v2CWpPnTpVcw35fF7JZNICE3fwlO/7ev3112vOUTeIBtV3EcOl9PV28k4pE40mrp7ce6l/FvXf38j5Xer+3b91QQ53nbg9eetBKvc9651onGNs8lLXdDtkXTm/rlNA+svtG1ifvjx58qQ5LJVKRcePHzcH2G2zgsNBj16mtc3Pz1s6FmeKCXAdHR3q6+uzwQxUFHOwcr2hUMgOfBAMnDg4oQjpdq7X7R3L/3Cj3NRrfapirYhLpGczPfHEE7rrrrtqipTchvquUzs/P6+xsTFJqunaADeNZ02anOdUrVY1MjJi6ykWi6m3t9d6FebzeWvmjkBLwUC4KXFoEXwtLbZ2I4Bx+/qyPugy0NLSYt0+cOYaUVzuL/pyI/nPfvazymQyGh0dtUCura1NY2Njuueee2xOPFP04NRKC8+LvpzoijUAXSGZTFqGA8NMOhzE0p3YVq1Wa4pvoGrgqFKQw3u6+xxjHYlEaoac4EC4PNR4PG6oGHx00Oe15Py6+/Gll16qyXYQgBNMjoyM1AyDoKgTShkdcFKplIaGhiyzgq74XTrjkM2CQ+3uX2mxjoNBJwRNIJlwkmdmZizjgCMt1SJO7r52EWHXCeNMQdaSHhGyIq7z6CJ3Ui3iK9V2E5CWdpRc1K/+d1yH6Sc/+UnN7y6FONY/c/c63Z+9U7RwqSzVWsvCuELgv1TA4gY2fO9SI26lsGwpwAW9uS0fOQPq15T7dX0WYSlxsxC3SxrzFH2PArLLQYXhch1fSYbK8TdPPfWUHWAHDx60dGp9yrZcLmtkZMQMOQiA7/tqaWkxp5vhBky4YnqK22oJWgSFHaC2LqLrTnVz27y4iy0WiymbzSqXyxlXleKCdxMNr7bgrLA5MHQ9PT1WDUzxEqnXVCplwURbW5s5Gjiro6OjNjQEfh/PbHZ2tsZRJlDq6Oiwhuy0taJ7BnpxNzXrji4ArEV+j/ZWbosgeKGgwvXt07LZrFpaWow/3Gi6dCN+vncDFt/3dejQIR09etRa6FCwyQjxLVu21NAW4OxSWQ7fWlrYzxSg0ppJqi00owUW7e2kRZ6Z29FFkvU1dYMPl4NGJoeUfUtLixXdYchx1ECZPW+heX5LS4sVf0C94GduJ4O1IKCivu9bX1R07NYW3HXXXcYvDIfDeuaZZ3TgwAHjvpMJOX36tA4ePGjPEXSffY8uQYo5pHGMXeSVjg9kSySZbabArb+/X+fOnbNrqHdsERf9cx0t6XoHbCmUdK0IiLlLV3DpC/WBwFL0JteJqac71TvK9fUn9eh6PWK81Pu4urqZs3wr+nA/y9XhWjwvXce3Xnf1a5l/9cjtzaQ+QHGDFveZL8UHrtfHUjav3lGvDy5vh6xL57c+SnEfouuMuMLDpwL1zJkzCofDxlGSFpo+d3d3KxqN2gjd+fl5a9FDWo9xyiBJ3d3dNnbY5RfBk6tUKurs7Kzpl0hfVBxkjH+xWFQmk6nhEYKAgXZns1njtLpUkLUkbCy3GXa5XLZeyhSouE3mKRJk86TTaaVSKXV3dxtnG3TQHUqRy+WM+gLFBD29+eabdtg2NzdbilySBRbxeFwTExOWbqUyHScIZzcUCml8fNy4oKC6VLqDQIP8MzYbRwBKTaMWvNV/7+47+haDxoEQlssLYzNpB+Z5njo7OxWPx9XR0WH8966uLqMcSbIuLYyFJqjBUR0eHq5pbVetLrQUZGANaHE2m7VCUtZR/T3QDouOLiCQ0J9AgOkn29bWZtSVmZkZ6wwDd5tCrUbs3PF2Uq9rt8CX/2lthIMsSS+++OKS74dj6orrlOzfv9+4oG7Qz+e673vs2DG9+OKLOnDggCTZdfCefFb9oVr/vXuPN7Kbb4dcrRWpd5DqX3el/l5v9Dc3Slvf6LPcwOJGf+uCIDf7+VLXeSO5UeC51hxfaennfKPfuRm94e0+41Ypbu77v5O9cit7b7lkXTq/NxPX8T148KB2795txhZuGL1ESZFLMmSWSTNMteHABgWEz8eIRXoIgxxS6AKym0wmDc3jsGbwBU4OaFKlUtHGjRsNiQQZhodKcY3LfaRhf6Omym8k9egM6FIotDBYBOQPvXDPV65cUXt7u9EDcCTd1jr0MKRKP5lMqq+vzxzlYrFoaVimSRFEUKCGY070TCBEb9hsNmvT5aTFQkXS+ZJMf1Afuru7lc1m7fqTyaSGh4eVSCSsvRKO1FoTkNyxsbEavUL5qFYXui1s2bLFgg2CN99f6DUJWiHJWgpKiy0D6aYxMTFRg9oSDMIvJFik00qhUND4+LgFlZOTk5ZdgENK5ojrod0T6DzXwT2Fw2HjLOdyOZseSUBMVmotHrS3U1w0MRKJWKHT6dOndejQIU1MTFjVv0uviUajVljlOsDYhbWEsAcSSCC3X9aWR7RMgiPz6KOPatu2bRofHzeagssvq++/St9Wqv/z+XxNNTidAcrlhX6IqVTKmtnjzNIiiEIPSTXcNnekJg4yLXlisZiGhobskOZgxxmXVNMmi+4CLlKyVsQNUuCmUQQVDoetcfrc3Jza2toUjUZr0HIcT2gm8KtzuZymp6eNlgDy7hbZ4OziHIGqExxJUiaTsXXk/g5OMA3/cbIpnqxWq9buDFoF7831u32DcZZwet1inrUkw8PDNuXw6tWrCofDNRzf5uZmGyAxNDRk1JJyuVyDqFPo5nLfc7mcIb4EIuXyQg9LaEVQIWZnZ+390FmlUjGUFmoKCDX7mZ9TlIdtAMVmzZCGB0HG8SKrQ/A8Pz+vycnJwCmrEzdtXk83O3XqlNVo1BfS1HMD4fUu1bkgkEACCeR95fzirOBY9fb2WsU4zgnOD+2sKDJiWpokq+xmYAUI4czMjHFI8/m85ufnlc/naziAtOmRFpv649xKi1XoOLcUgfA3oGKuQwai6Y6JlFTDm6PF1loRl0xPirhcLuu73/2u+vv7NT8/rzfffFP79u2zQkKm5hWLRXV0dOjChQuKRqP2PZSB8fFxc2Ky2awk1ThMFM2AyM7MzNiAFNBeUHp0DK2hXC4rn8/XFD+B+kciEeXzeb388ss6ePCgOfMvvviiHnjgAUUiEesFDVLodr1g1Opac5g8b6H3YyaT0fj4uLWBk6RsNquuri4NDw+rp6dHV69eVUtLiwYGBqwFksstO3/+vI3TLZVKyufzxv8kCHG7pJD5oMMHs+lBf90WPJlMxpB2ghM6AnAfvu/ryJEjtuZ837dBKwQ3w8PDhkY2NTWpra3NAuJkMmnjd2lVGMiCLMW1dVvVuTxs6fpetTjDbptE11EOnnUggQSCvK+cXxdNBCGcmJjQ5OSk0Qeq1ary+bz6+voUiUSUTCa1ceNGNTU1WfW953l20IHeSjI0B5RrYmLCeIpLVQmDLuHgklKlkr1QKBj6K8kQLHfIAo57Pp+3Qhw4wtIiP+52Vk3eLuFZ4VQiLk8QJ4ritWp1YZrQhQsXLC0aDod19epVG/U8PT1tE/hAA1OplMbGxqyQjeDCdXxxZHF2aD1GIRw9fjOZjLWyyuVyeu211+x69+/fr4MHD1oFOg3BQaNZY6D2ZBcY7cnQgLUoP/jBDyQtrMlDhw5pYGBATU1NGh8fV6VS0dWrV1Uul63fMuhqOp02Z7VYLGpsbEzVatUcaQILaSHwmZ+fVzweV2trq+0p9lM6ndalS5esyIq/Yx1kMhmVy+WawTNuEYYknTp1yhwtz/N0zz33qKWlRW+99ZZeeOEF9ff368yZMzp8+LCNNoeCxLVA9Wi0tnWrKfV8P7IekiyYqG8t5qK//I2L9q7lgrRAAgnk9sn7yvmVFo0hqOrg4KDeeusthUIhZTIZa39Duryvr88KXGh7RhodXiKTjKjS7+zsNM4mjtPs7KyNEcRhpV9suVy21kek3ikOoh0TCNWPf/xjHT161GgSHKSJREL5fN6QQ9KwoFzM4V6LwuHmTpGRFnT5zDPP1ByAPKfjx48bRzsWiymVSikej5sTTPuqTCZjOqSYCk42aD0OEYggg0rgfUuyYAiUnSlTqVRKDz74oF2nW3iDHD161FLjkmzwwujoqCH28EQpsFpLQmDgprFpc8R+c+/p0KFDxp9NJpPG6S4Wi1bcBn+e/cIkPknWCi2TydienZ+fVyaTsWEKdFYhuCoWizp58qT27t1r10Ogcc8991h/0v379+vs2bM1RRynT5+W53nat2+fpMXg7NSpU+rv77cMBsHZ9PS0zp07t+b0uBLior9ucbDb1sxdSy5Not4Rru/MEEgggQSCvK+cX1LR0oJxfPnll22EKv0mC4WCuru7Df3BoWWyG9QJKrsjkYi1omppaTHe78TEhKVq3bn3w8PD5pTy96B9pPihXtAwHlpFtVrVfffdJ2kBxWZ6m9sE3m3xxv1Cu1hL4vbddCfCuAdZ/f8uGvTjH/+45v2OHz9+XfsUAhFed1vvgCi7HTpAheF/u2gg10pxHb9PUd7dd9+tF1980Zy6ffv2KZlM6uTJkzp58qT6+/sVCoU0MjKizs5OK5qiI4jv++awrbWD3G0pCCe9frqP69C8+OKLxpPft2+fzp07pz179ujv/u7vND8/r/379xsKTGFjU1OTJiYmbA3Qz5f3JBjBWd23b58FNHDmq9Wqzp07Zw7wmTNndODAAZ05c8au7+zZs5aK595AHJmaJC2ikC+++KKq1aqOHj1qfcVv1GLr/S71Tmp92yp3sMGNHF6kvktMIIEEEogr3koepJ7nNcSp7faSfPDBBw2dbWtrM8SINGhXV5d27dplU8Ko1ofj29raqoGBAbW2tiqZTFqXBwqW6Cmay+Ws8TuN91OplBKJhFEnJiYmLM0+MzNjbbmgPTz99NM6fvy4yuWyTp8+rbm5OfX39+ull17SXXfdVeNEh0Iha81F+v+b3/zmKd/3jyzHM7zduqzvoykt3ePR7bHrOsr1h6bbu9IVXsMZhW7A8yfggQcMt5oASJJ125BkhYnT09PyPE8nTpyQ53k6ePCgpcLpA037LBwkt2F4f3+/Hd7ValUvvviiHnnkEZtONTExsWzRzErsSzeYcV+Tlm5pg/5uNJUIxxg+7sGDB/XSSy/VUBGkxRHRFDu62ZZwOKyTJ08amtvf36/Z2VmjD6FrCt1ef/31Gk5qfceBG00Gq19/9c/C9/1l0WWj2NflkKX2b73U95WVbr6WVkDWjH0N5G0l0OX6kSV1udLOb0HSa2/7i40vnZLGVvsi3oVs9X2/azneKNDlqsqy6VEKdLnKspx7clTSlNbeM1hKAl2uD12uRT1KwVm5lKwrXa407eG15YqmVlM8z3t+PdzHe5RAl+tHAl2uA/F9v2u9PIP1ch/vVtaLLtfDPSyDBPa1AaWxRkUFEkgggQQSSCCBBBLIbZTA+Q0kkEACCSSQQAIJ5H0jK+38/pcV/rzbJevlPt6LrJdnsF7u473IenkG6+U+3ousl2ewXu7jvch6eAbr4R7eq6yXZ7Be7kPSChe8BRJIIIEEEkgggQQSyGpKQHsIJJBAAgkkkEACCeR9Iyvm/Hqe9xHP817zPO8Nz/P+5Up97nKI53kXPM97yfO8M57nPX/ttXbP877ned7fXfu/bbWvc6Uk0OX6kbWqy0CPtbJW9SgFuqyXQJfrRwJdNq6siPPreV5Y0v9X0kcl7ZX0hOd5e1fis5dRHvV9v99p9fEvJX3f9/1dkr5/7ft1L4Eu14+sA10GetS60KMU6FJSoMv1JIEuG1tWCvk9KukN3/ff8n1/TtJ/k/SpFfrs2yWfkvTH177+Y0mfXr1LWVEJdLl+ZL3pMtDj+tCjFOgy0OXal0CXDSwr5fxulHTZ+f7KtdfWiviSvut53inP8z5/7bUe3/cHr309JKlndS5txSXQ5fqRtazLQI+Lspb1KAW6dCXQ5fqRQJcNLCs94W2tynHf9696ntct6Xue573q/tD3fT+Y371mJNDl+pBAj+tHAl2uHwl0uX5kXetypZDfq5I2O99vuvbamhDf969e+39E0te0kM4Y9jyvT5Ku/T+yele4ohLocv3ImtVloMcaWbN6lAJd1kmgy/UjgS4bWFbK+X1O0i7P87Z7nheT9IuSvrFCn/2exPO8pOd5ab6W9Liks1q4/s9e+7XPSvpfq3OFKy6BLtePrEldBnq8TtakHqVAl0tIoMv1I4EuG1hWhPbg+37Z87x/JOk7ksKS/l/f919eic9eBumR9DXP86SF5/Vnvu9/2/O85yT9ued5vyHpoqTPrOI1rpgEulw/soZ1GejRkTWsRynQZY0Eulw/EuiysSWY8BZIIIEEEkgggQQSyPtGgglvgQQSSCCBBBJIIIG8byRwfgMJJJBAAgkkkEACed9I4PwGEkgggQQSSCCBBPK+kcD5DSSQQAIJJJBAAgnkfSOB8xtIIIEEEkgggQQSyPtGAuc3kEACCSSQQAIJJJD3jQTObyCBBBJIIIEEEkgg7xsJnN9AAgkkkEACCSSQQN438v8HV81fbttVvKYAAAAASUVORK5CYII=", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Class: rock, path: C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/rock\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Class: scissor, path: C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/scissor\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Class: _unknown_, path: C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/_unknown_\n" ] }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "%matplotlib inline\n", "import matplotlib.pyplot as plt\n", "import matplotlib.image as mpimg\n", "\n", "# Collect 5 samples file paths for each class in the dataset\n", "class_samples = {}\n", "for class_name in os.listdir(dataset_dir):\n", " class_dir = f'{dataset_dir}/{class_name}'\n", " if not os.path.isdir(class_dir):\n", " continue\n", " if class_name not in class_samples:\n", " class_samples[class_name] = []\n", "\n", " for sample_filename in os.listdir(class_dir):\n", " if len(class_samples[class_name]) > 5:\n", " # We only want 5 samples from each class\n", " break\n", " if not sample_filename.endswith('.jpg'):\n", " continue\n", "\n", " sample_path = f'{class_dir}/{sample_filename}'\n", " class_samples[class_name].append(sample_path)\n", "\n", "# Display the class samples\n", "for class_name, sample_paths in class_samples.items():\n", " class_dir = f'{dataset_dir}/{class_name}'\n", " print(f'Class: {class_name}, path: {class_dir}')\n", " _, axs = plt.subplots(1, 6, figsize=(12, 3))\n", " axs = axs.flatten()\n", " for sample_path, ax in zip(sample_paths, axs):\n", " img = mpimg.imread(sample_path)\n", " ax.imshow(img, cmap=\"gray\")\n", " plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Update the Dataset\n", "\n", "Currently, the dataset contains less than 5k samples. This is quite small and will likely not produce a robust model.\n", "The best way to make a robust model is it add more __representative__ samples to the dataset.\n", "\n", "For this dataset, \"representative\" means:\n", "- Different people's hands making each gesture\n", "- Different lighting angles\n", "- Different backgrounds\n", "- Different distances from the camera\n", "- Use of \"left\" and \"right\" hand\n", "- Showing front and back of hand\n", "\n", "So basically, to improve the model we need to increase the size of the dataset by having different people record their hands performing \"rock\", \"paper\", \"scissors\" from different orientations.\n", "The more images we add, the more \"representative\" the dataset becomes, which should (hopefully) make the model more robust.\n", "\n", "Fortunately, the MLTK features a command that allows for recording images from the embedded device." ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Usage: mltk classify_image [OPTIONS] \n", "\n", " Classify images detected by a camera connected to an embedded device.\n", "\n", " NOTE: A supported embedded device must be locally connected to use this\n", " command.\n", "\n", "Arguments:\n", " On of the following:\n", " - MLTK model name \n", " - Path to .tflite file\n", " - Path to model archive file (.mltk.zip)\n", " NOTE: The model must have been previously trained for image classification [required]\n", "\n", "Options:\n", " -a, --accelerator Name of accelerator to use while executing the audio classification ML model\n", " --port Serial COM port of a locally connected embedded device.\n", " 'If omitted, then attempt to automatically determine the serial COM port\n", " -v, --verbose Enable verbose console logs\n", " -w, --window_duration \n", " Controls the smoothing. Drop all inference results that are older than minus window_duration.\n", " Longer durations (in milliseconds) will give a higher confidence that the results are correct, but may miss some images\n", " -c, --count The *minimum* number of inference results to\n", " average when calculating the detection value\n", " -t, --threshold Minimum averaged model output threshold for\n", " a class to be considered detected, 0-255.\n", " Higher values increase precision at the cost\n", " of recall\n", " -s, --suppression Number of samples that should be different\n", " than the last detected sample before\n", " detecting again\n", " -l, --latency This the amount of time in milliseconds\n", " between processing loops\n", " -i, --sensitivity FLOAT Sensitivity of the activity indicator LED.\n", " Much less than 1.0 has higher sensitivity\n", " -x, --dump-images Dump the raw images from the device camera to a directory on the local PC. \n", " NOTE: Use the --no-inference option to ONLY dump images and NOT run inference on the device\n", " Use the --dump-threshold option to control how unique the images must be to dump\n", " --dump-threshold FLOAT This controls how unique the camera images must be before they're dumped.\n", " This is useful when generating a dataset.\n", " If this value is set to 0 then every image from the camera is dumped.\n", " if this value is closer to 1. then the images from the camera should be sufficiently unique from\n", " prior images that have been dumped. [default: 0.1]\n", " --no-inference By default inference is executed on the\n", " device. Use --no-inference to disable\n", " inference on the device which can improve\n", " image dumping throughput\n", " -g, --generate-dataset Update the model's dataset.\n", " This will iterate through each data class used by the model and instruct the user\n", " the display the class in front of the camera. An image is captured from the device's camera\n", " and saved to the model's corresponding dataset sub-directory.\n", " This process will repeat until the user exits the command. \n", " Use the --sample-count option to specify the number of samples per class to collect\n", " NOTE: Device inference is disabled when using this option \n", " See the --dump-images option as an alternative to generating a dataset \n", " --sample-count INTEGER The number of samples to collect per class\n", " before iterating to the next class\n", " [default: 5]\n", " --app By default, the image_classifier app is automatically downloaded. \n", " This option allows for overriding with a custom built app.\n", " Alternatively, set this option to \"none\" to NOT program the image_classifier app to the device.\n", " In this case, ONLY the .tflite will be programmed and the existing image_classifier app will be re-used.\n", " --test Run as a unit test\n", " --help Show this message and exit.\n" ] } ], "source": [ "!mltk classify_image --help" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Using the command:\n", "\n", "```\n", "mltk rock_paper_scissors --dump-images --dump-threshold 0.01\n", "```\n", "\n", "Images from the embedded device will be saved to the local PC.\n", "The images can then by copied into the \"Rock, Paper, Scissors\" dataset directory." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Update Sequence\n", "\n", "The process for updating (i.e. adding more samples to) the \"Rock, Paper, Scissors\" dataset is as follows:\n", "\n", "__NOTE:__ A similar process can be used for other image based datasets\n", "\n", "1. Purchase an [ArduCAM](https://www.arducam.com/product/arducam-2mp-spi-camera-b0067-arduino) camera module\n", "2. Connect the ArduCAM to a supported development board as described in the [image_classifier](../../docs/cpp_development/examples/image_classifier.md) example application\n", "3. Issue the command: `mltk rock_paper_scissors --dump-images --dump-threshold 0.01`\n", "4. Open the dump directory that is printed in the terminal \n", " (which should be something like `~/.mltk/image_classifier_images/brd2601`), \n", " you should see images being dumped to this directory. \n", " __Ensure the background is a solid color__ (A _much_ larger dataset is required to use random backgrounds)\n", "5. Make the \"rock\" gesture in front of the camera, and move your hand is various orientations and distances from the camera\n", "6. Repeat step 4 showing the other side of your hand making the \"rock\" gesture \n", " (if possible, also change the lighting conditions to collect even more samples)\n", "7. Once enough images have been dumped (~100-200), review the images in the dump directory. \n", " Delete all images that do not clearly show your hand making the \"rock\" gesture\n", "8. Once the dump directory only contains images of your hand making the \"rock\" gesture, copy all of the images to the dataset directory: `~/.mltk/datasets/rock_paper_scissors/v2/rock`\n", "9. Repeat steps 5-8 using the \"paper\" gesture and then the \"scissors\" gesture\n", "\n", "Once this process is complete, the model should be retrained.e.g.: `mltk train rock_paper_scissors --clean` which will use the updated dataset." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Creating the Model Specification\n", "\n", "The model specification is a standard Python script containing everything needed to build, train, and evaluate a machine learning model in the MLTK.\n", "\n", "Refer to the [Model Specification Guide](../../docs/guides/model_specification.md) for more details about this file.\n", "\n", "The completed model specification used for this tutorial may be found on Github: [rock_paper_scissors.py](https://github.com/siliconlabs/mltk/blob/master/mltk/models/siliconlabs/rock_paper_scissors.py). \n", "\n", "The following sub-sections describe how to create this model specification from scratch." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Create the specification script\n", "\n", "From your favorite text editor, create a model specification Python script file, e.g: \n", "`my_rock_paper_scissors.py`\n", "\n", "The name of this file is the name given to the model. So all subsequent `mltk` commands will use the model name `my_rock_paper_scissors`, e.g:\n", "\n", "```shell\n", "mltk train my_rock_paper_scissors\n", "```\n", "You may use any name as long as it contains alphanumeric or underscore characters.\n", "\n", "When executing a command, the MLTK searches for the model specification script by model name. \n", "The MLTK commands search the current working directory then any configured paths. \n", "Refer to the [Model Search Path Guide](../../docs/guides/model_search_path.md) for more details.\n", "\n", "__NOTE:__ The commands below use the pre-defined model name: `rock_paper_scissors`, however, you should replace that with your model's name, e.g.: `my_rock_paper_scissors`." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Add necessary imports\n", "\n", "Next, open the newly created Python script: `my_rock_paper_scissors.py` \n", "in your favorite text editor and add the following to the top of the model specification script:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# Bring in the required Keras classes\n", "from tensorflow.keras.models import Sequential\n", "from tensorflow.keras.layers import Dense, Activation, Flatten, Dropout, BatchNormalization\n", "from tensorflow.keras.layers import Conv2D, MaxPooling2D\n", "\n", "from mltk.core.model import (\n", " MltkModel,\n", " TrainMixin,\n", " ImageDatasetMixin,\n", " EvaluateClassifierMixin\n", ")\n", "\n", "# By default, we use the ParallelImageDataGenerator\n", "# We could use the Keras ImageDataGenerator but it is slower\n", "from mltk.core.preprocess.image.parallel_generator import ParallelImageDataGenerator\n", "#from keras.preprocessing.image import ImageDataGenerator\n", "# Import the dataset\n", "from mltk.datasets.image import rock_paper_scissors_v2" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "These import various Tensorflow and MLTK packages we'll use throughout the script. \n", "Refer to the comments above each import for more details." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Define Model Object\n", "\n", "Next, add the following to the model specification script:" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# Instantiate the MltkModel object with the following 'mixins':\n", "# - TrainMixin - Provides classifier model training operations and settings\n", "# - ImageDatasetMixin - Provides image data generation operations and settings\n", "# - EvaluateClassifierMixin - Provides classifier evaluation operations and settings\n", "# @mltk_model # NOTE: This tag is required for this model be discoverable\n", "class MyModel(\n", " MltkModel, \n", " TrainMixin, \n", " ImageDatasetMixin, \n", " EvaluateClassifierMixin\n", "):\n", " pass\n", "my_model = MyModel()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "This defines and instantiates a custom MltkModel object with several model \"mixins\". \n", "\n", "The custom model object must inherit the [MltkModel](../../docs/python_api/mltk_model/index.md) object. \n", "Additionally, it inherits:\n", "- [TrainMixin](../../docs/python_api/mltk_model/train_mixin.md) so that we can train the model\n", "- [ImageDatasetMixin](../../docs/python_api/mltk_model/image_dataset_mixin.md) so that we can train the model with the [ParallelImageDataGenerator](../../docs/python_api/data_preprocessing/image_data_generator.md)\n", "- [EvaluateClassifierMixin](../../docs/python_api/mltk_model/evaluate_classifier_mixin.md) so that we can evaluate the trained model\n", "\n", "The rest of the model specification script configures the various properties of our custom model object." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Configure the general model settings" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# For better tracking, the version should be incremented any time a non-trivial change is made\n", "# NOTE: The version is optional and not used directly used by the MLTK\n", "my_model.version = 1\n", "# Provide a brief description about what this model models\n", "# This description goes in the \"description\" field of the .tflite model file\n", "my_model.description = 'Image classifier example for detecting Rock/Paper/Scissors hand gestures in images'" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Configure the basic training settings\n", "\n", "Refer to the [TrainMixin](../../docs/python_api/mltk_model/train_mixin.md) for more details about each property." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# This specifies the number of times we run the training\n", "# samples through the model to update the model weights.\n", "# Typically, a larger value leads to better accuracy at the expense of training time.\n", "# Set to -1 to use the early_stopping callback and let the scripts\n", "# determine how many epochs to train for (see below).\n", "# Otherwise set this to a specific value (typically 40-200)\n", "my_model.epochs = 125\n", "# Specify how many samples to pass through the model\n", "# before updating the training gradients.\n", "# Typical values are 10-64\n", "# NOTE: Larger values require more memory and may not fit on your GPU\n", "my_model.batch_size = 32\n", "# This specifies the algorithm used to update the model gradients\n", "# during training. Adam is very common\n", "# See https://www.tensorflow.org/api_docs/python/tf/keras/optimizers\n", "my_model.optimizer = 'adam'\n", "# List of metrics to be evaluated by the model during training and testing\n", "my_model.metrics = ['accuracy']\n", "# The \"loss\" function used to update the weights\n", "# This is a classification problem with more than two labels so we use categorical_crossentropy\n", "# See https://www.tensorflow.org/api_docs/python/tf/keras/losses\n", "my_model.loss = 'categorical_crossentropy'" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Configure the training callbacks\n", "\n", "Refer to the [TrainMixin](../../docs/python_api/mltk_model/train_mixin.md) for more details about each property." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# Generate checkpoints every time the validation accuracy improves\n", "# See https://www.tensorflow.org/api_docs/python/tf/keras/callbacks/ModelCheckpoint\n", "my_model.checkpoint['monitor'] = 'val_accuracy'\n", "\n", "# https://keras.io/api/callbacks/reduce_lr_on_plateau/\n", "# If the test loss doesn't improve after 'patience' epochs \n", "# then decrease the learning rate by 'factor'\n", "my_model.reduce_lr_on_plateau = dict(\n", " monitor='loss',\n", " factor = 0.95,\n", " min_delta=0.001,\n", " patience = 1\n", ")\n", "\n", "# If the accuracy doesn't improve after 35 epochs then stop training\n", "# https://keras.io/api/callbacks/early_stopping/\n", "my_model.early_stopping = dict( \n", " monitor = 'accuracy',\n", " patience = 25,\n", " verbose=1\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Configure the TF-Lite Converter settings\n", "\n", "The [Tensorflow-Lite Converter](https://www.tensorflow.org/lite/convert) is used to \"quantize\" the model. \n", "The quantized model is what is eventually programmed to the embedded device.\n", "\n", "Refer to the [Model Quantization Guide](../../docs/guides/model_quantization.md) for more details." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "my_model.tflite_converter['optimizations'] = ['DEFAULT']\n", "# Tell the TfliteConverter to generated int8 weights/filters\n", "my_model.tflite_converter['supported_ops'] = ['TFLITE_BUILTINS_INT8']\n", "# We want the input/output model data types to be float32\n", "# since we're using samplewise_std_normalization=True during training\n", "# With this, the TfliteConverter will automatically add quantize/dequantize\n", "# layers to the model to automatically convert the float32 data to int8\n", "my_model.tflite_converter['inference_input_type'] = 'float32'\n", "my_model.tflite_converter['inference_output_type'] = 'float32'\n", "# Generate a representative dataset from the validation data\n", "my_model.tflite_converter['representative_dataset'] = 'generate'" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Configure the dataset settings\n", "\n", "Next, we specify the dataset. In this tutorial we use the [Rock Paper Scissors v2](https://siliconlabs.github.io/mltk/docs/python_api/datasets/index.html#rock-paper-scissors-v2) dataset which comes as an MLTK package.\n", "\n", "__NOTE:__ While the MLTK comes with pre-defined datasets, any external dataset may also be specified. \n", "Refer to the [ImageDatasetMixin.dataset](../../docs/python_api/mltk_model/image_dataset_mixin.md) property for more details.\n", "\n", "__NOTE:__ While a dataset path can be hard coded, it is _strongly_ recommended that the script dynamically downloads the dataset from the internet. This allows for the model training and evaluating to be reproducible. It also enables remote training on cloud services like [Google Colab](https://colab.research.google.com/notebooks/welcome.ipynb) which need to download the dataset any time a virtual instance is created." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# The directory of the training data\n", "# NOTE: This can also be a directory path or a callback function\n", "my_model.dataset = rock_paper_scissors_v2\n", "# The classification type\n", "my_model.class_mode = 'categorical'\n", "# The class labels found in your training dataset directory\n", "my_model.classes = rock_paper_scissors_v2.CLASSES\n", "# The input shape to the model. The dataset samples will be resized if necessary\n", "my_model.input_shape = (84,84,1)\n", "# Shuffle the dataset directory once\n", "my_model.shuffle_dataset_enabled = True\n", "# The numbers of samples for each class is different\n", "# Then ensures each class contributes equally to training the model\n", "my_model.class_weights = 'balanced'" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Configure the data augmentation settings\n", "\n", "Next, we configure how we want to augment the dataset during training. \n", "See the [ParallelImageDataGenerator](../../docs/python_api/data_preprocessing/image_data_generator.md) API doc for more details.\n", "\n", "With these settings, random augmentations are done to the training subset samples during training.\n", "This effectively increases the size of the dataset." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "my_model.datagen = ParallelImageDataGenerator(\n", " cores=0.65,\n", " debug=False,\n", " max_batches_pending=32, \n", " validation_split= 0.15,\n", " validation_augmentation_enabled=False,\n", " rotation_range=15,\n", " width_shift_range=5,\n", " height_shift_range=5,\n", " brightness_range=(0.80, 1.10),\n", " contrast_range=(0.80, 1.10),\n", " noise=['gauss', 'poisson', 's&p'],\n", " zoom_range=(0.95, 1.05),\n", " rescale=None,\n", " horizontal_flip=True,\n", " vertical_flip=True,\n", " samplewise_center=True, # These settings require the model input to be float32\n", " # NOTE: With these settings, the embedded device must also convert the images at runtime\n", " samplewise_std_normalization=True,\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Data preprocessing\n", "\n", "The [ParallelImageDataGenerator](../../docs/python_api/data_preprocessing/image_data_generator.md) also features some data preprocessing settings:\n", "\n", "```\n", "samplewise_center=True\n", "samplewise_std_normalization=True\n", "```\n", "\n", "This normalizes the input images using:\n", "\n", "```\n", "norm_img = (img - mean(img)) / std(img)\n", "```\n", "\n", "This helps to ensure the model is not as dependent on camera and lighting variations.\n", "\n", "Alternatively, you could use:\n", "\n", "```\n", "rescale=1/255.\n", "```\n", "To scale each pixel between 0-1. This helps the model converge faster during training.\n", "\n", "Either way, any preprocessing that is done during training must also be done at runtime on the embedded device.\n", "\n", "The [image_classifier](../../docs/cpp_development/examples/image_classifier.md) example application demonstrates how to do these image preprocessing algorithms." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Define the model layout\n", "\n", "This defines the actual structure of the model that runs on the embedded device using the [Keras API](https://keras.io/about).\n", "The details of how to create the model structure are out-of-scope for this tutorial.\n", "\n", "The model used by this tutorial was taken from: [Building powerful image classification models using very little data](https://blog.keras.io/building-powerful-image-classification-models-using-very-little-data.html)" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# Build the ML Model\n", "# This model was adapted from:\n", "# https://blog.keras.io/building-powerful-image-classification-models-using-very-little-data.html\n", "#\n", "# This defines the actual model layout using the Keras API.\n", "# This particular model is a relatively standard\n", "# sequential Convolution Neural Network (CNN).\n", "#\n", "# It is important to the note the usage of the \n", "# \"model\" argument.\n", "# Rather than hardcode values, the model is\n", "# used to build the model, e.g.:\n", "# Dense(model.n_classes)\n", "#\n", "# This way, the various model properties above can be modified\n", "# without having to re-write this section.\n", "def my_model_builder(model: MyModel):\n", " keras_model = Sequential()\n", "\n", " # Increasing this value can increase model accuracy \n", " # at the expense of more RAM and execution latency\n", " filter_count = 16 \n", "\n", " # \"Feature Learning\" layers \n", " keras_model.add(Conv2D(filter_count, (3, 3), input_shape=model.input_shape))\n", " keras_model.add(Activation('relu'))\n", " keras_model.add(MaxPooling2D(pool_size=(2, 2)))\n", "\n", " keras_model.add(Conv2D(filter_count, (3, 3)))\n", " keras_model.add(Activation('relu'))\n", " keras_model.add(MaxPooling2D(pool_size=(2, 2)))\n", "\n", " keras_model.add(Conv2D(filter_count*2, (3, 3)))\n", " keras_model.add(Activation('relu'))\n", " keras_model.add(MaxPooling2D(pool_size=(2, 2)))\n", "\n", " # \"Classification\" layers\n", " keras_model.add(Flatten()) # this converts our 3D feature maps to 1D feature vectors\n", " keras_model.add(Dense(filter_count*2)) # This should be the same size at the previous Conv2D layer count\n", " keras_model.add(Activation('relu'))\n", " keras_model.add(Dropout(0.5))\n", " keras_model.add(Dense(model.n_classes, activation='softmax'))\n", "\n", " keras_model.compile(\n", " loss=model.loss, \n", " optimizer=model.optimizer, \n", " metrics=model.metrics\n", " )\n", "\n", " return keras_model" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "At this point, the model specification script should have everything needed to train, evaluate, and generate model file that can run on an embedded device. \n", "The following sections describe how to use the MLTK to perform these tasks." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Model Parameters\n", "\n", "It is extremely important that whatever transforms are done to the dataset during training are also done at run-time on the embedded device.\n", "\n", "To help with this, the MLTK allows for embedding parameters into the generated `.tflite` model file.\n", "\n", "Refer to the [Model Parameters Guide](../../docs/guides/model_parameters.md) for more details about how this works.\n", "\n", "This is useful for this tutorial as the MLTK will automatically embed [ImageDatasetMixin](../../docs/guides/model_parameters.md#imagedatasetmixin) parameters into the generated `.tflite` model file.\n", "Later, the Gecko SDK will read the settings from the `.tflite` model file when generating the project. \n", "\n", "__NOTE:__ The `mltk summarize --tflite` command prints all the parameters that are embedded into the `.tflite` model file." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Model Summary\n", "\n", "With the model specification complete, it is sometimes useful to generate a summary of the model before we spend the time to train it. \n", "This can be done using the `summarize` command.\n", "\n", "If you're using a local terminal, navigate to the same directory are your model specification script, e.g. `my_rock_paper_scissors.py` and modify the commands to use `my_rock_paper_scissors` or whatever you called your model.\n", "\n", "__NOTE:__ Since we have not trained our model yet, we must add the `--build` option to the command. \n", "Once the model is trained, this option is not required." ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/scissor/2022-04-29T23-01-25.981.jpg not found in existing index, re-generating index\n", "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/rock/2022-04-29T23-13-28.550.jpg not found in existing index, re-generating index\n", "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/_unknown_/2022-05-02T17-55-00.359.jpg not found in existing index, re-generating index\n", "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/paper/2022-04-29T23-06-01.204.jpg not found in existing index, re-generating index\n", "Model: \"sequential\"\n", "_________________________________________________________________\n", " Layer (type) Output Shape Param # \n", "=================================================================\n", " conv2d (Conv2D) (None, 82, 82, 16) 160 \n", " \n", " activation (Activation) (None, 82, 82, 16) 0 \n", " \n", " max_pooling2d (MaxPooling2D (None, 41, 41, 16) 0 \n", " ) \n", " \n", " conv2d_1 (Conv2D) (None, 39, 39, 16) 2320 \n", " \n", " activation_1 (Activation) (None, 39, 39, 16) 0 \n", " \n", " max_pooling2d_1 (MaxPooling (None, 19, 19, 16) 0 \n", " 2D) \n", " \n", " conv2d_2 (Conv2D) (None, 17, 17, 32) 4640 \n", " \n", " activation_2 (Activation) (None, 17, 17, 32) 0 \n", " \n", " max_pooling2d_2 (MaxPooling (None, 8, 8, 32) 0 \n", " 2D) \n", " \n", " flatten (Flatten) (None, 2048) 0 \n", " \n", " dense (Dense) (None, 32) 65568 \n", " \n", " activation_3 (Activation) (None, 32) 0 \n", " \n", " dropout (Dropout) (None, 32) 0 \n", " \n", " dense_1 (Dense) (None, 4) 132 \n", " \n", "=================================================================\n", "Total params: 72,820\n", "Trainable params: 72,820\n", "Non-trainable params: 0\n", "_________________________________________________________________\n", "\n", "Total MACs: 5.870 M\n", "Total OPs: 12.303 M\n", "Name: rock_paper_scissors\n", "Version: 1\n", "Description: Image classifier example for detecting Rock/Paper/Scissors hand gestures in images\n", "Classes: rock, paper, scissor, _unknown_\n", "hash: \n", "date: \n", "runtime_memory_size: 0\n", "detection_threshold: 175\n", "average_window_duration_ms: 500\n", "minimum_count: 2\n", "suppression_count: 1\n" ] } ], "source": [ "# Summarize the Keras Model\n", "# This is the non-quantized model used for training\n", "# NOTE: Running this the first time may take awhile since the dataset needs to be downloaded\n", "!mltk summarize rock_paper_scissors --build " ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/scissor/2022-04-29T23-01-25.981.jpg not found in existing index, re-generating index\n", "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/rock/2022-04-29T23-13-28.550.jpg not found in existing index, re-generating index\n", "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/paper/2022-04-29T23-05-47.387.jpg not found in existing index, re-generating index\n", "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/_unknown_/2022-04-29T22-04-13.350.jpg not found in existing index, re-generating index\n", "C:\\Users\\reed\\workspace\\silabs\\mltk\\.venv\\lib\\site-packages\\tensorflow\\lite\\python\\convert.py:746: UserWarning: Statistics for quantized inputs were expected, but not specified; continuing anyway.\n", " warnings.warn(\"Statistics for quantized inputs were expected, but not \"\n", "+-------+-----------------+-------------------+-----------------+-----------------------------------------------------+\n", "| Index | OpCode | Input(s) | Output(s) | Config |\n", "+-------+-----------------+-------------------+-----------------+-----------------------------------------------------+\n", "| 0 | quantize | 84x84x1 (float32) | 84x84x1 (int8) | BuiltinOptionsType=0 |\n", "| 1 | conv_2d | 84x84x1 (int8) | 82x82x16 (int8) | Padding:valid stride:1x1 activation:relu |\n", "| | | 3x3x1 (int8) | | |\n", "| | | 16 (int32) | | |\n", "| 2 | max_pool_2d | 82x82x16 (int8) | 41x41x16 (int8) | Padding:valid stride:2x2 filter:2x2 activation:none |\n", "| 3 | conv_2d | 41x41x16 (int8) | 39x39x16 (int8) | Padding:valid stride:1x1 activation:relu |\n", "| | | 3x3x16 (int8) | | |\n", "| | | 16 (int32) | | |\n", "| 4 | max_pool_2d | 39x39x16 (int8) | 19x19x16 (int8) | Padding:valid stride:2x2 filter:2x2 activation:none |\n", "| 5 | conv_2d | 19x19x16 (int8) | 17x17x32 (int8) | Padding:valid stride:1x1 activation:relu |\n", "| | | 3x3x16 (int8) | | |\n", "| | | 32 (int32) | | |\n", "| 6 | max_pool_2d | 17x17x32 (int8) | 8x8x32 (int8) | Padding:valid stride:2x2 filter:2x2 activation:none |\n", "| 7 | reshape | 8x8x32 (int8) | 2048 (int8) | BuiltinOptionsType=0 |\n", "| | | 2 (int32) | | |\n", "| 8 | fully_connected | 2048 (int8) | 32 (int8) | Activation:relu |\n", "| | | 2048 (int8) | | |\n", "| | | 32 (int32) | | |\n", "| 9 | fully_connected | 32 (int8) | 4 (int8) | Activation:none |\n", "| | | 32 (int8) | | |\n", "| | | 4 (int32) | | |\n", "| 10 | softmax | 4 (int8) | 4 (int8) | BuiltinOptionsType=9 |\n", "| 11 | dequantize | 4 (int8) | 4 (float32) | BuiltinOptionsType=0 |\n", "+-------+-----------------+-------------------+-----------------+-----------------------------------------------------+\n", "Total MACs: 5.870 M\n", "Total OPs: 12.050 M\n", "Name: rock_paper_scissors\n", "Version: 1\n", "Description: Image classifier example for detecting Rock/Paper/Scissors hand gestures in images\n", "Classes: rock, paper, scissor, _unknown_\n", "hash: 2482ff1c6e512f70479605f20e18e5fc\n", "date: 2022-05-03T23:33:50.754Z\n", "runtime_memory_size: 137176\n", "detection_threshold: 175\n", "average_window_duration_ms: 500\n", "minimum_count: 2\n", "suppression_count: 1\n", "samplewise_norm.rescale: 0.0\n", "samplewise_norm.mean_and_std: True\n", ".tflite file size: 80.2kB\n" ] } ], "source": [ "# Summarize the TF-Lite Model\n", "# This is the quantized model that eventually goes on the embedded device\n", "!mltk summarize rock_paper_scissors --tflite --build" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Model Visualization\n", "\n", "The MLTK also allows for visualizing the model in an interactive webpage.\n", "\n", "This is done using the `view` command.\n", "Refer to the [Model Visualization Guide](../../docs/guides/model_visualizer.md) for more details on how this works.\n", "\n", "__NOTES:__ \n", "- This will open a new tab to your web-browser\n", "- You must click the opened webpage's 'Accept' button the first time it runs (and possibly re-run the command)\n", "- Since we have not trained our model yet, we must add the `--build` option to the command. This is not required once the model is trained.\n", "- This command must run locally, it will not work from a remote terminal/notebook " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Visualize Keras model\n", "\n", "By default, the `view` command will visualize the [KerasModel](../../docs/python_api/keras_model.md), the model used for training (file extension `.h5`)." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# This will open a new tab in your web browser\n", "# Be sure the click the 'Accept' button in the opened webpage\n", "# (you may need to re-run this command after doing so)\n", "!mltk view rock_paper_scissors --build" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Visualize TF-Lite model\n", "\n", "Alternatively, the `--tflite` flag can be used to view the [TfliteModel](../../docs/python_api/tflite_model/index.md), the quantized model that is programmed to the embedded device (file extension `.tflite`).\n", "\n", "Note that the structure of the Keras and TfLite models are similar, but the TfLite model is a bit more simple. This is because the [TF-Lite Converter](https://www.tensorflow.org/lite/convert) optimized the model by merging/fusing as many layers as possible." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# This will open a new tab in your web browser\n", "# Be sure the click the 'Accept' button in the opened webpage\n", "# (you may need to re-run this command after doing so)\n", "!mltk view rock_paper_scissors --tflite --build" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Model Profiler\n", "\n", "Before spending the time and energy to train the model, it may be useful to profile the model to determine how efficiently it may run on the embedded device.\n", "If it's determined that the model does not fit within the time or memory constraints, then the model layout should be adjusted, the model input size should be reduced, and/or a different model should be selected.\n", "\n", "For this reason, th MLTK features a model profiler. Refer to the [Model Profiler Guide](../../docs/guides/model_profiler.md) for more details.\n", "\n", "__NOTE:__ The following examples use the `--build` flag since the model has not been trained yet. Once the model is trained this flag is no longer needed." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Profile in simulator\n", "\n", "The following command will profile our model in the MVP hardware simulator and return estimates about the time and energy the model might require on the embedded device. \n", "\n", "__NOTES:__ \n", "- An embedded device does not needed to be locally connected to run this command.\n", "- Remove the `--accelerator MVP` option if you are targeting a device that does not have an MVP hardware accelerator." ] }, { "cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/scissor/2022-04-29T23-01-25.981.jpg not found in existing index, re-generating index\n", "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/rock/2022-04-29T23-13-28.550.jpg not found in existing index, re-generating index\n", "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/paper/2022-04-29T23-05-47.387.jpg not found in existing index, re-generating index\n", "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/_unknown_/2022-04-29T22-04-13.350.jpg not found in existing index, re-generating index\n", "C:\\Users\\reed\\workspace\\silabs\\mltk\\.venv\\lib\\site-packages\\tensorflow\\lite\\python\\convert.py:746: UserWarning: Statistics for quantized inputs were expected, but not specified; continuing anyway.\n", " warnings.warn(\"Statistics for quantized inputs were expected, but not \"\n", "\n", "Profiling Summary\n", "Name: rock_paper_scissors\n", "Accelerator: MVP\n", "Input Shape: 1x84x84x1\n", "Input Data Type: float32\n", "Output Shape: 1x4\n", "Output Data Type: float32\n", "Flash, Model File Size (bytes): 80.2k\n", "RAM, Runtime Memory Size (bytes): 137.2k\n", "Operation Count: 12.3M\n", "Multiply-Accumulate Count: 5.9M\n", "Layer Count: 12\n", "Unsupported Layer Count: 0\n", "Accelerator Cycle Count: 11.0M\n", "CPU Cycle Count: 345.4k\n", "CPU Utilization (%): 3.1\n", "Clock Rate (hz): 80.0M\n", "Time (s): 141.5m\n", "Energy (J): 1.5m\n", "J/Op: 122.3p\n", "J/MAC: 256.8p\n", "Ops/s: 87.2M\n", "MACs/s: 41.5M\n", "Inference/s: 7.1\n", "\n", "Model Layers\n", "+-------+-----------------+--------+--------+------------+------------+------------+----------+-------------------------+--------------+-----------------------------------------------------+\n", "| Index | OpCode | # Ops | # MACs | Acc Cycles | CPU Cycles | Energy (J) | Time (s) | Input Shape | Output Shape | Options |\n", "+-------+-----------------+--------+--------+------------+------------+------------+----------+-------------------------+--------------+-----------------------------------------------------+\n", "| 0 | quantize | 28.2k | 0 | 0 | 254.7k | 41.3u | 3.2m | 1x84x84x1 | 1x84x84x1 | BuiltinOptionsType=0 |\n", "| 1 | conv_2d | 2.3M | 968.3k | 3.4M | 7.4k | 478.0u | 43.0m | 1x84x84x1,16x3x3x1,16 | 1x82x82x16 | Padding:valid stride:1x1 activation:relu |\n", "| 2 | max_pool_2d | 107.6k | 0 | 80.7k | 16.2k | 9.2u | 1.0m | 1x82x82x16 | 1x41x41x16 | Padding:valid stride:2x2 filter:2x2 activation:none |\n", "| 3 | conv_2d | 7.1M | 3.5M | 5.4M | 7.3k | 745.6u | 66.9m | 1x41x41x16,16x3x3x16,16 | 1x39x39x16 | Padding:valid stride:1x1 activation:relu |\n", "| 4 | max_pool_2d | 23.1k | 0 | 17.3k | 16.2k | 4.0u | 216.6u | 1x39x39x16 | 1x19x19x16 | Padding:valid stride:2x2 filter:2x2 activation:none |\n", "| 5 | conv_2d | 2.7M | 1.3M | 2.0M | 7.2k | 226.1u | 25.4m | 1x19x19x16,32x3x3x16,32 | 1x17x17x32 | Padding:valid stride:1x1 activation:relu |\n", "| 6 | max_pool_2d | 8.2k | 0 | 6.1k | 27.3k | 1.7u | 341.3u | 1x17x17x32 | 1x8x8x32 | Padding:valid stride:2x2 filter:2x2 activation:none |\n", "| 7 | reshape | 0 | 0 | 0 | 217.3 | 0.0p | 2.7u | 1x8x8x32,2 | 1x2048 | BuiltinOptionsType=0 |\n", "| 8 | fully_connected | 131.2k | 65.5k | 98.4k | 1.6k | 1.6u | 1.2m | 1x2048,32x2048,32 | 1x32 | Activation:relu |\n", "| 9 | fully_connected | 260.0 | 128.0 | 208.0 | 1.7k | 7.7n | 21.7u | 1x32,4x32,4 | 1x4 | Activation:none |\n", "| 10 | softmax | 20.0 | 0 | 0 | 4.1k | 16.5n | 51.8u | 1x4 | 1x4 | BuiltinOptionsType=9 |\n", "| 11 | dequantize | 8.0 | 0 | 0 | 1.4k | 159.1n | 17.5u | 1x4 | 1x4 | BuiltinOptionsType=0 |\n", "+-------+-----------------+--------+--------+------------+------------+------------+----------+-------------------------+--------------+-----------------------------------------------------+\n", "Generating profiling report at C:/Users/reed/.mltk/models/rock_paper_scissors-test/profiling\n", "Profiling time: 16.559938 seconds\n" ] } ], "source": [ "!mltk profile rock_paper_scissors --build --accelerator MVP" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Profile on physical device\n", "\n", "Alternatively, if we have a device locally connected, we can directly profile on that instead. This is useful as the returned profiling numbers are \"real\", they are not estimated as they would be in the simulator case. \n", "\n", "To profile on a physical device, simply added the `--device` command flag.\n", "\n", "__NOTES:__ \n", "- An embedded device must be locally connected to run this command.\n", "- Remove the `--accelerator MVP` option if you are targeting a device that does not have an MVP hardware accelerator." ] }, { "cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/scissor/2022-04-29T23-01-25.981.jpg not found in existing index, re-generating index\n", "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/rock/2022-04-29T23-13-28.550.jpg not found in existing index, re-generating index\n", "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/paper/2022-04-29T23-05-47.387.jpg not found in existing index, re-generating index\n", "File C:/Users/reed/.mltk/datasets/rock_paper_scissors/v2/_unknown_/2022-04-29T22-04-13.350.jpg not found in existing index, re-generating index\n", "C:\\Users\\reed\\workspace\\silabs\\mltk\\.venv\\lib\\site-packages\\tensorflow\\lite\\python\\convert.py:746: UserWarning: Statistics for quantized inputs were expected, but not specified; continuing anyway.\n", " warnings.warn(\"Statistics for quantized inputs were expected, but not \"\n", "\n", "Profiling Summary\n", "Name: rock_paper_scissors\n", "Accelerator: MVP\n", "Input Shape: 1x84x84x1\n", "Input Data Type: float32\n", "Output Shape: 1x4\n", "Output Data Type: float32\n", "Flash, Model File Size (bytes): 80.2k\n", "RAM, Runtime Memory Size (bytes): 137.1k\n", "Operation Count: 12.3M\n", "Multiply-Accumulate Count: 5.9M\n", "Layer Count: 12\n", "Unsupported Layer Count: 0\n", "Accelerator Cycle Count: 11.3M\n", "CPU Cycle Count: 358.9k\n", "CPU Utilization (%): 3.1\n", "Clock Rate (hz): 80.0M\n", "Time (s): 142.7m\n", "Ops/s: 86.4M\n", "MACs/s: 41.1M\n", "Inference/s: 7.0\n", "\n", "Model Layers\n", "+-------+-----------------+--------+--------+------------+------------+----------+-------------------------+--------------+-----------------------------------------------------+\n", "| Index | OpCode | # Ops | # MACs | Acc Cycles | CPU Cycles | Time (s) | Input Shape | Output Shape | Options |\n", "+-------+-----------------+--------+--------+------------+------------+----------+-------------------------+--------------+-----------------------------------------------------+\n", "| 0 | quantize | 28.2k | 0 | 0 | 254.8k | 3.1m | 1x84x84x1 | 1x84x84x1 | BuiltinOptionsType=0 |\n", "| 1 | conv_2d | 2.2M | 968.2k | 3.6M | 9.3k | 44.9m | 1x84x84x1,16x3x3x1,16 | 1x82x82x16 | Padding:valid stride:1x1 activation:relu |\n", "| 2 | max_pool_2d | 107.6k | 0 | 80.8k | 15.0k | 1.1m | 1x82x82x16 | 1x41x41x16 | Padding:valid stride:2x2 filter:2x2 activation:none |\n", "| 3 | conv_2d | 7.1M | 3.5M | 5.4M | 8.5k | 66.2m | 1x41x41x16,16x3x3x16,16 | 1x39x39x16 | Padding:valid stride:1x1 activation:relu |\n", "| 4 | max_pool_2d | 23.1k | 0 | 17.4k | 14.9k | 300.0u | 1x39x39x16 | 1x19x19x16 | Padding:valid stride:2x2 filter:2x2 activation:none |\n", "| 5 | conv_2d | 2.7M | 1.3M | 2.0M | 8.5k | 25.2m | 1x19x19x16,32x3x3x16,32 | 1x17x17x32 | Padding:valid stride:1x1 activation:relu |\n", "| 6 | max_pool_2d | 8.2k | 0 | 6.4k | 28.1k | 330.0u | 1x17x17x32 | 1x8x8x32 | Padding:valid stride:2x2 filter:2x2 activation:none |\n", "| 7 | reshape | 0 | 0 | 0 | 10.7k | 150.0u | 1x8x8x32,2 | 1x2048 | BuiltinOptionsType=0 |\n", "| 8 | fully_connected | 131.2k | 65.5k | 98.5k | 2.1k | 1.2m | 1x2048,32x2048,32 | 1x32 | Activation:relu |\n", "| 9 | fully_connected | 260.0 | 128.0 | 231.0 | 1.8k | 30.0u | 1x32,4x32,4 | 1x4 | Activation:none |\n", "| 10 | softmax | 20.0 | 0 | 0 | 4.1k | 60.0u | 1x4 | 1x4 | BuiltinOptionsType=9 |\n", "| 11 | dequantize | 8.0 | 0 | 0 | 1.1k | 0 | 1x4 | 1x4 | BuiltinOptionsType=0 |\n", "+-------+-----------------+--------+--------+------------+------------+----------+-------------------------+--------------+-----------------------------------------------------+\n", "Generating profiling report at C:/Users/reed/.mltk/models/rock_paper_scissors-test/profiling\n", "Profiling time: 49.481894 seconds\n" ] } ], "source": [ "!mltk profile rock_paper_scissors --build --device --accelerator MVP" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Note about CPU utilization\n", "\n", "An important metric the model profiler provides when using the MVP hardware accelerator is `CPU Utilization`.\n", "This gives an indication of how much CPU is required to run the machine learning model.\n", "\n", "If no hardware accelerator is used, then the CPU utilization is 100% as 100% of the machine learning model's calculations are executed on the CPU.\n", "With the hardware accelerator, many of the model's calculations can be offloaded to the accelerator freeing the CPU to do other tasks.\n", "\n", "The additional CPU cycles the hardware accelerator provides can be a major benefit, especially when other tasks such as real-time audio processing are required." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Note about model size and hardware constraints\n", "\n", "The model used in this tutorial has already been optimized for the embedded device.\n", "\n", "The original model [definition](https://gist.github.com/fchollet/f35fbc80e066a49d65f1688a7e99f069) used 32 Conv2D filters in the first layer and the input images were 96x96.\n", "\n", "If you revert your model specification to these parameters, e.g.:\n", "\n", "```python\n", "my_model.input_shape = (96,96,1)\n", "```\n", "\n", "```python\n", "def my_model_builder(model: MyModel):\n", " keras_model = Sequential()\n", "\n", " # Increasing this value can increase model accuracy \n", " # at the expense of more RAM and execution latency\n", " filter_count = 32 \n", "```\n", "\n", "Then profile the model:\n", "\n", "```\n", "mltk profile rock_paper_scissors --accelerator MVP --build\n", "```\n", "\n", "You'll find that the model exceeds the available RAM.\n", "\n", "Using the tips in the FAQ question [How can I reduce my model's size](../../docs/faq/how_to_reduce_model_size.md), the model parameters were reduced to:\n", "\n", "```python\n", "my_model.input_shape = (84,84,1)\n", "```\n", "\n", "```python\n", "def my_model_builder(model: MyModel):\n", " keras_model = Sequential()\n", "\n", " # Increasing this value can increase model accuracy \n", " # at the expense of more RAM and execution latency\n", " filter_count = 16 \n", "```\n", "\n", "Which produces a model that is much more suitable for the embedded hardware." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Model Training\n", "\n", "Now that we have our model fully specified and it fits within the constraints of the embedded device, we can train the model.\n", "\n", "The basic flow for model training is:\n", "1. Invoke the `train` command\n", "2. Tensorflow trains the model\n", "3. A [Model Archive](../../docs/guides/model_archive.md) containing the trained model is generated in the same directory as the model specification script \n", "\n", "Refer to the [Model Training Guide](../../docs/guides/model_training.md) for more details about this process." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Train as a \"dry run\"\n", "\n", "Before fully training the model, sometimes it is useful to train the model as a \"dry run\" to ensure the end-to-end training process works. Here, the model is trained for a few epochs on a subset of the dataset.\n", "\n", "To train as a dry run, append `-test` to the model name. \n", "At the end of training, a [Model Archive](../../docs/guides/model_archive.md) with `-test` appended to the archive name is generated in the same directory as the model specification script. " ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# Train as a dry run by appending \"-test\" to the model name\n", "!mltk train rock_paper_scissors-test" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Training locally\n", "\n", "One option for training your model is to run the `train` command in your local terminal. \n", "Most of the models used by embedded devices are small enough that this is a feasible option. \n", "Never the less, this is a very CPU intensive operation. Many times it's best to issue the `train` command and let it run over night." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# Be sure to replace \"rock_paper_scissors\"\n", "# with the name of your model\n", "# WARNING: This command may take several hours\n", "!mltk train rock_paper_scissors" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Train in cloud\n", "\n", "Alternatively, you can _vastly_ improve the model training time by training this model in the \"cloud\". \n", "See the tutorial: [Cloud Training with vast.ai](../../mltk/tutorials/cloud_training_with_vast_ai.md) for more details." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Model Evaluation\n", "\n", "With our model trained, we can now evaluate it to see how accurate it is.\n", "\n", "The basic idea behind model evaluation is to send test samples (i.e. new, unknown samples the model was _not_ trained with) through the model, and compare the model's predictions versus the expected values. If all the model predictions match the expected values then the model is 100% accurate, and every wrong prediction decreases the model accuracy, e.g.:\n", "\n", "![Model Accuracy](https://bit.ly/3w9xQXV) \n", "\n", "Assuming the test samples are _representative_ then the model accuracy should indicate how well it will perform in the real-world.\n", "\n", "Model evaluation is done using the `evaluate` MLTK command. Along with accuracy, the `evaluate` command generates other statistics such as [ROC-AUC](https://en.wikipedia.org/wiki/Receiver_operating_characteristic) and [Precision & Recall](https://en.wikipedia.org/wiki/Precision_and_recall). \n", "Refer to the [Model Evaluation Guide](../../docs/guides/model_evaluation.md) for more details about using the MLTK for model evaluation." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Command\n", "\n", "To evaluate the newly trained model, issue the following command:\n", "\n", "__NOTE:__ Be sure to replace `rock_paper_scissors` with the name of your model." ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "# Run the model evaluation command\n", "!mltk evaluate rock_paper_scissors --tflite --show" ] }, { "cell_type": "code", "execution_count": 24, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Name: rock_paper_scissors\n", "Model Type: classification\n", "Overall accuracy: 95.037%\n", "Class accuracies:\n", "- paper = 98.394%\n", "- scissor = 97.500%\n", "- rock = 92.476%\n", "- _unknown_ = 92.083%\n", "Average ROC AUC: 98.664%\n", "Class ROC AUC:\n", "- paper = 99.461%\n", "- _unknown_ = 99.042%\n", "- scissor = 98.554%\n", "- rock = 97.600%\n", "\n" ] } ], "source": [ "# For documentation purposes, we use the evaluate_model Python API so\n", "# the evaluation plots are generated inline with the docs\n", "from mltk.core import evaluate_model \n", "evaluation_results = evaluate_model('rock_paper_scissors', tflite=True, show=True)\n", "print(f'{evaluation_results}')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "So in this case, our model has a 95.0% overall accuracy. \n", "\n", "Once again, please refer to the [Model Evaluation Guide](../../docs/guides/model_evaluation.md) for more details about the various metrics generated by this command." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Note about model accuracy\n", "\n", "The following are things to keep in mind to improve the model accuracy: \n", "- __Verify the dataset__ - Ensure all the samples are properly labeled and in a consistent format\n", "- __Add more representative dataset__ - The more representative samples that are in the dataset, the better chance the model has at learning the important features in the samples\n", "- __Increase the model size__ - Increase the model size by adding more or wider layers (e.g. add more Conv2D filers)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Model Testing\n", "\n", "__NOTE:__ This section is __experimental__ and is optional for the rest of this tutorial.\n", "You may safely skip to the next section.\n", "\n", "To help evaluate the model's performance on real hardware, the MLTK offers the command: `classify_image`. With this command, the trained model can be used to classify images captured by an embedded camera. The `classify_image` command features: \n", "- Support for executing a model on a supported embedded development board\n", "- Support for dumping the captured images to the local PC\n", "- Support for displaying capture image on the local PC in real-time\n", "- Support for adjusting the detection threshold\n", "- Support for viewing the model prediction results in real-time\n", "\n", "__NOTE:__ The `classify_image` command must run locally. It will not work remotely (e.g. on Colab or remote SSH)\n", "\n", "See the output of the command help for more details:" ] }, { "cell_type": "code", "execution_count": 25, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Usage: mltk classify_image [OPTIONS] \n", "\n", " Classify images detected by a camera connected to an embedded device.\n", "\n", " NOTE: A supported embedded device must be locally connected to use this\n", " command.\n", "\n", "Arguments:\n", " On of the following:\n", " - MLTK model name \n", " - Path to .tflite file\n", " - Path to model archive file (.mltk.zip)\n", " NOTE: The model must have been previously trained for image classification [required]\n", "\n", "Options:\n", " -a, --accelerator Name of accelerator to use while executing the audio classification ML model\n", " --port Serial COM port of a locally connected embedded device.\n", " 'If omitted, then attempt to automatically determine the serial COM port\n", " -v, --verbose Enable verbose console logs\n", " -w, --window_duration \n", " Controls the smoothing. Drop all inference results that are older than minus window_duration.\n", " Longer durations (in milliseconds) will give a higher confidence that the results are correct, but may miss some images\n", " -c, --count The *minimum* number of inference results to\n", " average when calculating the detection value\n", " -t, --threshold Minimum averaged model output threshold for\n", " a class to be considered detected, 0-255.\n", " Higher values increase precision at the cost\n", " of recall\n", " -s, --suppression Number of samples that should be different\n", " than the last detected sample before\n", " detecting again\n", " -l, --latency This the amount of time in milliseconds\n", " between processing loops\n", " -i, --sensitivity FLOAT Sensitivity of the activity indicator LED.\n", " Much less than 1.0 has higher sensitivity\n", " -x, --dump-images Dump the raw images from the device camera to a directory on the local PC. \n", " NOTE: Use the --no-inference option to ONLY dump images and NOT run inference on the device\n", " Use the --dump-threshold option to control how unique the images must be to dump\n", " --dump-threshold FLOAT This controls how unique the camera images must be before they're dumped.\n", " This is useful when generating a dataset.\n", " If this value is set to 0 then every image from the camera is dumped.\n", " if this value is closer to 1. then the images from the camera should be sufficiently unique from\n", " prior images that have been dumped. [default: 0.1]\n", " --no-inference By default inference is executed on the\n", " device. Use --no-inference to disable\n", " inference on the device which can improve\n", " image dumping throughput\n", " -g, --generate-dataset Update the model's dataset.\n", " This will iterate through each data class used by the model and instruct the user\n", " the display the class in front of the camera. An image is captured from the device's camera\n", " and saved to the model's corresponding dataset sub-directory.\n", " This process will repeat until the user exits the command. \n", " Use the --sample-count option to specify the number of samples per class to collect\n", " NOTE: Device inference is disabled when using this option \n", " See the --dump-images option as an alternative to generating a dataset \n", " --sample-count INTEGER The number of samples to collect per class\n", " before iterating to the next class\n", " [default: 5]\n", " --app By default, the image_classifier app is automatically downloaded. \n", " This option allows for overriding with a custom built app.\n", " Alternatively, set this option to \"none\" to NOT program the image_classifier app to the device.\n", " In this case, ONLY the .tflite will be programmed and the existing image_classifier app will be re-used.\n", " --test Run as a unit test\n", " --help Show this message and exit.\n" ] } ], "source": [ "!mltk classify_image --help" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "To run this command with the trained model, issue the command:\n", "\n", "```shell\n", "# Run the image classifier with the trained model\n", "# Use the MVP hardware accelerator\n", "# Verbosely print the inference results\n", "# Dump images to the local PC\n", "# NOTE: This command must run from a local terminal\n", "mltk classify_image rock_paper_scissors --accelerator MVP --verbose --dump-images\n", "```" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Deploying the Model\n", "\n", "Now that we have a trained model, it is time to run it in on an embedded device.\n", "\n", "There are several different ways this can be done:\n", "\n", "\n", "### Using the MLTK\n", "\n", "The MLTK supports building [C++ Applications](../../docs/cpp_development/index.md).\n", "\n", "It also features an [image_classifier](../../docs/cpp_development/examples/image_classifier.md) C++ application\n", "which can be built using: \n", "- [Visual Studio Code](../../docs/cpp_development/vscode.md) \n", "- [Simplicity Studio](../../docs/cpp_development/simplicity_studio.md)\n", "- [Command Line](../../docs/cpp_development/command_line.md)\n", "\n", "Refer to the [image_classifier](../../docs/cpp_development/examples/image_classifier.md) application's documentation\n", "for how include your model into the built application." ] } ], "metadata": { "kernelspec": { "display_name": "Python 3.9.7 ('.venv': venv)", "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": "3.9.7" }, "orig_nbformat": 4, "vscode": { "interpreter": { "hash": "600e22ae316f8c315f552eaf99bb679bc9438a443c93affde9ac001991b79c8f" } } }, "nbformat": 4, "nbformat_minor": 2 }