commit e80de437552874bb25304b963e14670382e4ee07
parent 2d7379654f2f5b1fe99e02c111c5a1ba09b3a371
Author: Andrew <andrewlaack1@gmail.com>
Date: Wed, 23 Oct 2024 17:09:36 -0500
Did stuff in stream.
Diffstat:
2 files changed, 1693 insertions(+), 0 deletions(-)
diff --git a/nnPruning/NaivePruningFashion.ipynb b/nnPruning/NaivePruningFashion.ipynb
@@ -0,0 +1,761 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "* 89.84% Accuracy on test set before pruning\n",
+ "* 89.61% Accuracy on test set after pruning arbitrary neurons\n",
+ "\n",
+ "The degradation in accuracy here was .23% when pruning ~2.4% of neurons compared to the 1.25% degradation when not using dropout"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "FINDINGS:\n",
+ "\n",
+ "* Dropout increases polysemanticity and resilience (intuitive)\n",
+ "* Pruning neurons had a nominal impact upon accuracy (surprising)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 247,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import pandas as pd"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 248,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "train_set = pd.read_csv('../datasets/fashion/fashion-mnist_train.csv')\n",
+ "test_set = pd.read_csv('../datasets/fashion/fashion-mnist_test.csv')"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 249,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "y_train = train_set['label']\n",
+ "X_train = train_set.drop(axis=1, labels=['label'])\n",
+ "\n",
+ "y_test = test_set['label']\n",
+ "X_test = test_set.drop(axis=1, labels=['label'])"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 250,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from sklearn.preprocessing import OneHotEncoder\n",
+ "import numpy as np\n",
+ "\n",
+ "ohec = OneHotEncoder(sparse_output=False)\n",
+ "\n",
+ "y_train = np.array(y_train)\n",
+ "y_test = np.array(y_test)\n",
+ "\n",
+ "y_train = y_train.reshape(-1,1)\n",
+ "y_test = y_test.reshape(-1,1)\n",
+ "\n",
+ "\n",
+ "y_train = ohec.fit_transform(y_train)\n",
+ "y_test = ohec.fit_transform(y_test)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 251,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential_15\"</span>\n",
+ "</pre>\n"
+ ],
+ "text/plain": [
+ "\u001b[1mModel: \"sequential_15\"\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
+ "┃<span style=\"font-weight: bold\"> Layer (type) </span>┃<span style=\"font-weight: bold\"> Output Shape </span>┃<span style=\"font-weight: bold\"> Param # </span>┃\n",
+ "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
+ "│ dense_57 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">401,920</span> │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_58 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">262,656</span> │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ batch_normalization_11 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">2,048</span> │\n",
+ "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BatchNormalization</span>) │ │ │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_59 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">131,328</span> │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ batch_normalization_12 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">1,024</span> │\n",
+ "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BatchNormalization</span>) │ │ │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dropout_7 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_60 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">128</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">32,896</span> │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_61 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">10</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">1,290</span> │\n",
+ "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
+ "</pre>\n"
+ ],
+ "text/plain": [
+ "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
+ "┃\u001b[1m \u001b[0m\u001b[1mLayer (type) \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
+ "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
+ "│ dense_57 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m) │ \u001b[38;5;34m401,920\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_58 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m) │ \u001b[38;5;34m262,656\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ batch_normalization_11 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m) │ \u001b[38;5;34m2,048\u001b[0m │\n",
+ "│ (\u001b[38;5;33mBatchNormalization\u001b[0m) │ │ │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_59 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m) │ \u001b[38;5;34m131,328\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ batch_normalization_12 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m) │ \u001b[38;5;34m1,024\u001b[0m │\n",
+ "│ (\u001b[38;5;33mBatchNormalization\u001b[0m) │ │ │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dropout_7 (\u001b[38;5;33mDropout\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m) │ \u001b[38;5;34m0\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_60 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m128\u001b[0m) │ \u001b[38;5;34m32,896\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_61 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m10\u001b[0m) │ \u001b[38;5;34m1,290\u001b[0m │\n",
+ "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">833,162</span> (3.18 MB)\n",
+ "</pre>\n"
+ ],
+ "text/plain": [
+ "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m833,162\u001b[0m (3.18 MB)\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">831,626</span> (3.17 MB)\n",
+ "</pre>\n"
+ ],
+ "text/plain": [
+ "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m831,626\u001b[0m (3.17 MB)\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">1,536</span> (6.00 KB)\n",
+ "</pre>\n"
+ ],
+ "text/plain": [
+ "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m1,536\u001b[0m (6.00 KB)\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "import keras\n",
+ "import tensorflow as tf\n",
+ "\n",
+ "model = keras.Sequential(layers=[\n",
+ " keras.layers.Input(shape=(784,)),\n",
+ " keras.layers.Dense(512, activation='relu'),\n",
+ " keras.layers.Dense(512, activation='relu'),\n",
+ " keras.layers.BatchNormalization(),\n",
+ " keras.layers.Dense(256, activation='relu'),\n",
+ " keras.layers.BatchNormalization(),\n",
+ " keras.layers.Dropout(.2),\n",
+ " keras.layers.Dense(128, activation='relu'),\n",
+ " keras.layers.Dense(10, activation='softmax')\n",
+ " ]\n",
+ ")\n",
+ "model.summary()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 252,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "model.compile(loss=keras.losses.CategoricalCrossentropy, optimizer='adam', metrics=['accuracy'])"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 253,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Epoch 1/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 31ms/step - accuracy: 0.7238 - loss: 0.8328\n",
+ "Epoch 2/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 35ms/step - accuracy: 0.8638 - loss: 0.3770\n",
+ "Epoch 3/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 30ms/step - accuracy: 0.8770 - loss: 0.3335\n",
+ "Epoch 4/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 35ms/step - accuracy: 0.8874 - loss: 0.3060\n",
+ "Epoch 5/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 39ms/step - accuracy: 0.8958 - loss: 0.2791\n",
+ "Epoch 6/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 39ms/step - accuracy: 0.9008 - loss: 0.2675\n",
+ "Epoch 7/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 40ms/step - accuracy: 0.9040 - loss: 0.2596\n",
+ "Epoch 8/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 60ms/step - accuracy: 0.9065 - loss: 0.2471\n",
+ "Epoch 9/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 43ms/step - accuracy: 0.9051 - loss: 0.2535\n",
+ "Epoch 10/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 36ms/step - accuracy: 0.9045 - loss: 0.2558\n",
+ "Epoch 11/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9090 - loss: 0.2420\n",
+ "Epoch 12/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 34ms/step - accuracy: 0.9142 - loss: 0.2272\n",
+ "Epoch 13/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9238 - loss: 0.2070\n",
+ "Epoch 14/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 36ms/step - accuracy: 0.9238 - loss: 0.2045\n",
+ "Epoch 15/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 33ms/step - accuracy: 0.9202 - loss: 0.2110\n",
+ "Epoch 16/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 34ms/step - accuracy: 0.9266 - loss: 0.1918\n",
+ "Epoch 17/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 35ms/step - accuracy: 0.9311 - loss: 0.1818\n",
+ "Epoch 18/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 39ms/step - accuracy: 0.9359 - loss: 0.1736\n",
+ "Epoch 19/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 34ms/step - accuracy: 0.9357 - loss: 0.1710\n",
+ "Epoch 20/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 37ms/step - accuracy: 0.9377 - loss: 0.1657\n",
+ "Epoch 21/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 41ms/step - accuracy: 0.9394 - loss: 0.1588\n",
+ "Epoch 22/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 37ms/step - accuracy: 0.9391 - loss: 0.1605\n",
+ "Epoch 23/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 39ms/step - accuracy: 0.9446 - loss: 0.1479\n",
+ "Epoch 24/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 36ms/step - accuracy: 0.9472 - loss: 0.1403\n",
+ "Epoch 25/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 41ms/step - accuracy: 0.9478 - loss: 0.1344\n",
+ "Epoch 26/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 42ms/step - accuracy: 0.9446 - loss: 0.1471\n",
+ "Epoch 27/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 37ms/step - accuracy: 0.9513 - loss: 0.1278\n",
+ "Epoch 28/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 39ms/step - accuracy: 0.9532 - loss: 0.1221\n",
+ "Epoch 29/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9531 - loss: 0.1228\n",
+ "Epoch 30/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 35ms/step - accuracy: 0.9559 - loss: 0.1166\n",
+ "Epoch 31/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9491 - loss: 0.1314\n",
+ "Epoch 32/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 43ms/step - accuracy: 0.9562 - loss: 0.1157\n",
+ "Epoch 33/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9587 - loss: 0.1093\n",
+ "Epoch 34/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 31ms/step - accuracy: 0.9582 - loss: 0.1092\n",
+ "Epoch 35/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 30ms/step - accuracy: 0.9612 - loss: 0.1025\n",
+ "Epoch 36/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 29ms/step - accuracy: 0.9637 - loss: 0.0972\n",
+ "Epoch 37/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 30ms/step - accuracy: 0.9649 - loss: 0.0905\n",
+ "Epoch 38/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9647 - loss: 0.0953\n",
+ "Epoch 39/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 30ms/step - accuracy: 0.9659 - loss: 0.0911\n",
+ "Epoch 40/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 36ms/step - accuracy: 0.9687 - loss: 0.0828\n",
+ "Epoch 41/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 33ms/step - accuracy: 0.9691 - loss: 0.0804\n",
+ "Epoch 42/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 35ms/step - accuracy: 0.9658 - loss: 0.0893\n",
+ "Epoch 43/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 34ms/step - accuracy: 0.9642 - loss: 0.0943\n",
+ "Epoch 44/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 36ms/step - accuracy: 0.9681 - loss: 0.0822\n",
+ "Epoch 45/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 36ms/step - accuracy: 0.9703 - loss: 0.0773\n",
+ "Epoch 46/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 38ms/step - accuracy: 0.9723 - loss: 0.0719\n",
+ "Epoch 47/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 38ms/step - accuracy: 0.9716 - loss: 0.0723\n",
+ "Epoch 48/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 39ms/step - accuracy: 0.9732 - loss: 0.0703\n",
+ "Epoch 49/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 37ms/step - accuracy: 0.9717 - loss: 0.0743\n",
+ "Epoch 50/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 40ms/step - accuracy: 0.9760 - loss: 0.0639\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "<keras.src.callbacks.history.History at 0x7f746f4076d0>"
+ ]
+ },
+ "execution_count": 253,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "model.fit(x=X_train, y=y_train, batch_size=1000, epochs=50)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 254,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step\n"
+ ]
+ }
+ ],
+ "source": [
+ "y_test_pred = model.predict(x=X_test)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 255,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "array([9.9999624e-01, 1.6376542e-12, 1.1062847e-06, 5.7214750e-11,\n",
+ " 5.1883814e-07, 1.2025164e-10, 2.0674804e-06, 2.9590450e-13,\n",
+ " 7.2374801e-11, 7.2594625e-11], dtype=float32)"
+ ]
+ },
+ "execution_count": 255,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "y_test_pred[0]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 256,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "array([1., 0., 0., 0., 0., 0., 0., 0., 0., 0.])"
+ ]
+ },
+ "execution_count": 256,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "y_test[0]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 257,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "y_test_pred_proper = []\n",
+ "\n",
+ "for i in y_test_pred:\n",
+ " y_test_pred_proper.append(i.argmax())"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 258,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "trueY = []\n",
+ "for i in y_test:\n",
+ " trueY.append(i.argmax())"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 259,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "[0, 1, 2, 2, 3, 6, 8, 6, 5, 0]"
+ ]
+ },
+ "execution_count": 259,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "y_test_pred_proper[:10]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 260,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "[0, 1, 2, 2, 3, 2, 8, 6, 5, 0]"
+ ]
+ },
+ "execution_count": 260,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "trueY[:10]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 261,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "0.8984"
+ ]
+ },
+ "execution_count": 261,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "from sklearn.metrics import accuracy_score\n",
+ "\n",
+ "accuracy_score(y_true=trueY, y_pred=y_test_pred_proper)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 262,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "[[864 4 21 14 2 1 89 0 5 0]\n",
+ " [ 0 988 1 8 1 0 1 0 1 0]\n",
+ " [ 22 0 858 11 56 1 49 0 3 0]\n",
+ " [ 26 10 7 898 39 1 17 0 2 0]\n",
+ " [ 2 3 90 14 850 0 39 0 2 0]\n",
+ " [ 0 0 0 1 0 953 0 26 3 17]\n",
+ " [137 4 79 20 69 1 683 0 7 0]\n",
+ " [ 0 0 0 0 0 15 0 972 0 13]\n",
+ " [ 5 2 2 2 2 0 9 4 972 2]\n",
+ " [ 0 0 0 0 0 10 0 44 0 946]]\n"
+ ]
+ }
+ ],
+ "source": [
+ "from sklearn.metrics import confusion_matrix\n",
+ "\n",
+ "cm = confusion_matrix(y_true=trueY, y_pred=y_test_pred_proper)\n",
+ "print(cm)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 263,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ "<Figure size 640x480 with 2 Axes>"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "import matplotlib.pyplot as plt\n",
+ "import seaborn as sns\n",
+ "\n",
+ "sns.heatmap(cm, annot=True, fmt=\"d\")\n",
+ "plt.xlabel('Predicted')\n",
+ "plt.ylabel('Actual')\n",
+ "plt.title('Confusion Matrix')\n",
+ "plt.show()\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "* 0 T-shirt/top\n",
+ "* 1 Trouser\n",
+ "* 2 Pullover\n",
+ "* 3 Dress\n",
+ "* 4 Coat\n",
+ "* 5 Sandal\n",
+ "* 6 Shirt\n",
+ "* 7 Sneaker\n",
+ "* 8 Bag\n",
+ "* 9 Ankle boot"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "1408 neurons * .024 percent ~= 34 neurons"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "# get second hidden layer (input not included in layer calculation)\n",
+ "\n",
+ "\n",
+ "weights = model.get_layer(index=1).get_weights()\n",
+ "\n",
+ "# Weights:\n",
+ "# First dimension is spec neuron\n",
+ "# Second dimension are the associated weights\n",
+ "\n",
+ "weights"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 271,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import random \n",
+ "# 34/512 neurons to pick\n",
+ "\n",
+ "specs = []\n",
+ "\n",
+ "for i in range(0, 34):\n",
+ " specs.append(random.randint(0,511))\n",
+ "\n",
+ "specs.sort()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "# Zero out specified weights\n",
+ "\n",
+ "for i in range(0,len(weights)): \n",
+ " for neuron_index in specs:\n",
+ " weights[i][neuron_index] = 0"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 274,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "model.get_layer(index=1).set_weights(weights)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "* Trained model\n",
+ "* Selected random neurons from second hidden layer (512 dense)\n",
+ "* Zeroed all of their weights (outside of model)\n",
+ "* Put that back into the model"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 275,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step\n"
+ ]
+ }
+ ],
+ "source": [
+ "y_test_pred = model.predict(x=X_test)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 276,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "y_test_pred_proper = []\n",
+ "\n",
+ "for i in y_test_pred:\n",
+ " y_test_pred_proper.append(i.argmax())"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 277,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "0.8961"
+ ]
+ },
+ "execution_count": 277,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "from sklearn.metrics import accuracy_score\n",
+ "\n",
+ "accuracy_score(y_true=trueY, y_pred=y_test_pred_proper)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 278,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "[[885 3 20 7 1 1 76 0 7 0]\n",
+ " [ 1 988 0 8 1 0 1 0 1 0]\n",
+ " [ 22 0 842 8 84 1 40 0 3 0]\n",
+ " [ 37 10 10 876 41 1 20 0 5 0]\n",
+ " [ 4 2 75 14 862 0 38 0 5 0]\n",
+ " [ 0 0 0 1 0 948 0 28 5 18]\n",
+ " [168 3 77 16 63 1 663 0 9 0]\n",
+ " [ 0 0 0 0 0 12 0 972 0 16]\n",
+ " [ 7 1 2 2 1 1 6 4 974 2]\n",
+ " [ 0 0 0 0 0 8 0 41 0 951]]\n"
+ ]
+ }
+ ],
+ "source": [
+ "from sklearn.metrics import confusion_matrix\n",
+ "\n",
+ "cm = confusion_matrix(y_true=trueY, y_pred=y_test_pred_proper)\n",
+ "print(cm)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 279,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ "<Figure size 640x480 with 2 Axes>"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "import matplotlib.pyplot as plt\n",
+ "import seaborn as sns\n",
+ "\n",
+ "sns.heatmap(cm, annot=True, fmt=\"d\")\n",
+ "plt.xlabel('Predicted')\n",
+ "plt.ylabel('Actual')\n",
+ "plt.title('Confusion Matrix')\n",
+ "plt.show()"
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": ".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": "3.11.2"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}
diff --git a/nnPruning/NaivePruningFashionNoDropout.ipynb b/nnPruning/NaivePruningFashionNoDropout.ipynb
@@ -0,0 +1,932 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "* 89.38% Accuracy on test set before pruning\n",
+ "* 88.13% Accuracy on test set after pruning arbitrary neurons\n",
+ "\n",
+ "The degradation in accuracy here was 1.25% when pruning ~2.4% of neurons compared to the .23% when using dropout."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 27,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "1.25"
+ ]
+ },
+ "execution_count": 27,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "89.38 - 88.13"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import pandas as pd"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "train_set = pd.read_csv('../datasets/fashion/fashion-mnist_train.csv')\n",
+ "test_set = pd.read_csv('../datasets/fashion/fashion-mnist_test.csv')"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "y_train = train_set['label']\n",
+ "X_train = train_set.drop(axis=1, labels=['label'])\n",
+ "\n",
+ "y_test = test_set['label']\n",
+ "X_test = test_set.drop(axis=1, labels=['label'])"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from sklearn.preprocessing import OneHotEncoder\n",
+ "import numpy as np\n",
+ "\n",
+ "ohec = OneHotEncoder(sparse_output=False)\n",
+ "\n",
+ "y_train = np.array(y_train)\n",
+ "y_test = np.array(y_test)\n",
+ "\n",
+ "y_train = y_train.reshape(-1,1)\n",
+ "y_test = y_test.reshape(-1,1)\n",
+ "\n",
+ "\n",
+ "y_train = ohec.fit_transform(y_train)\n",
+ "y_test = ohec.fit_transform(y_test)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 5,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "2024-10-23 16:21:47.399066: I external/local_tsl/tsl/cuda/cudart_stub.cc:32] Could not find cuda drivers on your machine, GPU will not be used.\n",
+ "2024-10-23 16:21:47.402499: I external/local_tsl/tsl/cuda/cudart_stub.cc:32] Could not find cuda drivers on your machine, GPU will not be used.\n",
+ "2024-10-23 16:21:47.437251: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n",
+ "To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
+ "2024-10-23 16:21:48.269808: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT\n",
+ "2024-10-23 16:21:48.798517: I external/local_xla/xla/stream_executor/cuda/cuda_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355\n",
+ "2024-10-23 16:21:48.799002: W tensorflow/core/common_runtime/gpu/gpu_device.cc:2251] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at https://www.tensorflow.org/install/gpu for how to download and setup the required libraries for your platform.\n",
+ "Skipping registering GPU devices...\n"
+ ]
+ },
+ {
+ "data": {
+ "text/html": [
+ "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential\"</span>\n",
+ "</pre>\n"
+ ],
+ "text/plain": [
+ "\u001b[1mModel: \"sequential\"\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
+ "┃<span style=\"font-weight: bold\"> Layer (type) </span>┃<span style=\"font-weight: bold\"> Output Shape </span>┃<span style=\"font-weight: bold\"> Param # </span>┃\n",
+ "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
+ "│ dense (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">401,920</span> │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">262,656</span> │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ batch_normalization │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">2,048</span> │\n",
+ "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BatchNormalization</span>) │ │ │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">131,328</span> │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ batch_normalization_1 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">1,024</span> │\n",
+ "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BatchNormalization</span>) │ │ │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_3 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">128</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">32,896</span> │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_4 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>) │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">10</span>) │ <span style=\"color: #00af00; text-decoration-color: #00af00\">1,290</span> │\n",
+ "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
+ "</pre>\n"
+ ],
+ "text/plain": [
+ "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
+ "┃\u001b[1m \u001b[0m\u001b[1mLayer (type) \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
+ "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
+ "│ dense (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m) │ \u001b[38;5;34m401,920\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_1 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m) │ \u001b[38;5;34m262,656\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ batch_normalization │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m) │ \u001b[38;5;34m2,048\u001b[0m │\n",
+ "│ (\u001b[38;5;33mBatchNormalization\u001b[0m) │ │ │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_2 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m) │ \u001b[38;5;34m131,328\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ batch_normalization_1 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m) │ \u001b[38;5;34m1,024\u001b[0m │\n",
+ "│ (\u001b[38;5;33mBatchNormalization\u001b[0m) │ │ │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_3 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m128\u001b[0m) │ \u001b[38;5;34m32,896\u001b[0m │\n",
+ "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
+ "│ dense_4 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m10\u001b[0m) │ \u001b[38;5;34m1,290\u001b[0m │\n",
+ "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">833,162</span> (3.18 MB)\n",
+ "</pre>\n"
+ ],
+ "text/plain": [
+ "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m833,162\u001b[0m (3.18 MB)\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">831,626</span> (3.17 MB)\n",
+ "</pre>\n"
+ ],
+ "text/plain": [
+ "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m831,626\u001b[0m (3.17 MB)\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">1,536</span> (6.00 KB)\n",
+ "</pre>\n"
+ ],
+ "text/plain": [
+ "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m1,536\u001b[0m (6.00 KB)\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "import keras\n",
+ "import tensorflow as tf\n",
+ "\n",
+ "model = keras.Sequential(layers=[\n",
+ " keras.layers.Input(shape=(784,)),\n",
+ " keras.layers.Dense(512, activation='relu'),\n",
+ " keras.layers.Dense(512, activation='relu'),\n",
+ " keras.layers.BatchNormalization(),\n",
+ " keras.layers.Dense(256, activation='relu'),\n",
+ " keras.layers.BatchNormalization(),\n",
+ " keras.layers.Dense(128, activation='relu'),\n",
+ " keras.layers.Dense(10, activation='softmax')\n",
+ " ]\n",
+ ")\n",
+ "model.summary()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "model.compile(loss=keras.losses.CategoricalCrossentropy, optimizer='adam', metrics=['accuracy'])"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Epoch 1/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 34ms/step - accuracy: 0.7436 - loss: 0.7600\n",
+ "Epoch 2/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 31ms/step - accuracy: 0.8742 - loss: 0.3466\n",
+ "Epoch 3/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 31ms/step - accuracy: 0.8841 - loss: 0.3191\n",
+ "Epoch 4/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 31ms/step - accuracy: 0.8898 - loss: 0.3002\n",
+ "Epoch 5/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 30ms/step - accuracy: 0.8963 - loss: 0.2792\n",
+ "Epoch 6/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 30ms/step - accuracy: 0.9011 - loss: 0.2654\n",
+ "Epoch 7/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 36ms/step - accuracy: 0.9096 - loss: 0.2427\n",
+ "Epoch 8/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 36ms/step - accuracy: 0.9152 - loss: 0.2296\n",
+ "Epoch 9/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 36ms/step - accuracy: 0.9202 - loss: 0.2136\n",
+ "Epoch 10/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 38ms/step - accuracy: 0.9195 - loss: 0.2145\n",
+ "Epoch 11/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 34ms/step - accuracy: 0.9051 - loss: 0.2516\n",
+ "Epoch 12/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 37ms/step - accuracy: 0.9167 - loss: 0.2240\n",
+ "Epoch 13/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 35ms/step - accuracy: 0.9199 - loss: 0.2136\n",
+ "Epoch 14/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 34ms/step - accuracy: 0.9212 - loss: 0.2095\n",
+ "Epoch 15/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 30ms/step - accuracy: 0.9244 - loss: 0.2002\n",
+ "Epoch 16/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 33ms/step - accuracy: 0.9306 - loss: 0.1820\n",
+ "Epoch 17/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 31ms/step - accuracy: 0.9370 - loss: 0.1666\n",
+ "Epoch 18/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9393 - loss: 0.1610\n",
+ "Epoch 19/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 29ms/step - accuracy: 0.9433 - loss: 0.1514\n",
+ "Epoch 20/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 39ms/step - accuracy: 0.9451 - loss: 0.1455\n",
+ "Epoch 21/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 37ms/step - accuracy: 0.9469 - loss: 0.1426\n",
+ "Epoch 22/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 31ms/step - accuracy: 0.9520 - loss: 0.1276\n",
+ "Epoch 23/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 30ms/step - accuracy: 0.9542 - loss: 0.1229\n",
+ "Epoch 24/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 32ms/step - accuracy: 0.9575 - loss: 0.1158\n",
+ "Epoch 25/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 41ms/step - accuracy: 0.9605 - loss: 0.1082\n",
+ "Epoch 26/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 34ms/step - accuracy: 0.9628 - loss: 0.1022\n",
+ "Epoch 27/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9627 - loss: 0.1009\n",
+ "Epoch 28/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 33ms/step - accuracy: 0.9649 - loss: 0.0921\n",
+ "Epoch 29/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 34ms/step - accuracy: 0.9661 - loss: 0.0931\n",
+ "Epoch 30/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 35ms/step - accuracy: 0.9677 - loss: 0.0866\n",
+ "Epoch 31/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 29ms/step - accuracy: 0.9709 - loss: 0.0800\n",
+ "Epoch 32/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9671 - loss: 0.0872\n",
+ "Epoch 33/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 34ms/step - accuracy: 0.9708 - loss: 0.0770\n",
+ "Epoch 34/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9742 - loss: 0.0705\n",
+ "Epoch 35/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 32ms/step - accuracy: 0.9739 - loss: 0.0672\n",
+ "Epoch 36/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 37ms/step - accuracy: 0.9783 - loss: 0.0586\n",
+ "Epoch 37/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 33ms/step - accuracy: 0.9752 - loss: 0.0657\n",
+ "Epoch 38/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 34ms/step - accuracy: 0.9815 - loss: 0.0523\n",
+ "Epoch 39/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 35ms/step - accuracy: 0.9785 - loss: 0.0556\n",
+ "Epoch 40/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 33ms/step - accuracy: 0.9809 - loss: 0.0521\n",
+ "Epoch 41/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 34ms/step - accuracy: 0.9812 - loss: 0.0519\n",
+ "Epoch 42/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9805 - loss: 0.0522\n",
+ "Epoch 43/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 37ms/step - accuracy: 0.9803 - loss: 0.0527\n",
+ "Epoch 44/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9816 - loss: 0.0484\n",
+ "Epoch 45/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 35ms/step - accuracy: 0.9838 - loss: 0.0462\n",
+ "Epoch 46/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 35ms/step - accuracy: 0.9857 - loss: 0.0394\n",
+ "Epoch 47/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 36ms/step - accuracy: 0.9832 - loss: 0.0448\n",
+ "Epoch 48/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 35ms/step - accuracy: 0.9876 - loss: 0.0357\n",
+ "Epoch 49/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9842 - loss: 0.0421\n",
+ "Epoch 50/50\n",
+ "\u001b[1m60/60\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 32ms/step - accuracy: 0.9821 - loss: 0.0478\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "<keras.src.callbacks.history.History at 0x7fee5a966bd0>"
+ ]
+ },
+ "execution_count": 7,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "model.fit(x=X_train, y=y_train, batch_size=1000, epochs=50)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step\n"
+ ]
+ }
+ ],
+ "source": [
+ "y_test_pred = model.predict(x=X_test)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 9,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "array([9.9859405e-01, 7.9753104e-10, 6.4784399e-07, 9.1048644e-09,\n",
+ " 1.3203351e-05, 6.1672427e-09, 1.3919529e-03, 3.4279146e-10,\n",
+ " 1.6498035e-07, 1.3554866e-10], dtype=float32)"
+ ]
+ },
+ "execution_count": 9,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "y_test_pred[0]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 10,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "array([1., 0., 0., 0., 0., 0., 0., 0., 0., 0.])"
+ ]
+ },
+ "execution_count": 10,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "y_test[0]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 11,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "y_test_pred_proper = []\n",
+ "\n",
+ "for i in y_test_pred:\n",
+ " y_test_pred_proper.append(i.argmax())"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 12,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "trueY = []\n",
+ "for i in y_test:\n",
+ " trueY.append(i.argmax())"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 13,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "[0, 1, 2, 2, 3, 6, 8, 6, 5, 0]"
+ ]
+ },
+ "execution_count": 13,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "y_test_pred_proper[:10]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 14,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "[0, 1, 2, 2, 3, 2, 8, 6, 5, 0]"
+ ]
+ },
+ "execution_count": 14,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "trueY[:10]"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 15,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "0.8938"
+ ]
+ },
+ "execution_count": 15,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "from sklearn.metrics import accuracy_score\n",
+ "\n",
+ "accuracy_score(y_true=trueY, y_pred=y_test_pred_proper)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 16,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "[[839 0 19 23 5 2 107 0 5 0]\n",
+ " [ 2 993 1 4 0 0 0 0 0 0]\n",
+ " [ 13 1 768 15 123 0 79 0 1 0]\n",
+ " [ 12 24 6 885 47 0 23 1 2 0]\n",
+ " [ 1 1 47 22 886 0 42 0 1 0]\n",
+ " [ 1 0 0 1 0 956 1 16 3 22]\n",
+ " [ 96 5 62 26 75 0 732 0 4 0]\n",
+ " [ 0 0 0 0 0 17 0 941 1 41]\n",
+ " [ 3 0 4 3 2 2 9 1 976 0]\n",
+ " [ 0 0 0 0 0 6 1 28 3 962]]\n"
+ ]
+ }
+ ],
+ "source": [
+ "from sklearn.metrics import confusion_matrix\n",
+ "\n",
+ "cm = confusion_matrix(y_true=trueY, y_pred=y_test_pred_proper)\n",
+ "print(cm)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 17,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "image/png": "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",
+ "text/plain": [
+ "<Figure size 640x480 with 2 Axes>"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "import matplotlib.pyplot as plt\n",
+ "import seaborn as sns\n",
+ "\n",
+ "sns.heatmap(cm, annot=True, fmt=\"d\")\n",
+ "plt.xlabel('Predicted')\n",
+ "plt.ylabel('Actual')\n",
+ "plt.title('Confusion Matrix')\n",
+ "plt.show()\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "* 0 T-shirt/top\n",
+ "* 1 Trouser\n",
+ "* 2 Pullover\n",
+ "* 3 Dress\n",
+ "* 4 Coat\n",
+ "* 5 Sandal\n",
+ "* 6 Shirt\n",
+ "* 7 Sneaker\n",
+ "* 8 Bag\n",
+ "* 9 Ankle boot"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "1408 neurons * .024 percent ~= 34 neurons"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 18,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "[array([[ 0.10132942, -0.09994747, 0.0544187 , ..., -0.01199088,\n",
+ " -0.06292131, 0.01987181],\n",
+ " [-0.03616156, -0.07451397, 0.01944981, ..., 0.01166514,\n",
+ " 0.02126038, -0.0748473 ],\n",
+ " [ 0.01550752, 0.0909953 , 0.04380812, ..., -0.07962617,\n",
+ " 0.01898236, 0.07236466],\n",
+ " ...,\n",
+ " [ 0.0108054 , 0.05861497, 0.17087871, ..., 0.03839699,\n",
+ " 0.09123042, -0.06567019],\n",
+ " [ 0.00497013, -0.08016764, -0.06656981, ..., -0.00515878,\n",
+ " 0.00221761, 0.05097136],\n",
+ " [-0.02290533, -0.03354137, 0.04545495, ..., 0.01482573,\n",
+ " 0.00098484, -0.04111394]], dtype=float32),\n",
+ " array([ 3.48618440e-02, -5.41638955e-03, 1.31103210e-02, 7.22463280e-02,\n",
+ " -2.43876502e-02, -1.06870309e-02, 2.91081183e-02, 2.90576406e-02,\n",
+ " 2.62515862e-02, 1.45024937e-02, -1.57963522e-02, 1.49091119e-02,\n",
+ " 8.93487409e-03, 4.12793346e-02, 2.95282118e-02, 5.19331312e-03,\n",
+ " 1.32530916e-03, -3.14759538e-02, 4.28747246e-03, 7.10726976e-02,\n",
+ " -2.24813707e-02, 2.92180516e-02, -5.23487013e-03, 8.60791206e-02,\n",
+ " -2.18736026e-02, 1.45187313e-02, -1.80636141e-02, 6.63637221e-02,\n",
+ " 1.71002783e-02, -1.99741162e-02, 6.77409815e-03, -2.29119249e-02,\n",
+ " -2.11065151e-02, -4.21732739e-02, 5.66550791e-02, -7.87500665e-03,\n",
+ " 4.96934354e-02, -2.62451507e-02, -3.14179659e-02, 6.30049482e-02,\n",
+ " 8.19519460e-02, 1.55785831e-03, 4.29074429e-02, 3.24573219e-02,\n",
+ " 1.79940443e-02, 1.36072130e-03, 5.86371124e-02, 3.61156873e-02,\n",
+ " -1.09881293e-02, 4.32750061e-02, -2.25960705e-02, -8.04545358e-03,\n",
+ " -1.53334010e-02, -3.28587070e-02, 3.13267857e-03, -1.86315607e-02,\n",
+ " -4.37643602e-02, 3.81392762e-02, 3.66224125e-02, 1.48072718e-02,\n",
+ " 3.44659165e-02, -1.13196466e-02, -2.27169134e-02, 8.48719776e-02,\n",
+ " 3.10068321e-03, 1.39202345e-02, 9.27911773e-02, 7.53409788e-02,\n",
+ " -2.16330979e-02, -2.20414903e-02, -3.15212552e-03, 1.59236938e-02,\n",
+ " -2.42054686e-02, -1.75593980e-02, -2.21934468e-02, -3.70963216e-02,\n",
+ " 6.72085881e-02, 3.21805626e-02, 2.90451963e-02, -1.29152099e-02,\n",
+ " -1.29320752e-02, 4.36900444e-02, -2.44683400e-02, 3.23325917e-02,\n",
+ " -1.76756456e-02, -5.36627956e-02, 4.35329936e-02, 2.40565222e-02,\n",
+ " 4.36999649e-02, -2.26819180e-02, -8.69834423e-03, 6.07701577e-02,\n",
+ " -2.18553673e-02, -3.17634009e-02, 2.49652676e-02, -4.14558090e-02,\n",
+ " 4.81878147e-02, -3.10613662e-02, 6.24261843e-03, 4.90439422e-02,\n",
+ " -1.63377635e-02, -9.88943875e-03, 5.74486889e-02, -3.52662057e-02,\n",
+ " -3.21774445e-02, 3.78112569e-02, -3.94279137e-02, -4.42866087e-02,\n",
+ " 9.20713786e-03, -2.36647036e-02, -1.01757804e-02, -5.47396392e-02,\n",
+ " -6.01586420e-03, 1.01622276e-01, -1.10045085e-02, -4.08708602e-02,\n",
+ " -4.97367457e-02, -7.52684176e-02, 6.09636456e-02, -1.71669777e-02,\n",
+ " 5.91361187e-02, -2.88355686e-02, 3.53775546e-03, -3.64997461e-02,\n",
+ " -2.26949845e-02, -1.02284960e-02, -1.37392487e-02, 7.15957507e-02,\n",
+ " -2.68650427e-02, 8.53226632e-02, 5.14466465e-02, 4.93848957e-02,\n",
+ " 8.59091058e-02, -2.62902882e-02, 1.00378595e-01, -1.65355932e-02,\n",
+ " -1.12617807e-02, -1.07267024e-02, -3.48392874e-02, -7.95916282e-03,\n",
+ " 1.00653268e-01, 5.97646572e-02, 7.75867561e-03, 1.15043588e-01,\n",
+ " 2.07628310e-02, -1.76335238e-02, 1.49275819e-02, 5.37896603e-02,\n",
+ " 7.00431988e-02, 8.98461267e-02, 9.10779536e-02, 5.60465502e-03,\n",
+ " 4.89570200e-02, 1.72273871e-02, 6.33122772e-02, 2.81463861e-02,\n",
+ " -1.87164899e-02, -4.43356372e-02, -1.96359158e-02, 6.37755021e-02,\n",
+ " 5.42709902e-02, -8.60076584e-03, -2.25404557e-02, 3.58947404e-02,\n",
+ " 5.20031601e-02, 1.08685963e-01, -1.22991828e-02, -1.04562687e-02,\n",
+ " 3.02043371e-02, 3.07208095e-02, 4.24020141e-02, -1.82511788e-02,\n",
+ " -9.11675568e-04, -3.51112783e-02, 6.27857372e-02, -2.22748592e-02,\n",
+ " 2.46507693e-02, 4.26435359e-02, 5.98271750e-03, -2.23289598e-02,\n",
+ " -2.76342370e-02, 2.70636491e-02, -1.39639527e-03, -3.58550772e-02,\n",
+ " 5.28606214e-02, 3.11384685e-02, 5.14203608e-02, 5.45845442e-02,\n",
+ " 1.19109713e-02, 8.52998719e-02, -4.34485935e-02, 1.28477793e-02,\n",
+ " -2.05492657e-02, -3.66462097e-02, 9.34293121e-03, 6.71103895e-02,\n",
+ " 6.07372411e-02, 2.44732462e-02, -1.44067034e-02, 7.23934174e-02,\n",
+ " -2.92091519e-02, 4.71341871e-02, 8.01225677e-02, -1.20827099e-02,\n",
+ " 4.47092671e-03, -1.86639261e-02, 1.82584543e-02, 4.94396463e-02,\n",
+ " 7.74530992e-02, -3.29862945e-02, -3.13315839e-02, -7.30312197e-03,\n",
+ " 7.01288553e-03, -2.70273201e-02, -1.52204782e-02, -2.32145302e-02,\n",
+ " -1.05082309e-02, -1.92536754e-05, -1.81905385e-02, -9.96621698e-03,\n",
+ " 1.09827740e-03, 8.68956521e-02, 1.66933890e-02, 6.86416961e-03,\n",
+ " -1.36355804e-02, 6.13272749e-03, -4.08143923e-03, 6.13516010e-03,\n",
+ " -4.06875797e-02, 1.20015517e-01, -1.68200638e-02, 5.30381538e-02,\n",
+ " -1.80107821e-02, -8.72184616e-03, -1.38499169e-02, 4.68643233e-02,\n",
+ " 9.19964090e-02, 1.01643559e-02, 5.90647422e-02, -4.65446375e-02,\n",
+ " 1.23580657e-02, -1.08336322e-02, -1.94771849e-02, 1.24923378e-01,\n",
+ " 4.38378043e-02, 4.30674516e-02, 4.88052331e-02, 1.19037889e-01,\n",
+ " 1.41229006e-02, 4.44606319e-02, -1.27055962e-02, 9.96087026e-03,\n",
+ " -4.08994928e-02, 1.04974672e-01, -2.27738936e-02, 2.94191055e-02,\n",
+ " -7.52164470e-03, 1.45411575e-02, -3.27894352e-02, -3.44525538e-02,\n",
+ " -1.79617628e-02, 7.10037127e-02, 2.80745570e-02, 4.35190760e-02,\n",
+ " -4.07501981e-02, -5.86949894e-03, 5.25610968e-02, -2.21862085e-02,\n",
+ " 4.24077436e-02, 5.35300560e-02, 1.01719853e-02, 5.97279631e-02,\n",
+ " 4.16470729e-02, -2.26495452e-02, 1.12023562e-01, 1.15930304e-01,\n",
+ " 6.27541617e-02, -1.05424039e-02, -3.84119377e-02, -4.32710610e-02,\n",
+ " 2.88346666e-03, 4.88664694e-02, 3.65142450e-02, 3.03427354e-02,\n",
+ " -2.83553582e-02, -3.89167946e-03, -6.82092551e-03, 6.12990037e-02,\n",
+ " -2.25962736e-02, 1.20927487e-02, -1.14525799e-02, 7.77647048e-02,\n",
+ " -2.94107408e-03, -7.81376287e-03, 2.62177419e-02, 4.43499200e-02,\n",
+ " 4.61728647e-02, -4.55445573e-02, 9.62291658e-02, 2.39728652e-02,\n",
+ " 7.95497745e-03, -5.06303972e-03, 1.64592341e-02, 4.28285971e-02,\n",
+ " 4.94796261e-02, 3.43723111e-02, 2.94336770e-02, 4.68909554e-02,\n",
+ " 7.18486868e-03, 2.99673546e-02, -3.39506529e-02, 4.21292381e-03,\n",
+ " 3.34585048e-02, 3.79088186e-02, 6.38425276e-02, -7.89680693e-04,\n",
+ " 2.56242584e-02, 3.94742750e-02, 1.40315751e-02, 4.60222661e-02,\n",
+ " 2.42276583e-02, 4.81235385e-02, 2.12334413e-02, 3.24134156e-02,\n",
+ " 5.22319693e-03, 1.51581809e-01, -2.08536666e-02, 5.09871505e-02,\n",
+ " -2.71328781e-02, 5.51932771e-03, 1.83137301e-02, 5.57270758e-02,\n",
+ " 4.80883941e-02, 6.50618374e-02, -1.34619158e-02, 1.22141302e-01,\n",
+ " -1.71008296e-02, 3.81552847e-03, 7.77109936e-02, 7.29411319e-02,\n",
+ " 3.79633270e-02, -6.00689724e-02, -9.96629149e-03, 2.48741861e-02,\n",
+ " -1.47205703e-02, -1.55409258e-02, 1.02969883e-02, 4.04569283e-02,\n",
+ " 3.13799903e-02, 1.16074421e-01, 3.65990065e-02, 3.31641100e-02,\n",
+ " -1.01007754e-06, 7.60934576e-02, -1.75704136e-02, 1.24804899e-02,\n",
+ " 1.18262857e-01, -3.16509157e-02, 1.05599836e-01, -3.54418345e-02,\n",
+ " 7.37690032e-02, -4.21324465e-03, -1.89842116e-02, -5.67948520e-02,\n",
+ " 4.46014712e-03, -2.98472904e-02, 6.52025342e-02, 9.10847858e-02,\n",
+ " 2.15716530e-02, -2.97974572e-02, 1.73324645e-02, 3.95461395e-02,\n",
+ " 3.12089212e-02, 4.31387834e-02, 6.34427816e-02, 3.88964042e-02,\n",
+ " 5.78227676e-02, 3.57266329e-02, 3.58240455e-02, -5.85102215e-02,\n",
+ " 6.58234283e-02, 4.79356572e-02, -3.79503556e-02, 9.29015651e-02,\n",
+ " 9.08859149e-02, 3.59453931e-02, -1.46479048e-02, 2.31869463e-02,\n",
+ " 3.59964892e-02, -2.23172419e-02, -1.54903410e-02, -1.83392055e-02,\n",
+ " 1.03566885e-01, 4.57644612e-02, -3.77948098e-02, 3.33484635e-02,\n",
+ " 1.28243286e-02, 2.67320629e-02, 8.60852096e-03, -4.81675379e-03,\n",
+ " 5.97773604e-02, 4.21931632e-02, -2.67935526e-02, -2.17138021e-03,\n",
+ " -2.12545926e-03, 3.86224315e-02, 8.93745758e-03, 1.20874859e-01,\n",
+ " -2.89826393e-02, -4.28759791e-02, 4.41658124e-02, 1.68996360e-02,\n",
+ " -3.31505165e-02, 5.13882004e-02, 5.88960387e-02, -1.76181737e-02,\n",
+ " 1.48313271e-03, 5.79988658e-02, 8.94222558e-02, -1.36769097e-02,\n",
+ " 9.34284274e-03, -1.84453707e-02, -1.26828887e-02, 8.16332735e-03,\n",
+ " 1.40198991e-02, -2.84167286e-02, 6.19943962e-02, -1.19751494e-03,\n",
+ " -3.54143754e-02, 2.86288578e-02, 7.49624521e-02, 2.02866849e-02,\n",
+ " 1.08794076e-02, 4.54496685e-03, 2.41601411e-02, 4.14725691e-02,\n",
+ " -2.01792410e-03, 7.59529248e-02, -2.71637514e-02, 5.82897440e-02,\n",
+ " 3.52679119e-02, -1.26215005e-02, -1.30356280e-02, 4.40822244e-02,\n",
+ " -1.54802725e-02, 1.93001963e-02, 2.34842468e-02, 1.62283089e-02,\n",
+ " 2.27487423e-02, 6.47596344e-02, 1.06462287e-02, -1.79562205e-03,\n",
+ " 2.33289460e-03, 4.68608066e-02, 6.58840733e-03, 1.15877278e-02,\n",
+ " -1.65016465e-02, -1.56839415e-02, 3.75386402e-02, 5.15719280e-02,\n",
+ " 4.46450040e-02, 9.76308808e-03, 3.82444868e-03, 5.87814413e-02,\n",
+ " 1.14726231e-01, 4.75812741e-02, 3.50318104e-02, 2.91183833e-02,\n",
+ " -1.01490710e-02, 2.14113072e-02, 6.38290169e-03, 5.89702800e-02,\n",
+ " 7.55651370e-02, -4.28581722e-02, 3.85477953e-02, -2.33158339e-02,\n",
+ " 8.83768313e-03, -1.96045693e-02, 5.51499948e-02, 2.92682145e-02,\n",
+ " -2.69281492e-02, -2.41599735e-02, 3.44376490e-02, 7.64978603e-02,\n",
+ " 5.74806742e-02, 3.19192410e-02, 4.75810543e-02, 2.57085152e-02,\n",
+ " 1.20992931e-02, -2.29395423e-02, 4.98514511e-02, 3.49956192e-02,\n",
+ " 4.24854718e-02, 3.13469879e-02, -3.78861034e-04, 6.45802449e-03,\n",
+ " 9.36235711e-02, 5.45675009e-02, 2.89321337e-02, 3.20924097e-03,\n",
+ " 1.14332768e-03, 7.16858879e-02, 6.88569993e-02, 2.04472039e-02,\n",
+ " -4.26226849e-04, -2.12573763e-02, -2.53489371e-02, -3.60967629e-02,\n",
+ " 4.41990867e-02, 1.84635643e-03, 2.98514590e-02, 1.83002884e-03],\n",
+ " dtype=float32)]"
+ ]
+ },
+ "execution_count": 18,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "# get second hidden layer (input not included in layer calculation)\n",
+ "\n",
+ "\n",
+ "weights = model.get_layer(index=1).get_weights()\n",
+ "\n",
+ "# Weights:\n",
+ "# First dimension is spec neuron\n",
+ "# Second dimension are the associated weights\n",
+ "\n",
+ "weights"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 19,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import random \n",
+ "# 34/512 neurons to pick\n",
+ "\n",
+ "specs = []\n",
+ "\n",
+ "for i in range(0, 34):\n",
+ " specs.append(random.randint(0,511))\n",
+ "\n",
+ "specs.sort()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 20,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "# Zero out specified weights\n",
+ "\n",
+ "for i in range(0,len(weights)): \n",
+ " for neuron_index in specs:\n",
+ " weights[i][neuron_index] = 0"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 21,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "model.get_layer(index=1).set_weights(weights)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "* Trained model\n",
+ "* Selected random neurons from second hidden layer (512 dense)\n",
+ "* Zeroed all of their weights (outside of model)\n",
+ "* Put that back into the model"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 22,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step\n"
+ ]
+ }
+ ],
+ "source": [
+ "y_test_pred = model.predict(x=X_test)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 23,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "y_test_pred_proper = []\n",
+ "\n",
+ "for i in y_test_pred:\n",
+ " y_test_pred_proper.append(i.argmax())"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 24,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "0.8813"
+ ]
+ },
+ "execution_count": 24,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "from sklearn.metrics import accuracy_score\n",
+ "\n",
+ "accuracy_score(y_true=trueY, y_pred=y_test_pred_proper)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 25,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "[[805 5 48 35 6 1 92 0 8 0]\n",
+ " [ 2 992 0 5 1 0 0 0 0 0]\n",
+ " [ 9 2 768 12 127 0 79 0 3 0]\n",
+ " [ 9 24 13 878 56 0 17 0 3 0]\n",
+ " [ 1 1 61 26 863 0 48 0 0 0]\n",
+ " [ 1 1 1 2 0 972 2 7 5 9]\n",
+ " [ 99 7 77 40 67 0 706 0 4 0]\n",
+ " [ 0 0 3 0 1 41 3 893 10 49]\n",
+ " [ 3 0 5 0 5 1 9 1 976 0]\n",
+ " [ 0 0 0 0 0 16 3 17 4 960]]\n"
+ ]
+ }
+ ],
+ "source": [
+ "from sklearn.metrics import confusion_matrix\n",
+ "\n",
+ "cm = confusion_matrix(y_true=trueY, y_pred=y_test_pred_proper)\n",
+ "print(cm)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 26,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhsAAAHHCAYAAAAWM5p0AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy80BEi2AAAACXBIWXMAAA9hAAAPYQGoP6dpAAChYElEQVR4nOzdd1hUxx7G8e8iHSkq3d57xS723rsxamLX2Huisbdgb4lGTYxYo1Fj74q9IhawFxQVpYmAgCBl7x9e16wUQXfZRX+f+5znhjln57wcdmWYmTNHoVQqlQghhBBCaImBrgMIIYQQ4ssmjQ0hhBBCaJU0NoQQQgihVdLYEEIIIYRWSWNDCCGEEFoljQ0hhBBCaJU0NoQQQgihVdLYEEIIIYRWSWNDCCGEEFoljQ0htOjevXs0atQIa2trFAoFO3bs0Gj9jx49QqFQ4O7urtF6M7M6depQp04dXccQQvyHNDbEF+/Bgwf079+fAgUKYGpqipWVFTVq1GDx4sW8fv1aq+fu3r07Pj4+zJw5k3Xr1lGxYkWtni8j9ejRA4VCgZWVVbLX8d69eygUChQKBfPmzUt3/c+ePWPKlClcvXpVA2mFELpkqOsAQmjT3r176dixIyYmJnz//feUKlWKN2/ecPr0acaMGcONGzdYuXKlVs79+vVrzp07x/jx4xk8eLBWzpE3b15ev36NkZGRVur/GENDQ6Kjo9m9ezedOnVS27dhwwZMTU2JiYn5pLqfPXvG1KlTyZcvH+XKlUvz6w4dOvRJ5xNCaI80NsQX6+HDh3Tu3Jm8efPi4eGBk5OTat+gQYO4f/8+e/fu1dr5g4ODAbCxsdHaORQKBaamplqr/2NMTEyoUaMGf//9d5LGxsaNG2nevDnbtm3LkCzR0dGYm5tjbGycIecTQqSdDKOIL9acOXOIjIxk1apVag2NdwoVKsSwYcNUX8fHxzN9+nQKFiyIiYkJ+fLl4+effyY2Nlbtdfny5aNFixacPn2aypUrY2pqSoECBVi7dq3qmClTppA3b14AxowZg0KhIF++fMDb4Yd3//1fU6ZMQaFQqJUdPnwYV1dXbGxsyJo1K0WLFuXnn39W7U9pzoaHhwc1a9bEwsICGxsbWrduza1bt5I93/379+nRowc2NjZYW1vTs2dPoqOjU76wH+jSpQv79+8nLCxMVebp6cm9e/fo0qVLkuNDQ0MZPXo0pUuXJmvWrFhZWdG0aVOuXbumOub48eNUqlQJgJ49e6qGY959n3Xq1KFUqVJ4eXlRq1YtzM3NVdflwzkb3bt3x9TUNMn337hxY7Jly8azZ8/S/L0KIT6NNDbEF2v37t0UKFCA6tWrp+n4Pn36MGnSJCpUqMDChQupXbs2bm5udO7cOcmx9+/fp0OHDjRs2JD58+eTLVs2evTowY0bNwBo164dCxcuBODbb79l3bp1LFq0KF35b9y4QYsWLYiNjWXatGnMnz+fVq1acebMmVRfd+TIERo3bkxQUBBTpkxh5MiRnD17lho1avDo0aMkx3fq1IlXr17h5uZGp06dcHd3Z+rUqWnO2a5dOxQKBf/++6+qbOPGjRQrVowKFSokOd7X15cdO3bQokULFixYwJgxY/Dx8aF27dqqX/zFixdn2rRpAPTr149169axbt06atWqparnxYsXNG3alHLlyrFo0SLq1q2bbL7FixdjZ2dH9+7dSUhIAGDFihUcOnSIX3/9FWdn5zR/r0KIT6QU4gsUHh6uBJStW7dO0/FXr15VAso+ffqolY8ePVoJKD08PFRlefPmVQLKkydPqsqCgoKUJiYmylGjRqnKHj58qASUc+fOVauze/fuyrx58ybJMHnyZOV/P5ILFy5UAsrg4OAUc787x+rVq1Vl5cqVU9rb2ytfvHihKrt27ZrSwMBA+f333yc5X69evdTqbNu2rTJHjhwpnvO/34eFhYVSqVQqO3TooKxfv75SqVQqExISlI6OjsqpU6cmew1iYmKUCQkJSb4PExMT5bRp01Rlnp6eSb63d2rXrq0ElMuXL092X+3atdXKDh48qASUM2bMUPr6+iqzZs2qbNOmzUe/RyGEZkjPhvgiRUREAGBpaZmm4/ft2wfAyJEj1cpHjRoFkGRuR4kSJahZs6bqazs7O4oWLYqvr+8nZ/7Qu7keO3fuJDExMU2vef78OVevXqVHjx5kz55dVV6mTBkaNmyo+j7/64cfflD7umbNmrx48UJ1DdOiS5cuHD9+nICAADw8PAgICEh2CAXezvMwMHj7T09CQgIvXrxQDRFdvnw5zec0MTGhZ8+eaTq2UaNG9O/fn2nTptGuXTtMTU1ZsWJFms8lhPg80tgQXyQrKysAXr16labj/fz8MDAwoFChQmrljo6O2NjY4Ofnp1aeJ0+eJHVky5aNly9ffmLipL755htq1KhBnz59cHBwoHPnzvzzzz+pNjze5SxatGiSfcWLFyckJISoqCi18g+/l2zZsgGk63tp1qwZlpaWbN68mQ0bNlCpUqUk1/KdxMREFi5cSOHChTExMcHW1hY7Ozu8vb0JDw9P8zlz5syZrsmg8+bNI3v27Fy9epUlS5Zgb2+f5tcKIT6PNDbEF8nKygpnZ2euX7+ertd9OEEzJVmyZEm2XKlUfvI53s0neMfMzIyTJ09y5MgRvvvuO7y9vfnmm29o2LBhkmM/x+d8L++YmJjQrl071qxZw/bt21Ps1QD45ZdfGDlyJLVq1WL9+vUcPHiQw4cPU7JkyTT34MDb65MeV65cISgoCAAfH590vVYI8XmksSG+WC1atODBgwecO3fuo8fmzZuXxMRE7t27p1YeGBhIWFiY6s4STciWLZvanRvvfNh7AmBgYED9+vVZsGABN2/eZObMmXh4eHDs2LFk636X886dO0n23b59G1tbWywsLD7vG0hBly5duHLlCq9evUp2Uu07W7dupW7duqxatYrOnTvTqFEjGjRokOSapLXhlxZRUVH07NmTEiVK0K9fP+bMmYOnp6fG6hdCpE4aG+KL9eOPP2JhYUGfPn0IDAxMsv/BgwcsXrwYeDsMACS5Y2TBggUANG/eXGO5ChYsSHh4ON7e3qqy58+fs337drXjQkNDk7z23eJWH96O+46TkxPlypVjzZo1ar+8r1+/zqFDh1TfpzbUrVuX6dOn89tvv+Ho6JjicVmyZEnSa7Jlyxb8/f3Vyt41ipJrmKXXTz/9xOPHj1mzZg0LFiwgX758dO/ePcXrKITQLFnUS3yxChYsyMaNG/nmm28oXry42gqiZ8+eZcuWLfTo0QOAsmXL0r17d1auXElYWBi1a9fm4sWLrFmzhjZt2qR4W+Wn6Ny5Mz/99BNt27Zl6NChREdH8/vvv1OkSBG1CZLTpk3j5MmTNG/enLx58xIUFMSyZcvIlSsXrq6uKdY/d+5cmjZtSrVq1ejduzevX7/m119/xdramilTpmjs+/iQgYEBEyZM+OhxLVq0YNq0afTs2ZPq1avj4+PDhg0bKFCggNpxBQsWxMbGhuXLl2NpaYmFhQVVqlQhf/786crl4eHBsmXLmDx5supW3NWrV1OnTh0mTpzInDlz0lWfEOIT6PhuGCG07u7du8q+ffsq8+XLpzQ2NlZaWloqa9Soofz111+VMTExquPi4uKUU6dOVebPn19pZGSkzJ07t3LcuHFqxyiVb299bd68eZLzfHjLZUq3viqVSuWhQ4eUpUqVUhobGyuLFi2qXL9+fZJbX48ePaps3bq10tnZWWlsbKx0dnZWfvvtt8q7d+8mOceHt4ceOXJEWaNGDaWZmZnSyspK2bJlS+XNmzfVjnl3vg9vrV29erUSUD58+DDFa6pUqt/6mpKUbn0dNWqU0snJSWlmZqasUaOG8ty5c8nesrpz505liRIllIaGhmrfZ+3atZUlS5ZM9pz/rSciIkKZN29eZYUKFZRxcXFqx40YMUJpYGCgPHfuXKrfgxDi8ymUynTMAhNCCCGESCeZsyGEEEIIrZLGhhBCCCG0ShobQgghhNAqaWwIIYQQQquksSGEEEIIrZLGhhBCCCG0ShobQgghhNCqL3IF0ddrx+k6Qqqs+qzVdYRUWZtq59kZmhCXqLkHkGlD9JsYXUdIlT4vqpPXykHXEVLlF5F0yXt9YqDBZ8loWqKeL+cU/8b/4wd9prgQX43UY2Rb4OMH6SHp2RBCCCGEVn2RPRtCCCGEXtHzXlltk8aGEEIIoW3KRF0n0ClpbAghhBDalvh1NzZkzoYQQgghtEp6NoQQQggtU8owihBCCCG0SoZRhBBCCCG0R3o2hBBCCG2TYRQhhBBCaNVXvs6GDKMIIYQQQqukZ0MIIYTQtq98GOWr7NlISFSy9PgNmv12gCqzd9Bi6UFWnrqF8j8PC1IqlSw7cZMGi/ZSZfYO+m84hV9opFo9TX87QLmZ/6ptf529o/X8EyeOJO6Nv9rm43NC6+dNi6Ej+hEScZcZs35Wldnb27Js5Vxu3DuD3/OreJzcTotWjTIkT+8+XThzfi9Pnl3lybOrHD66hQYNa6v279m/gfDIB2rbwsXTMyRbSpydHVnjvoSA59eJCL/PlctHcKlQRqeZ/qumaxV2bHfn8SMv4t/406pVY51lschqzoQZozl5ZS83npxly77VlC5fAgBDQ0N+nDSUfSc34+N3hrPXDzJv6TTsHW11lhdgwA/duX/3PJERDzh7ejeVKpbTaZ53DAwMmDJ5NHfunCU87D63bp3m53HDdB0rCX29fh+VmKiZLZP6Kns2Vp+7w5bLD5nW0oWCdlbcfB7G5D1eZDU1okulQgC4n7vLRs8HTG/pQk4bC5aduMnAv0/zb/+GmBhmUdU1sFZx2pXPr/rawjhjLun1G7dp0qSz6uv4+PgMOW9qylcoTfee33Dd57Za+dKVc7C2tqJb5wGEvnhJ+44tWLVmMQ1qt8PH+5ZWM/n7BzBl0lwePHiEQgFdurbn783LqVmjFbdv3QPAffUmZk5fqHrN69e6e3KrjY01J47v4MSJs7Rs2Y3gkBcUKpSfl2HhOsv0IQsLc7y9b7LafRPbtqzSaRa3RZMoXKwgowZOJCggmNYdm7Fu2+80rt6BqKjXlCxTjN/m/8mtG3extrZi4i+jWbl+EW0adNNJ3o4dWzFv7mQGDhrLRc8rDB3Sh317N1CiVC2Cg1/oJNM7Y0YPpF+/7+ndZzg3b97FpUJZ/vhjPuERr1i69C+dZntHn6+fSN1X2bNx7WkodYo4UauwEzltLGhYPCfV8ttz/dlL4G2vxoaL9+nrWpS6RZ0p4mDN9FYVCX4Vw7E7z9TqMjcxwjarqWozy6DGRkJ8AoGBwartxYuXGXLelFhYmLP8z3mMGDqR8A9+MVaqXJ4/Vqzjipc3fo+esGDu74SHR1C2XCmt5zqw34PDh47j++ARD+4/YvrU+URFRlOpUjnVMdHRrwkKClFtr15Fplyhlo0ZM5CnT5/Rp+9IPC9d5dGjJxw5chJfXz+dZfrQgYPHmDR5Djt3HtBpDhNTExq3qMfsqYvxPHcZv4dPWDJnBX4Pn9K1Z0ciX0XSvcNA9u08zMP7flz18mHK2NmULlcCp5yOOsk8Ylhf/ly1kTVr/+HWrXsMHDSW6OjX9OzR+eMv1rKq1Sqye/ch9u/3wM/vKf9u38uRIyf1qudAn6/fxyiViRrZMiudNjZCQkKYM2cObdu2pVq1alSrVo22bdsyd+5cgoODtXbesrmyc+FRMH4vXgFwJzCMK09fUKOgAwD+YdGERMVSJZ+96jWWpkaUzpmda/6hanWtPnuH2gv28M2fR3E/d5f4DOrmKlQoP36PvLhz+yxr1/xK7tzOGXLelMyeP5nDB49z8vjZJPs8L16hbbtm2GSzRqFQ0LZ9c0xMTDhz+kKGZjQwMKB9hxaYW5hx8eIVVXmnb1rh6+fJuYv7mTxlNGZmphma679atGiEl5c3f/+9Av+n1/C8eJDevbroLI8+MzTMgqGhIW9i3qiVx7yOwaVquWRfY2mZlcTERF6Fv8qAhOqMjIyoUKEMRz1OqcqUSiVHPU5TtapLhuf50Plzl6hbtwaFC7/tqS1TujjVq1fi4MFjOk72lr5fv4+SYRTd8PT0pHHjxpibm9OgQQOKFCkCQGBgIEuWLGHWrFkcPHiQihUravzcvaoXJSo2njbLD5PFQEFCopLBdUrSvFQeAEKi3naj57AwUXtddgsTXkS+72LvUqkgxRxtsDY15trTFyw5foOQyBhGN9Tu+PrFi1fo3WcEd+8+wNHRnokTRnLMYzvlytcjMjJKq+dOTtv2zSlTtgQN67RPdn/v7sP4030R9/08iYuL43V0DN27DuKh7+MMyVeiZBEOH92KqakJkZHRdP12IHdu3wdg6z+7efLYn+cBgZQsWYyp03+kcJECdOsyMEOyfahA/jz07/8dixb/wezZS6joUo6FC6fxJi6Odeu26CSTvoqKjObyxWsMGt2H+/d8CQkKpWX7JpSvVAa/h0+SHG9sYsxPk4ex+98DOvmc2Npmx9DQkKDAELXyoKBgihUtmOF5PjRn7lKsrCzx8T5BQkICWbJkYdKk2fy9abuuowH6f/0+KhP3SmiCzhobQ4YMoWPHjixfvhyFQqG2T6lU8sMPPzBkyBDOnTuXaj2xsbHExsaqlSXGxWNilPK3dujmU/Zdf4Jbm0oUtLPiTmA4cw97Y2dpSqsyedP8PXxXpbDqv4s4WGOUxYAZ+68wtG5JjP8zr0PT/vuXho/PLS5evMKD+xfo2KElq903ae28yXHO6cjM2ePp0LonsbFvkj1m3IThWFtb0bZld0JfvKRZiwascl9MiyZduHXzrtYz3rv7kJrVW2JlZUnrNk1YvnIOzZp04c7t+7ivfn+9bt64S2BgMLv3rid//jw8fJgxjaH/MjAwwMvLm4kTZwFw9eoNSpYsSr++30ljIxmjBk5k1pLJnLt+iPj4eG5432b3vwcpVba42nGGhob8umo2KGDSaDcdpdVvHTu0pHPntnz//WBu3rxL2bIlmTdvCs+fB7Ju/VZdxxOZnM4aG9euXcPd3T1JQwNAoVAwYsQIypcv/9F63NzcmDp1qlrZz21qMKFtzRRfs/DodXpWL0KTkrkBKGxvzfPwaP46e4dWZfJia/G2G/1FVCx2lmaq14VGxVLEwTrFekvlzE58opJn4dHky2H50eyaEh4ewb17vhQslC/DzvlO2XKlsLe3xePU+79+DA0NqVajEn36daOqS2P69v+OGpWbqXoTbly/TdVqFendtyujR0zWesa4uDjVnIerV69TwaUMAwb2YPjQCUmOveR5FYACBfLqpLHx/HkQt26pN8Bu375P27bNMjxLZvD40VO6tOqLmbkpWS2zEhwYwpI/Z/HE76nqmLcNjVnkzOVEt7b9ddKrARASEkp8fDz2Dup3w9jb2xEQqL1h47Ryc5vA3HlL+WfLLuDtJPQ8eXLy44+D9aKxoe/X76NkUS/dcHR05OLFiynuv3jxIg4ODh+tZ9y4cYSHh6ttY1pUS/U1MfEJGHzQyDFQKEj8/52vOW3MsbUw4eKj92/gyNg4fPxDKZsze4r13gkMw0AB2c1NUjxGGywszClQIC8Bz4My9LwAp06cw7VKc+rUaK3arlz2Yes/u6lTozVmZm8ba4mJSrXXJSQmYGCgm7efgYEBxsbGye4rXebtbZMBARl/LQHOnvOkSBH1LuHChQvw+LG/TvJkFq+jYwgODMHK2pKadatxZP/bW8HfNTTyFcjD9+1/IOyl7u7qiYuL4/Jlb+rVdVWVKRQK6tV15fx5L53lesfc3IzED+YEJCTo7nP6IX2/fh+lTNTMlknprGdj9OjR9OvXDy8vL+rXr69qWAQGBnL06FH++OMP5s2b99F6TExMMDFR/+X+OpUhFIBahR3588xtHK3M3g6jBISx/uI9WpfNB7x9A3etXIg/ztwmT3YLctpYsPTETewsTalb9O1EzGtPX+Dz7CWV8tpiYWzENf8XzDvsQ7NSebAyS/4XmabMnjWRPXsP8/jxU5ydHJk0aRQJCYls2rxDq+dNTmRklOoW0neio6IJDX3J7Vv3MDQ0xPfBIxYsnsakCbN5GfqSZs0bUqduDbp06q/1fJOnjObw4RM8ffKMrJYWdOzYCteaVWjXugf58+ehQ6dWHD54nNDQl5QsVQy3WeM5ffoCN25of72U5CxZ/AcnT+7kp5+GsHXrbipVKkefPl0ZMPBHneRJjoWFOYUKvb/dO3++PJQtW5LQ0Jc8efIslVdqXs261VAoFPjef0Te/LkZO2U4D+49YuvGXRgaGvLb6jmUKlOMPl2GYZAlC7b2OQAIfxlOXFzG3y6+cPEfrF61EK/L3nh6XmHokL5YWJjhvmZzhmf50N69hxn701CePPHn5s27lCtbimHD+rFGD7K9o8/XT6ROZ42NQYMGYWtry8KFC1m2bBkJCW+7mLJkyYKLiwvu7u506tRJK+ce26gsS0/cxO3AVUKjY7HLakb78vnpX/P9OG+PakV4HZfA9H1XeBUTR/ncOVjWuYZqjQ3jLAYcvPGE5SdvEZeQQE4bC7pVLsR3VQppJfN/5czlxPp1S8mRIxvBwaGcOXsR15otCQkJ/fiLM1h8fDydO/Rl4pTRbNi8HAsLcx76PmbQDz9x5JD2FyKzs8vB8pXzcHS0IyIikhvXb9OudQ+OHTtDzpxO1KlbnYEDe2BuYY7/0+fs2nmQuXOWaj1XSi55XaNDxz7MnDGWCeOH8/DRE0aNmszff+vHJD2Aii5lOXrkfbf6/HlTAFiz9h969xmRoVksrbIyesJgHJ0dCA8L58BuD+bPXEp8fDw5czvRsGkdAPaeUP9l1KV1Xy6cyfi/hrds2YWdbXamTBqNo6Md167doHmLbgQFhXz8xVo2fMREpkwZw5LFv2Bvb8uz5wH8+ed6ZsxcpOtoKvp8/T4qE99JogkK5X+XzdSRuLg4QkLevllsbW0xMjL6rPperx2niVhaY9Vnra4jpMra1ELXEVIUp+fjntFvdLcgWFro/MOeirxWHx821SW/iEBdR0jVh0PD+iRR979mUhX/RvvDlLHXD2ukHpNSDTVST0bTixVEjYyMcHJy0nUMIYQQQmiBXjQ2hBBCiC/aVz6MIo0NIYQQQsuUSv0eAtY2/binSQghhBBfLOnZEEIIIbQtE6+RoQnS2BBCCCG0TeZsCCGEEEKrvvKeDZmzIYQQQgitkp4NIYQQQtv0fEFCbZPGhhBCCKFtMowihBBCCKE90rMhhBBCaJvcjSKEEEIIrfrKh1G+yMaGpZ4/VfX1s1O6jpAqM+eauo6QIv19ruVb+v1sS/2m709V1Xf6/mRV8XX7IhsbQgghhF6RYRQhhBBCaNVX3tiQu1GEEEIIoVXSsyGEEEJo2df+iHlpbAghhBDa9pUPo0hjQwghhNC2r/zWV5mzIYQQQgitkp4NIYQQQttkGEUIIYQQWiXDKEIIIYQQ2iM9G0IIIYS2yTCKEEIIIbRKhlFEcn76cTDnzu7l5Ys7PHt6jW1bV1GkSMEMOXdUVDSzFi2nYbvuuNRtTdf+I/G5dUe1PyT0JeNnzKduq65UrNeG/iMn4PfEX7U/POIVvyxYRovOfXCp25oG7b7nl4W/8yoyKkPyvzPgh+7cv3ueyIgHnD29m0oVy2Xo+VMyceJI4t74q20+Pid0HUulpmsVdmx35/EjL+Lf+NOqVWNdR0pCX3+27+hzPn3OBpJPaIc0NlJQq2ZVfv99DTVqtqRJs28xMjRi/96NmJubaf3ck2Yt5pznFdwmjWb7ut+pXrkCfYf9TGBwCEqlkmFjp/H0WQBLZk9iy+rfcHa0p8+wn4l+HQNAUMgLgkJCGT24D9vX/c7M8SM5c8GLSW4LtZ79nY4dWzFv7mSmz1hApSpNuOZ9k317N2BnlyPDMqTm+o3b5MpdTrXVqdNG15FULCzM8fa+yZBh43UdJVn6/rPV53z6nA0kn1YlJmpmy6QUSuWX91xiQ+OcGq/T1jY7Ac98qFuvHadOX/isulJ7xHxMbCxVGrZjyazJ1K5eWVXeqdcQXKtWpFWT+rT4ti871i2nUIG8ACQmJlKnZReG9u9Bh1ZNkq33oMcpxk6bg+eRHRgaZkk1nyYeMX/29G48L11j2PAJACgUCh75erJ02WrmzF36yfVq4hHzEyeOpHWrJlSs1EgDtanT9Icp/o0/7Tr0Yteugxqu+dNp62erKfqcT5+zwdebL/6N/8cP+kyv9y7SSD1mzYdrpJ6MJj0baWRtbQVA6MswrZ4nIT6BhIRETIyN1MpNTIy57H2DN3FxABj/Z7+BgQFGxkZc8b6RYr2vIqPIamH+0YaGJhgZGVGhQhmOerxvVCmVSo56nKZqVRetnz8tChXKj98jL+7cPsvaNb+SO7ezriNlCvr+s9XnfPqcDSSf0C69bmw8efKEXr166ToGCoWCBfOmcubMRW7cuPPxF3wGCwtzypYqznL3vwkKfkFCQgK7D3pw7fptQkJCyZ83N04O9ixe4U54xCvi4uJYtf4fAoNCCH4RmmydL8PCWeH+Nx1aNdVq9ndsbbNjaGhIUGCIWnlQUDCODnYZkiE1Fy9eoXefEbRo2Y3BQ8aRL18ejnlsJ2tWC11H03v6/rPV53z6nA0kn9YpEzWzZVJ63dgIDQ1lzZo1qR4TGxtLRESE2qbpkaFfl/xCyZJF6dJtoEbrTYnbxNGgVFKvTTcq1G3Fhi07adqgNgoDA4wMDVn0ywQePfanRtNOVKzfhouXvalZtSIGBkl/nJFRUQwcM5mC+fMwsHe3DMmv7w4ePMa2bXvw8bnF4cMnaNnqO2xsrOjYoaWuowkhvlRf+ZwNnd76umvXrlT3+/r6frQONzc3pk6dqlamMMiKIovVZ2V7Z/GiGTRv1oC69dvh7/9cI3V+TJ5czrgvnUv06xiioqKxs83OqIlu5HJ2BKBkscJsW7OUV5FRxMXFkT2bDd/2HU7JYoXV6omKiqb/yIlYmJux+JeJGBlmzI87JCSU+Ph47B1s1crt7e0ICAzOkAzpER4ewb17vhQslE/XUfSevv9s9TmfPmcDyad1mbhXQhN02rPRpk0b2rZtS5s2bZLdRo4c+dE6xo0bR3h4uNqmMLDUSL7Fi2bQpnUTGjbuxKNHTzRSZ3qYm5liZ5ud8IhXnL3oRb2aVdX2W2a1IHs2G/ye+HPj9j3qur7fHxkVRb8R4zEyMuTX2ZMxMTHOsNxxcXFcvuxNvbquqjKFQkG9uq6cP++VYTnSysLCnAIF8hLwPEjXUfSevv9s9TmfPmcDySe0S6c9G05OTixbtozWrVsnu//q1au4uKQ+8cfExAQTExO1MoXi8+9Z+HXJL3zbuQ3t2vfi1atIHP4/Jhge/oqYmJjPrj81Zy54oVQqyZcnF4+fPmP+0lXkz5OLNs3f3j1x0OMU2WyscXKw457vI2YtWk69mtWoUeXttYqMiqLf8PG8jo1l8aQxREVFExUVDUA2G2uyZNH+JNGFi/9g9aqFeF32xtPzCkOH9MXCwgz3NZu1fu6PmT1rInv2Hubx46c4OzkyadIoEhIS2bR5h66jAW8bP4UK5Vd9nT9fHsqWLUlo6EuePHmmw2Rv6fPPFvQ7nz5nA8mnVZl4CEQTdNrYcHFxwcvLK8XGhkKh0Pj8i7Qa8EN3ADyOblMr79V7BGvX/aPVc7+KjGLR8tUEBodgbWVJw9quDO3fXTUMEvwilDm/ruRFaBh2ObLTqkl9fuj5rer1N+88wPvm24mszb7prVb3wa3u5HRy0Gp+gC1bdmFnm50pk0bj6GjHtWs3aN6iG0FBIR9/sZblzOXE+nVLyZEjG8HBoZw5exHXmi0JCUl+gm1Gq+hSlqNHtqq+nj9vCgBr1v5D7z4jdJTqPX3+2YJ+59PnbCD5tOorH0bR6Tobp06dIioqiiZNkl8bIioqikuXLlG7du101auNdTY0KbV1NvSBJtbZ0BZNrLOhTV/cojVCfAUyZJ2Nf3/RSD1m7X7WSD0ZTac9GzVrpv5LzcLCIt0NDSGEEELvyDCKEEIIIbTqK29s6PU6G0IIIYTI/KRnQwghhNC2L+8xZOkijQ0hhBBC22QYRQghhBBCe6SxIYQQQmibDp6NkpCQwMSJE8mfPz9mZmYULFiQ6dOnq61fpVQqmTRpEk5OTpiZmdGgQQPu3bunVk9oaChdu3bFysoKGxsbevfuTWRkZLqySGNDCCGE0DYdPPV19uzZ/P777/z222/cunWL2bNnM2fOHH799VfVMXPmzGHJkiUsX76cCxcuYGFhQePGjdVWyu7atSs3btzg8OHD7Nmzh5MnT9KvX790ZdHpol7aIot6fR5Z1OvTfXEfJiG+AhmyqNfacRqpx+x7tzQf26JFCxwcHFi1apWqrH379piZmbF+/XqUSiXOzs6MGjWK0aNHAxAeHo6DgwPu7u507tyZW7duUaJECTw9PalYsSIABw4coFmzZjx9+hRnZ+c0ZZGeDSGEEOILVL16dY4ePcrdu3cBuHbtGqdPn6Zp06YAPHz4kICAABo0aKB6jbW1NVWqVOHcuXMAnDt3DhsbG1VDA6BBgwYYGBhw4cKFNGeRu1GEEEIIbdPQIEJsbCyxsbFqZck9kBRg7NixREREUKxYMbJkyUJCQgIzZ86ka9euAAQEBADg4KD+vCwHBwfVvoCAAOzt7dX2Gxoakj17dtUxaSE9G0IIIYS2aWiCqJubG9bW1mqbm1vyQyv//PMPGzZsYOPGjVy+fJk1a9Ywb9481qxZk8HfvPRsCCGEEJnGuHHjGDlypFpZcr0aAGPGjGHs2LF07twZgNKlS+Pn54ebmxvdu3fH0dERgMDAQJycnFSvCwwMpFy5cgA4OjoSFBSkVm98fDyhoaGq16fFF9nYMFDo9zRCfZ6ACRAxv7WuI6Qo25jduo6QqhI2eXQdIVU+oY90HSFFOS1z6DpCqvxfvdB1BJGZaWhRr5SGTJITHR2NgYH6AEaWLFlI/H+W/Pnz4+joyNGjR1WNi4iICC5cuMCAAQMAqFatGmFhYXh5eeHi4gKAh4cHiYmJVKlSJc25v8jGhhBCCKFX0nnbqia0bNmSmTNnkidPHkqWLMmVK1dYsGABvXr1AkChUDB8+HBmzJhB4cKFyZ8/PxMnTsTZ2Zk2bdoAULx4cZo0aULfvn1Zvnw5cXFxDB48mM6dO6f5ThSQxoYQQgjxRfr111+ZOHEiAwcOJCgoCGdnZ/r378+kSZNUx/z4449ERUXRr18/wsLCcHV15cCBA5iamqqO2bBhA4MHD6Z+/foYGBjQvn17lixZkq4sX+Q6G8YmuXQdIVWJen7JZRjl08kwyqeTYRShKxmxzkb0yhEaqce830KN1JPRpGdDCCGE0DZ5EJsQQgghhPZIz4YQQgihbTqYIKpPpLEhhBBCaFuifs/V0zZpbAghhBDaJnM2hBBCCCG0R3o2hBBCCG37yns2pLEhhBBCaJuer6+kbTKMIoQQQgitksZGKrJmtWDevCncu3ue8LD7nDi+AxeXsrqOxU8/Dubc2b28fHGHZ0+vsW3rKooUKZgh52625gzlfzuaZHM7cVt1zLXn4fTbfplqy4/huuI4vf71IiY+QbXf72U0w/deo+6fJ3FdcZye2y7h+TRUa5ldXavw77a/eOh7idiYJ7Rq2Vi1z9DQkJkzxuF16TChL+7w0PcSq1YtxMnJQStZKlQty6K1szl0dSdXAs5Qp8n7h/IZGmZh6IQB/HNsLWd9j3Do6k6m/zoBOwdb1TEu1ctzJeBMsluJcsW0kjk5A37ozv2754mMeMDZ07upVLFchp37v05f2Y/fC+8k2/Q5PwOQJ18uVqxdyOU7x7n+6CxLV83F1i67TrK+oy/X7kP9+33PZa/DhIbcJjTkNqdP7qJJ47q6jpWEvl6/j9LQI+YzK2lspGLF8rk0qF+Tnr2GUcGlAUeOnOTA/r9xdk77Y3W1oVbNqvz++xpq1GxJk2bfYmRoxP69GzE3N9P6udd3qsThnq6q7ffW5QFoWPDtL+drz8MZvPsKVfNkZ33HSqzvVInOpXOpPYl36J6rJCQqWdGmPBu+qUwR26wM3XONkKhYrWS2MDfD2+cWw4ZPSLLP3NyM8uVL8YvbYqpWbco3nftSpHBBtm39SytZzMzNuHvjPm7j5ifZZ2pmSvHSRfljoTvfNuzFqF4/k7dgHhatna065pqnDw1Kt1Tb/l2/i6d+/ty8ejtJndrQsWMr5s2dzPQZC6hUpQnXvG+yb+8G7OwyfrnxVg26ULF4XdXWpV1fAPbuPISZuRnrt64ApZJv2/SlfdPuGBkbsWrjryh09GRofbp2H/L3f8748W5UrtqUKtWacez4Gf7d9hclShTRdTQVfb5+H5Wo1MyWScmzUVJgampK6IvbtO/Qi/37PVTl58/t4+DBY0yeMveT69b0s1FsbbMT8MyHuvXacer0hc+uLz3PRpl76i6nHoWws1s1FAoF32/xpEru7AyqmnxPy8vXb6i36hSr2lWggnM2AKLexOO68gS/ty5P1dyp/9X5uc9GiY15QseOfdi1+2CKx7i4lOXsmT0UKlyFJ0+epav+9Dwb5UrAGUb0GMvxA6dSrq9cMTYcWEVTl3YE+Acm2W9omIWDV3eyadVW/ljo/tFzauLZKGdP78bz0jVV402hUPDI15Oly1YzZ+7ST65XE89GmTTzR+o3qkXtSi2oWacaa/5ZRpmCrkS+igLA0jIr3r6n6dahP2dOpO+zoolno2jr2mlLUMB1fho7g9Xum3QdBdDe9cuQZ6PM66OResxH/6mRejKa9GykwNAwC4aGhsTEqP+1/fp1DNWrV9ZRquRZW1sBEPoyLEPPG5eQyL47AbQu7oxCoSA0+g0+gRFkNzOm+9ZL1F91kt7/enHl2ftcNqZG5LMxZ8/tAF7HJRCfmMi26/5kNzOihJ1lhuZPibW1JYmJiYSFReg6CpaWWUlMTORV+Ktk99duXBPrbFbs3LQ3Q/IYGRlRoUIZjnq8byAplUqOepymalWXDMmQEiMjQ9p2bM4/G3cAYGxijFKp5E3sG9UxsbGxJCYmUqlKBR3k099r9yEDAwM6dWqFhYU55y946ToOkLmuX7KUiZrZMimdNzZev37N6dOnuXnzZpJ9MTExrF27VgepIDIyinPnLvHzuOE4OTlgYGBAl2/bUbWqC05O9jrJlByFQsGCeVM5c+YiN27cydBzH/MN5lVsPC2LOQHwNOI1ACsu+tKuhDNLW5WnuJ0l/Xdcxi8sWpV3eZvy3A5+RY0Vx6n6+3HWXX3M0lblsTI1ytD8yTExMWHmjHFs/mcnr15F6jSLsYkxQycM4MD2I0RFRid7TJsuLTh3/CJBz4MzJJOtbXYMDQ0JCgxRKw8KCsbRwS5DMqSkUbN6WFlbsuXvnQBcueRNdPRrxk4egamZKWbmZoyfNgpDQ0Ps/zMPJqPo87V7p1SpYoSF3iU68iHLfptFh459uHXrnq5jAZnj+qXqKx9G0Wlj4+7duxQvXpxatWpRunRpateuzfPnz1X7w8PD6dmzZ6p1xMbGEhERobZpamSoZ69hKBQK/B55EfnKl0GDerF5804S9WiSzq9LfqFkyaJ06TYww8+94+YzauTNgX1WE+D98FD7UjlpXcKZYnaWjK5ZhHzZLNh58+1whFKpxO3EHbKbG/NXexfWdaxI3QJ2DNtzjWAtzdlIK0NDQzZu+B2FQsGQIT/rOEsW5qycjkKh4Jefkh+ys3eyo1qdyuzYuCeD0+mnb7q15fiRMwQFvG14hb54ycCeo2nQuDa3Hp/n+sMzWFlb4nP1psaHMr8Ud+48wKVSI6rXaMGKlWv5a9UiihcvrOtY4gug08bGTz/9RKlSpQgKCuLOnTtYWlpSo0YNHj9+nOY63NzcsLa2VtsSE5Lvck4vX18/GjTsgE22whQoWJkari0wMjLE92Ha82nT4kUzaN6sAQ0adcTf//nHX6BBzyJec+FpKG1KOKvK7CzeNjoKZLdQOzZ/NnMCImMAuPj0JacehTCrcSnKOdlQ3N6Kn+sUw8TQgN23M/Z7+K93DY08eXLSrHkXnfZqGBpmYfbK6TjlcmDAN8NT7NVo3bk54S8jOHEw5TkfmhYSEkp8fHySngF7ezsCAjOmdyU5OXM54Vq7KpvWb1MrP3X8HLUqNqdC0TqUL1ybEQPG4+Bkz5NHTzM8o75eu/+Ki4vjwYNHXL7iw/gJs/D2vsmQwZqZa/C5MsP1S40yMVEjW2al08bG2bNncXNzw9bWlkKFCrF7924aN25MzZo18fX1TVMd48aNIzw8XG0zyKLZsf/o6NcEBARhY2NNw4a12b37kEbr/xSLF82gTesmNGzciUePnmT4+Xfdek52M2Nq5ns/qc/Z0hQ7CxMevVT/5egXFo2TpSmA6hbYD994BgqFzta8edfQKFQoP02bfUtoaJhugvC+oZGnQG5+6DSc8Jcpzxtp1bkZe7bsJ/4/txVrW1xcHJcve1OvrquqTKFQUK+uK+fP625sv2OXNrwIDsXjUPINr5ehYUREvKJ6zcrY2mXn8IHjGRsQ/b12qTEwMMDExFjXMYDMef3UfOXDKDpdQfT169cYGr6PoFAo+P333xk8eDC1a9dm48aNH63DxMQEExMTtTJN3dbWsGFtFAoFd+8+oGDBfMxym8CdOw9Ys2azRur/VL8u+YVvO7ehXftevHoVicP/xyvDw18RExOj9fMnKpXsvP2cFsWcMDR432xQKBR0L5+H5Rd9KWKblaK2luy+/ZxHL6OZ27Q0AGUcrbEyMWLikZv0q5wf0yxZ+PemP/4Rr3HNp53b1ywszClYMJ/q63z5clOmTAlevgzj+fMgNv29gnLlS9G2bQ+yZMmiup6hoWHExcVpNIuZuRm587+/WypnHmeKlCxMRFgEIYEhzP1zJsVKF2HYdz9iYGBAjv+vCREeFkF8XLzqdZVdXciVNyfbN3ze3TmfYuHiP1i9aiFel73x9LzC0CF9sbAww11HnwuFQkHHLq3ZunkXCQnqDa+OXVpz/+5DXoSE4lKpLJN/+YlVv6/D9/4jnWTVt2v3XzNnjOXAgWM8fuKPpWVWvu3chtq1q9GseRddR1PR5+v3UZl4cqcm6LSxUaxYMS5dukTx4sXVyn/77TcAWrVqpYtYKtZWlkyfMZZcOZ0IDQ1j+479TJo0m/j4+I+/WIsG/NAdAI+j6l3GvXqPYO26f7R+/gtPQgl4FUOb4s5J9nUtl4fYhETmn75HeEwcRWwt+b11eXJbmwOQzcyY31qWY+n5B/Tffpn4RCUFsluwsHkZitpq524UF5cyHD60RfX13LmTAVi7bgszZiygZctGAFzyVO+xatioIydPntdolhLlivHnv7+pvh49bSgAuzbvY/m8VapFvjZ7rFF7XZ92g/E6e0X1dZsuLbh60ZtH9zN+SG/Lll3Y2WZnyqTRODrace3aDZq36EZQUMjHX6wFrrWrkiu3M/9s2JFkX4FC+fhxwjBsslnz9LE/vy34gz9/X5fxIf9P367df9nZ2bL6r8U4OdkTHv4KH59bNGvehSNHM26Y7mP0+fqJ1Ol0nQ03NzdOnTrFvn37kt0/cOBAli9fnu4JmZpYZ0Ob9H1yWnrW2chon7vOhralZ50NXdDEOhvaool1NrRJE+tsCP2UEetsRE3rqpF6LCZt0Eg9GU2nczbGjRuXYkMDYNmyZXp154cQQgjxSWS5ciGEEEII7ZFHzAshhBDalonvJNEEaWwIIYQQ2vaV340iwyhCCCGE0Crp2RBCCCG0TYZRhBBCCKFNmXmpcU2QYRQhhBBCaJX0bAghhBDaJsMoQgghhNAqaWwIIYQQQqvk1lchhBBCCO2Rng0hhBBC22QYRWQ0E0MjXUdIlT4/WTXswFRdR0hVvtZzdB0h0wqMCtN1BCG0RvmVNzZkGEUIIYQQWiU9G0IIIYS2feU9G9LYEEIIIbRNVhAVQgghhNAe6dkQQgghtE2GUYQQQgihVV95Y0OGUYQQQgihVdKzIYQQQmiZUvl192xIY0MIIYTQtq98GEUaG0IIIYS2feWNDZmzIYQQQgitkp4NIYQQQsvk2SgiRVmzWjBv3hTu3T1PeNh9ThzfgYtL2QzPMXr0QE6e2klA4HUePbrEps0rKVy4QIrHb9/hTlT0I1q0bJQh+Vxdq/Dvtr946HuJ2JgntGrZWG3/hAkj8L52jNAXdwh47sP+fRupVKmcVrIkJCaydOdJmv28nCqD59Ni/ApW7j2jNjmrXP/ZyW7uBy+ojvELDGX4sm3UGbmEGsMW0mPOejzv+Gk87+ixgwkMv622nfbcp3ZMxUrl2LbbnYfPLnP/ySV27FuHqamJxrOkx4AfunP/7nkiIx5w9vRuKlUsp5Mcrq6V2bbtL3x9PYmJeUzLD97zMTGPk91GjOivk7ygP9fuQ/37fc9lr8OEhtwmNOQ2p0/uoknjurqOlYS+Xr+PSlRqZsukpLGRihXL59Kgfk169hpGBZcGHDlykgP7/8bZ2TFDc7jWrMLKFeuoW6ctLVt+h5GRIbt2r8Xc3CzJsYMH987wWc8W5mZ4+9xi2PAJye6/d+8hw0dMxKViQ+rWa88jv6fs3bMBW9vsGs+y+sAFtpy4ythvG/LvlD4Ma1cb94MX+fuYl+qYI3MGqW1Tvm+KQgENKhRVHTPkt63EJySycmRnNv7cnSK57Bny2zZCwiM1nvn2zbuUKuyq2lo17qLaV7FSOf7e9gfHPc7QpF4nGtftyF9/bCBRh0sfd+zYinlzJzN9xgIqVWnCNe+b7Nu7ATu7HBmexdzcHB+fmwxP4b2XN6+L2tav3ygSExPZsWN/Bid9S5+u3Yf8/Z8zfrwblas2pUq1Zhw7foZ/t/1FiRJFdB1NRZ+vn0idQvkF3o9jbJLrs+swNTUl9MVt2nfoxf79Hqry8+f2cfDgMSZPmfvJdRtl+bzRK1vb7Pg9vkyjhp04c+aiqrxMmRJs3baKmq6t8H3oyTff9GPP7kPprj8+MeGTs8XGPKFjxz7s2n0wxWMsLbMSEnyLJk07c+zYmXTV/7FHzA/5bSs5rMyZ8n0zVdmo5dsxMTLkl94tk33N8GX/Eh3zhpUjOwPwMjKauqN+5a/RXahQODcAUTGx1Bi2iOXDv6Fq8Xwpnj+9j5gfPXYwTZvXp37Ntsnu33dkEyeOnWX2zCXpqjclL16/+uw6zp7ejeela6rGpUKh4JGvJ0uXrWbO3KWfXK+hQZbPyhUT85iOHfuwO5X3/D///IGlZVaaNv023fV/zufiHW1dO20JCrjOT2NnsNp9k66jANq7fvFv/DUVMUXh39XXSD3W645qpJ6MJj0bKTA0zIKhoSExMbFq5a9fx1C9emUdpXrLysoSgJcvw1RlZmam/LV6MSNGTCIwMFhHyT7OyMiIPr27EhYWjrf3TY3XX7ZATi7c9sMvMBSAO0+CuHL/KTVKJT/s9CIiitM+D2jjWkZVZmNhRj6H7Ow+f53XsW+IT0hk68mrZLc0p0QezfdqFSiYl2u3T3Lx2mGW/TGXnLmcgLeNSpdK5QgJDmXPob+5fu802/euo3LVChrPkFZGRkZUqFCGox6nVGVKpZKjHqepWtVFZ7nSwt7elqZN6+Guo1+cmenaGRgY0KlTKywszDl/wevjL8gAmen6JUeZqNTIllnpfILorVu3OH/+PNWqVaNYsWLcvn2bxYsXExsbS7du3ahXr16qr4+NjSU2Vr1BoFQqUSgUn5UrMjKKc+cu8fO44dy+fZ/AwGA6f9OGqlVdePDg0WfV/TkUCgVz5k7i7FlPbt68qyqfPWcSFy54sXfPYZ1lS02zpvVZt24p5uZmPH8eRLPmXXnx4qXGz9OrSVWiYmJpM/kPsigMSFAmMrh1LZpXKZns8bvOXcfc1Jj65d93FSsUClaM+IYRy7ZTfdhCDBQKsltasGxoJ6wsTDWa9/KlawwdOI4H9x5i72jP6J8GsXP/empXa0XefG97VUaPG8zUCXO47nOLTp1bs3WXO7WrtuShr+bnkHyMrW12DA0NCQoMUSsPCgqmWNGCGZ4nPbp168CrV1Hs2HFAJ+fPDNeuVKlinD65C1NTEyIjo+jQsQ+3bt3TdSwgc1w/kTKdNjYOHDhA69atyZo1K9HR0Wzfvp3vv/+esmXLkpiYSKNGjTh06FCqDQ43NzemTlXvWjcwsCSLodVn5+vZaxgrV8zH75EX8fHxXLlync2bd1KhQunPrvtTLVw0nRIlitKgQQdVWbPmDahduxrVqzXXWa6POX7iLJUrNyGHbTZ69erCxg3LcK3ZiuDgFxo9zyGvW+y7eBO33i0p6GzHnSeBzP3nKHY2WWlVLenPbecZb5pVLoGJ0fuPglKpxO3vw2SzMuev0V0xNTbk39PeDF26lQ0/d8fOOqvG8nocef9X2s0bd7l86RpePh60btuEu3d8AVi3ejObNvwLwHXvW9SsXY0u37Vn5tQFGsvxNejevRObNm1P8seJeO/OnQe4VGqEtZUl7ds3569Vi6jXoL3eNDgytUzcK6EJOh1GmTZtGmPGjOHFixesXr2aLl260LdvXw4fPszRo0cZM2YMs2bNSrWOcePGER4errYZZLHUSD5fXz8aNOyATbbCFChYmRquLTAyMsT34WON1J9e8xdMpWnTejRt0pln/gGq8jq1q1OgQF6ePfcmPOI+4RH3Adi48Xf2H9CPsdbo6Nc88H3ExYtX+OGHMcTHJ9CjR2eNn2fhtuP0bFyVJpVKUDinHS2qlqJb/Ur8tf98kmMv33vCo8BQ2rqq32F08bYfJ70fMLtPK8oXykXxPI6M79IIE2Mjdp+7rvHM/xUR/ooHDx6Rv0BeggKDALhz+77aMffuPlANtWS0kJBQ4uPjsXewVSu3t7cjQI+H72rUqEzRooVYvVp3n4fMcO3i4uJ48OARl6/4MH7CLLy9bzJkcB9dxwIyx/VLVaKGtkxKp42NGzdu0KNHDwA6derEq1ev6NDh/V/sXbt2xdvbO9U6TExMsLKyUts+dwjlQ9HRrwkICMLGxpqGDWunOgFNW+YvmEqrVo1p1rQLfn5P1ffN/50qlZtQrWoz1Qbw04/T+aH/6AzPmhYGBgaYmBhrvN6YN3EYGKj//A0MFCQmMw96+xlvSuRxpGhu+w/qiH/7ug/eRwYKBYla/uvE3MKcfPlzExgQzGM/f54/C6RQ4fxqxxQolI+nT55pNUdK4uLiuHzZm3p1XVVlCoWCenVdOX9eP8b2k9Ojxzd4eXnj43NLZxky47XT1uf0U2TG6yfe0/mcjXcNAwMDA0xNTbG2tlbts7S0JDw8XFfRaNiwNgqFgrt3H1CwYD5muU3gzp0HrFmzOUNzLFw0nU6dWvNNp75ERkbh4GAHQHh4BDExsQQGBic7KfTJ02dJGibaYGFhTsGC+VRf58uXmzJlSvDyZRgvXrxk7Nih7NlziICAIHLkyM4PP3TH2dmBbdv2ajxLrTKF+HPfWRyzW1HQyZY7TwJZf8ST1tXLqB0X+TqWw153GNUh6ToCZQo6Y2VuykT3vfRrXgNTY0O2nbqGf0gYNUtrdmx48owfObT/GE+fPMPB0Z4ffx5MQkIi27fuAWDZklWMGTeEG9fvcN3nFt9824ZChQvQ+/thGs2RHgsX/8HqVQvxuuyNp+cVhg7pi4WFGe4Z/LmA1N97T/7fILO0zEq7ds356acZGZ7vQ/p07T40c8ZYDhw4xuMn/lhaZuXbzm2oXbsazZp3+fiLM4g+X7+PycyTOzVBp42NfPnyce/ePQoWfPsP+Llz58iTJ49q/+PHj3Fy0k13MYC1lSXTZ4wlV04nQkPD2L5jP5MmzSY+Pj5Dc/Tr9x0ABw+pf6D69xvN+vVbMzRLclxcynD40BbV13PnTgZg7botDB48jqJFCtLt75XY2mbjxYswvLyuUa9+B27duptSlZ9sbOcGLN15CreNhwh9FY2ddVba1yxH/xY11I474HkLlEqaVC6RpI5sWc1ZOrQjv+08Sb+FfxOfkEhBJ1sWDWyXpBfkczk7O7B81XyyZbfhRUgoF8970azBN6rJsyt/X4uJqQnTfhlLtmzW3Lh+h05teuH38IlGc6THli27sLPNzpRJo3F0tOPatRs0b9GNoKCQj79Yw1xcynDo0D+qr9+999at20LfvqMA6NSpFQqFgn/+2Znh+T6kT9fuQ3Z2tqz+azFOTvaEh7/Cx+cWzZp34cjRUx9/cQbR5+v3UZl4CEQTdLrOxvLly8mdOzfNmyc/sfHnn38mKCiIP//8M131amKdDW363HU2tE0T6wloy8fW2dC19K6zkdE0sc6GtnzuOhvaps+fC/F5MmKdjdC2tTVST/btJzRST0bT6W+9H374IdX9v/zySwYlEUIIIYS26Pef2EIIIcSX4CsfRpHGhhBCCKFlyq+8sSHLlQshhBBCq6RnQwghhNC2r7xnQxobQgghhJbJMIoQQgghhBZJY0MIIYTQNh09G8Xf359u3bqRI0cOzMzMKF26NJcuXVLtVyqVTJo0CScnJ8zMzGjQoAH37qk/eC80NJSuXbtiZWWFjY0NvXv3JjIyMl05pLEhhBBCaJkyUTNberx8+ZIaNWpgZGTE/v37uXnzJvPnzydbtmyqY+bMmcOSJUtYvnw5Fy5cwMLCgsaNGxMTE6M6pmvXrty4cYPDhw+zZ88eTp48Sb9+/dKVRacriGqLrCD6efR5pURZQfTzyAqin06fPxfi82TECqJB9TWzgqj90bSvIDp27FjOnDnDqVPJLzmvVCpxdnZm1KhRjB799qGd4eHhODg44O7uTufOnbl16xYlSpTA09OTihUrAnDgwAGaNWvG06dPcXZ2TlMW6dkQQgghMonY2FgiIiLUttjY2GSP3bVrFxUrVqRjx47Y29tTvnx5/vjjD9X+hw8fEhAQQIMGDVRl1tbWVKlShXPnzgFvn1lmY2OjamgANGjQAAMDAy5cuJDm3NLYEEIIIbRMU8Mobm5uWFtbq21ubm7JntPX15fff/+dwoULc/DgQQYMGMDQoUNZs2YNAAEBAQA4ODiovc7BwUG1LyAgAHt79QdQGhoakj17dtUxaaHf/flCCCHEl0Cp0Eg148aNY+TIkWplJiYmyR6bmJhIxYoVVc8ZK1++PNevX2f58uV0795dI3nS6otsbCTq+TSU2Pg4XUdIla25la4jpMi22TRdR0hVyI4fdR0hVZYtZuo6QoqyGpvqOkKqwmKidB1BCExMTFJsXHzIycmJEiVKqJUVL16cbdu2AeDo6AhAYGAgTk5OqmMCAwMpV66c6pigoCC1OuLj4wkNDVW9Pi1kGEUIIYTQMl3cjVKjRg3u3LmjVnb37l3y5s0LQP78+XF0dOTo0aOq/REREVy4cIFq1aoBUK1aNcLCwvDy8lId4+HhQWJiIlWqVElzli+yZ0MIIYTQJ8pEzQyjpMeIESOoXr06v/zyC506deLixYusXLmSlStXAqBQKBg+fDgzZsygcOHC5M+fn4kTJ+Ls7EybNm2Atz0hTZo0oW/fvixfvpy4uDgGDx5M586d03wnCkhjQwghhPgiVapUie3btzNu3DimTZtG/vz5WbRoEV27dlUd8+OPPxIVFUW/fv0ICwvD1dWVAwcOYGr6flhzw4YNDB48mPr162NgYED79u1ZsmRJurJ8ketsGBrn1HWETE2f52y8evNa1xFSJXM2Pp2NqYWuI6RK5mx8uTJinY1n1etqpB7ns8c0Uk9Gk54NIYQQQsuUGrobJbOSCaJCCCGE0Crp2RBCCCG07Gt/xLw0NoQQQggt08XdKPpEGhtCCCGEln15t2Kkj8zZEEIIIYRWSc+GEEIIoWUyjCKEEEIIrfraGxsyjJKCmq5V2LHdncePvIh/40+rVo11HUmNvuVzdLLntxWzuel7jofPr3DszE7Kliup2t+sZUM2/fsnN33PERB2i5Kli2VIrtGjB3Ly1E4CAq/z6NElNm1eSeHCBZIcV7lyBfbt20hQ8E2eB/hw8NBmTE3T9rCj9EhITGTpnnM0m7yaKiN/o8VUd1YeuMCHa+v5BoQybMUuXMf8TtVRS+ky92+eh0ao9k/fdJQWU92pMvI36o5byfCVu3kYEKrxvCkZ8EN37t89T2TEA86e3k2liuUy7NwpGTqiHyERd5kx62dVmb29LctWzuXGvTP4Pb+Kx8nttGjVSIcp9fPa/ZfkE9ogjY0UWFiY4+19kyHDxus6SrL0KZ+1tRW7D24kPj6erh36UbtqC6ZMmE1Y2PtfjubmZlw8f5kZk+dnaDbXmlVYuWIddeu0pWXL7zAyMmTX7rWYm5upjqlcuQI7drpz9OgpatdqTa2arVmxfC2JiZqf0bX68CW2nPZmbMc6/Dv+e4a1qoH7ES/+PnFNdcyT4DB6LtxCPofs/Dm0PVvGdqVfkyqYGL3viCye256pXRvy7/jvWTawDUqlkgHLtpOQqP376zp2bMW8uZOZPmMBlao04Zr3Tfbt3YCdXQ6tnzsl5SuUpnvPb7juc1utfOnKORQqnJ9unQdQq1pL9uw+xKo1iyldprhOcurjtfsvyac9SqVmtsxK75YrVyqVKBSf192k6eXK49/4065DL3btOqjRejVF0/nSu1z5+MkjqVSlPG2afffRY3PnccbT+yj1a7blxge/GNLic5crt7XNjt/jyzRq2IkzZy4CcOz4djw8TjF92oLPqhs+vlz5kOU7yWFpzpSuDVVlo/7cg4mRIb90bwLAT6v3Y5jFgJnfp7236q5/MJ1mbWT3pO7ktrNJ8ThNLFd+9vRuPC9dY9jwCcDbhzk98vVk6bLVzJm79JPr/dTlyi0szPE4tZ0xI6cyaswAfHxuMWHsLwA8enaFMSOnsGXTTtXxdx9dYNqkeaxfuyVd59HEcuXaunaa8rXmy4jlyn1La6ZHrYDPIY3Uk9H0rmfDxMSEW7du6TqGSIfGTety7eoN/nBfyPV7pzl8chtdv++o61jJsrKyBODlyzAA7OxyULlyeYKDXnDUYxsPH3py4OBmqlWrqJXzl83vxIW7T/ALegnAnafBXPF9Ro0S+QBITFRy6sZD8trbMGDpduqOW0m3eZvwuPYgxTpfx8ax8/xNcuawwjGbpVZyv2NkZESFCmU46nFKVaZUKjnqcZqqVV20eu6UzJ4/mcMHj3Py+Nkk+zwvXqFtu2bYZLNGoVDQtn1zTExMOHP6Qobn1Mdr91+ST2iTziaIjhw5MtnyhIQEZs2aRY4cb7vFFixI/a/N2NhYYmNj1co00Tsi0i5Pvtx079WZFUvdWbxgJeXKl2LG7J+Ji3vDP3/v/HgFGUShUDBn7iTOnvXk5s27AOTLlweAn8cPZ/zPv+DtfZMuXdqxd98GKlVszIMHjzSaoVfDSkTFvKHNjLVkURiQoExkcIvqNK/0dg5LaGQ00bFx/HX4EoOaV2NYa1fO3nzEqFV7+GNIeyoWzqWqa/PJayzaeYbXb+LIZ5+N5YPaYmSYRaN5P2Rrmx1DQ0OCAkPUyoOCgilWtKBWz52ctu2bU6ZsCRrWaZ/s/t7dh/Gn+yLu+3kSFxfH6+gYuncdxEPfxxmcVP+u3Yckn3Z97c9GSVNjY9euXWmusFWrVmk6btGiRZQtWxYbGxu1cqVSya1bt7CwsEhTg8HNzY2pU6eqlSkMsqLIor9PLv3SGBgouHblBm7TFwFw3fsWxUoU5vuenfWqsbFw0XRKlChKgwYdVGUGBm/fY3/9tZF16952q1+7doM6darz/fedmDx5jkYzHLpyl32X7uDWvQkFnXJw52kwc7edxM7aglZVSpD4/1HNOqUL8F29CgAUy2XHtYfP2XraR62x0axSMaoWy0NIRDRrj3rx4+r9uI/oqDa340vmnNORmbPH06F1T2Jj3yR7zLgJw7G2tqJty+6EvnhJsxYNWOW+mBZNunDr/w1OITKCLFeeBm3atElTZQqFgoSEhDQd+8svv7By5Urmz59PvXr1VOVGRka4u7tTokSJNNUzbty4JL0k2XJkzJ0O4q2gwBDu3lHv5r93x5fmLXU76/+/5i+YStOm9WjUsBPP/ANU5QEBQQDcvnVP7fjbdx6QO7ezxnMs3HGang0r0sSlKACFnW15HvqKvw5dolWVEmSzMMPQwICCjuoT3vI7ZufKg2dqZZZmJliamZDXPhtl8jlS86fleFx7QNOKRTWe+52QkFDi4+Oxd7BVK7e3tyMgMFhr501O2XKlsLe3xePUdlWZoaEh1WpUok+/blR1aUzf/t9Ro3Iz7ty+D8CN67epWq0ivft2ZfSIyRmaV5+uXXIkn9CmNM3ZSExMTNOW1oYGwNixY9m8eTMDBgxg9OjRxMXFfdI3YGJigpWVldomQygZ6+L5yxQslE+trEChfDx98iz5F2Sw+Qum0qpVY5o17YKf31O1fX5+T3n2LIDCRdRvhy1cOD+Pn2h+0ljMm3gMPnh/GhgoVD0aRoZZKJHXgUf/n9OhyhkUhlP2lOdjKJVKUMKb+LR/Bj9FXFwcly97U6+uq6pMoVBQr64r5897afXcHzp14hyuVZpTp0Zr1Xblsg9b/9lNnRqtMTN7e8fRh3cVJSQmYGCQ8dPV9OnaJUfyaVeiUqGRLbPSaX9rpUqV8PLyYtCgQVSsWJENGzboTUPBwsKcQoXyq77Ony8PZcuWJDT0JU/04JeoPuVbuWwNuw9tZOjIfuzafoDyLqX5rntHRg9//5ejjY01OXM74ehoD6DKHhQYQnBQSLL1asLCRdPp1Kk133TqS2RkFA4OdgCEh0cQE/N2rs+ihSsZP2E4Pt638Pa+Sddu7SlSpCBduwzQeJ5apfLz5yFPHLNZ/n8YJYj1x67Quur7nrwe9Svw4+r9VCiYk0pFcnH2ph8nr/vy59C38xKehoRz8PJdqhXLQ7asZgSGRbL68CVMjAypWTKfxjN/aOHiP1i9aiFel73x9LzC0CF9sbAww33NZq2f+78iI6OS9EhFR0UTGvqS27fuYWhoiO+DRyxYPI1JE2bzMvQlzZo3pE7dGnTp1D9Ds76jL9cuJZJPe772ORufdOtrVFQUJ06c4PHjx7x5oz5WOnTo0E8KsmnTJoYPH05wcDA+Pj5pHkZJjiZufa1dqxpHj2xNUr5m7T/07jPis+v/XNrMl95bXwEaNq7Dz5NGkL9gXh77PWXF0jVs+M+thd90acPiZW5JXjdv1m/Mm5X2W9bSe+trVPSjZMv79xvN+vXvr9+oUQPo1/87smWzeXvr5Hg3zp27lK5zwcdvfY2KecPSvec4du0BoZHR2FlnpYlLEfo3qaI2uXPHuRusOuxJUFgkee2zMaBZVeqWeTsJLig8kqkbj3DrSRAR0bHksDSnQqGc9G9ShXwO2VI9vyZufQUYOKAHo0YOwNHRjmvXbjB8xCQuel75rDo/9dbX/9q5d53ara8FCuZl4pTRVKnmgoWFOQ99H7P011Vqt8KmlSZufQXtXDtN+hrzZcStr7eLNNNIPcXu7tNIPRkt3Y2NK1eu0KxZM6Kjo4mKiiJ79uyEhIRgbm6Ovb09vr6+nxzm6dOneHl50aBBAywsPv0fHk2vs/G1+ZTGRkb53HU2tO1jjQ1d01RjQxs00djQJk01NoT+kcaG9qV74HLEiBG0bNmSly9fYmZmxvnz5/Hz88PFxYV58+Z9VphcuXLRunXrz2poCCGEEPrma19BNN2NjatXrzJq1CgMDAzIkiULsbGx5M6dmzlz5vDzzz9/vAIhhBDiK6NMVGhky6zS3dgwMjJSzeS2t7fn8eO3i+NYW1vz5MkTzaYTQgghRKaX7rtRypcvj6enJ4ULF6Z27dpMmjSJkJAQ1q1bR6lSpbSRUQghhMjUMvNtq5qQ7p6NX375BScnJwBmzpxJtmzZGDBgAMHBwaxcuVLjAYUQQojMTqlUaGTLrNLds1Gx4vsHVNnb23PgwAGNBhJCCCHEl+XreIiCEEIIoUOZ+U4STUh3YyN//vyprvL5OetsCCGEEF+ir33ORrobG8OHD1f7Oi4ujitXrnDgwAHGjBmjqVxCCCGE+EKku7ExbNiwZMuXLl3KpUvpX95ZCCGE+NJl5smdmqCxRx82bdqUbdu2aao6IYQQ4ovxta8gqrEJolu3biV79uyaqk4IIYT4YsicjXQqX7682gRRpVJJQEAAwcHBLFu2TKPhhBBCCJH5pbux0bp1a7XGhoGBAXZ2dtSpU4dixYppNJzQjZDoCF1HyLT0+amqANG++rsujnmBJrqOkKqv++/Sz5PaHYxfi699zka6GxtTpkzRQgwhhBDiy/W1D6Oke4JolixZCAoKSlL+4sULsmTJopFQQgghhPhypLtnQ5nCdNjY2FiMjY0/O5AQQgjxpcnEN5JoRJobG0uWLAHejr39+eefZM2aVbUvISGBkydPypwNIYQQIhlf+zBKmhsbCxcuBN72bCxfvlxtyMTY2Jh8+fKxfPlyzScUQgghRKaW5sbGw4cPAahbty7//vsv2bJl01ooIYQQ4ksid6Ok07Fjx7SRQwghhPhiJeo6gI6l+26U9u3bM3v27CTlc+bMoWPHjhoJJYQQQogvR7obGydPnqRZs2ZJyps2bcrJkyc1EkoIIYT4kihRaGTLrNI9jBIZGZnsLa5GRkZERMjKk0IIIcSHEr/ye1/T3bNRunRpNm/enKR806ZNlChRQiOhhBBCiC9JIgqNbJlVuns2Jk6cSLt27Xjw4AH16tUD4OjRo2zcuJGtW7dqPKAQQgghMrd092y0bNmSHTt2cP/+fQYOHMioUaPw9/fHw8ODQoUKaSOjTtR0rcKO7e48fuRF/Bt/WrVqrOtIaiTfp/vpx8GcO7uXly/u8OzpNbZtXUWRIgV1HUvNgB+6c//ueSIjHnD29G4qVSyXIeeNin7N7KWrafTtD1Rs2oVuQ37m+u37qv2l63dIdlu9eScA/gFBTJq7jCZdB1KxaReadhvEUvfNxMXFZUh+ff/Z3rt7nrg3/km2JYt1/wC/iRNHJsnl43NC17HUZM1qwbx5U7h39zzhYfc5cXwHLi5ldR0rTb72ORvpbmwANG/enDNnzhAVFYWvry+dOnVi9OjRlC2bOX7oaWFhYY63902GDBuv6yjJknyfrlbNqvz++xpq1GxJk2bfYmRoxP69GzE3N9N1NAA6dmzFvLmTmT5jAZWqNOGa90327d2AnV0OrZ978vzfOed1jV/GDeXfP+dTvWJZ+v44jcDgFwAc2/KH2jZtzEAUCgUNalYF4OFjfxKVSiaN6Mf2VQv5cWAP/tl9iMWrNmo9O+j/z7Za9Wbkyl1OtTVu0hmArdv26DjZW9dv3FbLV6dOG11HUrNi+Vwa1K9Jz17DqODSgCNHTnJg/984OzvqOtpHJWpoy6wUypQedvIRJ0+eZNWqVWzbtg1nZ2fatWtH+/btqVSpkqYzppuhcU6N1hf/xp92HXqxa9dBjdarKZLv89jaZifgmQ9167Xj1OkLuo7D2dO78bx0jWHDJwBvHxHwyNeTpctWM2fu0s+qO7VHzMfExlK1xXcsmf4Ttaq6qMo7/fAjrpXLM7TXt0leM3TibKJfv+bPeVNSrHf15p1s3n2QA+uXpZpNG4+Y1+TPVht/U86fN5VmzepTvISrFmpPn4kTR9K6VRMqVmqk8bo18Yh5U1NTQl/cpn2HXuzf76EqP39uHwcPHmPylLmfXPeb2Kefne9jDjt8o5F6GgYmnTOZGaSrZyMgIIBZs2ZRuHBhOnbsiJWVFbGxsezYsYNZs2bpRUNDiPSytrYCIPRlmG6D8PaurgoVynDU45SqTKlUctTjNFX/0wDQhoSERBISEzE2NlIrNzUx5sr1W0mODwkN49SFy7RtWj/Vel9FRWNtmTXVY7RFn362HzIyMqJLl3a4r9GfXx6FCuXH75EXd26fZe2aX8md21nXkVQMDbNgaGhITEysWvnr1zFUr15ZR6nSToZR0qhly5YULVoUb29vFi1axLNnz/j11181GiYqKorVq1czfvx4fvvtN168ePHR18TGxhIREaG2fWJnjfgKKRQKFsybypkzF7lx446u42Brmx1DQ0OCAkPUyoOCgnF0sNPquS3MzShboggr1m8lKCSUhIQEdh8+ybWbdwl5EZbk+F2HjmNubkaDmlVSrPOx/3P+3rGfji0aajF58vTtZ/uh1q2bYGNjxdq1/+g6CgAXL16hd58RtGjZjcFDxpEvXx6OeWwna1YLXUcDIDIyinPnLvHzuOE4OTlgYGBAl2/bUbWqC05O9rqO91Ff+zBKmhsb+/fvp3fv3kydOpXmzZurPYjtU5UoUYLQ0FAAnjx5QqlSpRgxYgSHDx9m8uTJlChRQvVMlpS4ublhbW2ttikTX312NvF1+HXJL5QsWZQu3QbqOopecBs3FKUS6n/TD5cm37Jx+z6a1q2BwiDpX1TbD3jQvH5NTJJZdwcgMPgFP4ydSaNa1ejQPOMbG/r+s+3ZozMHDh7j+fNAXUcB4ODBY2zbtgcfn1scPnyClq2+w8bGio4dWuo6mkrPXsNQKBT4PfIi8pUvgwb1YvPmnSQmZuZfw1+HNDc2Tp8+zatXr3BxcaFKlSr89ttvhISEfPyFqbh9+zbx8fEAjBs3DmdnZ/z8/Lh48SJ+fn6UKVOG8eNTn2A4btw4wsPD1TaFgeVn5RJfh8WLZtC8WQMaNOqIv/9zXccBICQklPj4eOwdbNXK7e3tCAgM1vr5czs74r5wGhf2rOfwphX8vWwW8QkJ5HJyUDvOy/smj548o32z5IdQgkJC6T1qCuVKFmHyyP5az/0hffzZ/leePDmpX78mf/2VMRNnP0V4eAT37vlSsFA+XUdR8fX1o0HDDthkK0yBgpWp4doCIyNDfB8+1nW0j5KejTSqWrUqf/zxB8+fP6d///5s2rQJZ2dnEhMTOXz4MK9efV5vwrlz55gyZQrW1tYAZM2alalTp3L69OlUX2diYoKVlZXaponJSOLLtnjRDNq0bkLDxp149OiJruOoxMXFcfmyN/Xqvp8wqFAoqFfXlfPnvTIsh7mZKXY5shH+KpKznlepW119Pta/+z0oUaQARQvmS/LawOAX9Bo5mRJFCjB9zCAMDD7pprdPpq8/2//q3v0bgoJC2LfvqK6jpMjCwpwCBfIS8DxI11GSiI5+TUBAEDY21jRsWJvduw/pOtJHfe1zNtK9qJeFhQW9evWiV69e3Llzh1WrVjFr1izGjh1Lw4YN2bVrV7rqe9cwiImJwcnJSW1fzpw5CQ7W/l9zybGwMKdQofyqr/Pny0PZsiUJDX3JkyfPdJLpvyTfp/t1yS9827kN7dr34tWrSBz+PxciPPwVMTExOs0GsHDxH6xetRCvy954el5h6JC+WFiYZchEwjOeV1EqleTL7cxj/wAWrFxH/jw5adOkruqYyKhoDp88x+gfvk/y+sDgF/QaNRknBztG9f+el+HvH2Fgmz2b1vPr+88W3v6b1/37b1i3fgsJCQm6jqMye9ZE9uw9zOPHT3F2cmTSpFEkJCSyafMOXUdTadiwNgqFgrt3H1CwYD5muU3gzp0HrNGjSbYieelubPxX0aJFmTNnDm5ubuzevZu//vor3XXUr18fQ0NDIiIiuHPnDqVKlVLt8/PzI0cO7a8tkJyKLmU5euT9iqjz/39r35q1/9C7zwidZPovyffpBvzQHQCPo9vUynv1HsHadbqfrLdlyy7sbLMzZdJoHB3tuHbtBs1bdCMo6POGLdPiVVQ0i//cQGDIC6wts9KgZlWG9voWI8P3/1TsP3YGpVJJ07pJb9c85+XNY/8AHvsH0KCz+vCJz1HtrzCs7z9bgPr1a5I3by7c3fXrF2TOXE6sX7eUHDmyERwcypmzF3Gt2ZKQkFBdR1OxtrJk+oyx5MrpRGhoGNt37GfSpNmq4Xh9lph5OyU04pPX2dCEqVOnqn1dtWpVGjd+v9LkmDFjePr0KX///Xe66tX0OhtCfClSW2dD17SxzoYmfeW/Kz6Lvg9tZ8Q6Gzsdu2ikntYB+jvPJzU6bWxoizQ2hEieNDY+nX7/utRv0tiAHRpqbLTJpI2NjJ25JYQQQoivzmfN2RBCCCHEx2Xm21Y1QRobQgghhJYl6vlQkrbJMIoQQgghtEp6NoQQQggt++LuxEgnaWwIIYQQWva1z9mQYRQhhBDiKzBr1iwUCgXDhw9XlcXExDBo0CBy5MhB1qxZad++PYGB6g8HfPz4Mc2bN8fc3Bx7e3vGjBmT7oXUpLEhhBBCaFmiQjPbp/L09GTFihWUKVNGrXzEiBHs3r2bLVu2cOLECZ49e0a7du1U+xMSEmjevDlv3rzh7NmzrFmzBnd3dyZNmpSu80tjQwghhNCyRBQa2T5FZGQkXbt25Y8//iBbtvfPKAoPD2fVqlUsWLCAevXq4eLiwurVqzl79iznz58H4NChQ9y8eZP169dTrlw5mjZtyvTp01m6dClv3rxJcwZpbAghhBBfsEGDBtG8eXMaNGigVu7l5UVcXJxaebFixciTJw/nzp0D3j6RvXTp0jg4OKiOady4MREREdy4cSPNGWSCqBBCCKFlmrobJTY2ltjYWLUyExMTTExMkj1+06ZNXL58GU9PzyT7AgICMDY2xsbGRq3cwcGBgIAA1TH/bWi82/9uX1pJz4YQQgihZZqas+Hm5oa1tbXa5ubmluw5nzx5wrBhw9iwYQOmpqYZ/B2r+yJ7NgpYO+k6Qqoehj/XdYRUOWXNrusIKXr15rWuI6TK3Cj5vy70hT4/7CxsRBVdR0iVzcILuo6QKn1enzLxy3veZ7pp6tbXcePGMXLkSLWylHo1vLy8CAoKokKFCqqyhIQETp48yW+//cbBgwd58+YNYWFhar0bgYGBODo6AuDo6MjFixfV6n13t8q7Y9JCejaEEEKITMLExAQrKyu1LaXGRv369fHx8eHq1auqrWLFinTt2lX130ZGRhw9elT1mjt37vD48WOqVasGQLVq1fDx8SEoKEh1zOHDh7GysqJEiRJpzv1F9mwIIYQQ+kQXfTuWlpaUKlVKrczCwoIcOXKoynv37s3IkSPJnj07VlZWDBkyhGrVqlG1alUAGjVqRIkSJfjuu++YM2cOAQEBTJgwgUGDBqXYyEmONDaEEEIILfucNTK0aeHChRgYGNC+fXtiY2Np3Lgxy5YtU+3PkiULe/bsYcCAAVSrVg0LCwu6d+/OtGnT0nUeaWwIIYQQX4njx4+rfW1qasrSpUtZunRpiq/Jmzcv+/bt+6zzSmNDCCGE0LKv/dko0tgQQgghtOxrb2zI3ShCCCGE0Crp2RBCCCG0TKmnE0QzijQ2hBBCCC2TYRQhhBBCCC2Sng0hhBBCy772ng1pbAghhBBa9rU/HUaGUf7PwsKcn2eM5Njl3Xg/Ps2mvasoXe79uu857LIz69fJnPLZzzW/0/y5eQl5C+TWWd57d88T98Y/ybZk8cwMz3Lm6gEeh/ok2abPGU+u3M7J7nsc6kPz1o0yPOvwkf15GXmfX2aPV5WZmBgzd8EUHvh58iTgGms2/IadfY4My+ToZM+SFbO4/uAM9595ceTMdsqUK6na7//yRrLbD0N6ZljGDw34oTv3754nMuIBZ0/vplLFclo/p/lPy8k6+98km3Hrvm8PMDTCuHVfLCatwWLaBky7jUGR1TpJPYYudTEbvgCLGZswn7j6/esziC6u3acYM2YQcW/8mT9vqq6jqMks1+9Dmnrqa2YlPRv/N3PRBAoXK8iYQZMICgimdYdmuG9bRrMaHQkMCGbZmnnEx8cz8LtRRL6KoueArrhvXUYz1468jo7J8LzVqjcjS5Ysqq9LlizGwQOb2LptT4ZnaVn/W7Jked9uLVq8MBu3/8HenQd55h+AS7E6asd36d6R/oN7cOzIqQzNWb5CaXr06sx1n1tq5b/MHk+jxnXp8f0QIsJfMWf+FNZtWEaTht9oPZO1tRU7Dqzn7KmLdOv4Ay9CQslfMC/hYRGqY8oVra32mroNXJn/63T27Tqs9XzJ6dixFfPmTmbgoLFc9LzC0CF92Ld3AyVK1SI4+IXWzhv9248oFO/fZwaOeTDrO4UEn7MAmLToSZbiLsRsmIsyJhqT1n0x/e4nXv/+s+o1RjVbYlSzFW/2rSXh8V0UxqYostlrLfOHdHXt0quiS1n69umGt/dNXUdRk1mun0hKejYAE1MTGrWox9xpS7h07gqPHz7l17kr8Xv4hG97diBfgTyUr1SGyWNm4XP1Jg8f+DF5jBumpia0aNdYJ5lDQkIJDAxWbc2bNeD+/YecPHkuw7OEvnhJcNAL1Va/cS0e+T7m/JlLJCYmqu0LDnpB4+b12LPzINFRGfe4eAsLc1auWsCwweMJ+88vciurrHT7viPjx/3CqRPnuXb1BoMH/ESVai5UrFRO67kGDu/NM/8ARg6ewNXLPjx57M/JY2fxe/REdUxwUIja1rhZPc6eushjv6daz5ecEcP68ueqjaxZ+w+3bt1j4KCxREe/pmePzto9cVQEysgw1ZaleEUSQ56T4HsDTM0xrFSf2D3uJDy4TqK/LzFbfiNLvmIY5Cny9vVmFhg36kLs5iXEXz2FMjSQxAA/Em55ajf3f+js2qWDhYU5a9b+xg8DfuTlyzBdx1GTGa5fShI1tGVW0tgADLNkwdDQkNiYN2rlsTGxuFQph7GJ0duvY2NV+5RKJW/evMGlSrmMjJosIyMjunRph/uazbqOgpGRIW07tmDzhu3J7i9dtgSlyhRn8/p/MzTX3AVTOHTwOCeOn1UrL1u+FMbGxhw/dkZVdu+uL08e+1Opcnmt52rUpC7eV26wYvUCrt09ycETW+nyfYcUj7e1y0H9RrX4O4Ov3ztGRkZUqFCGox7ve6WUSiVHPU5TtapLxgXJYohR+VrEXfJ4+2XOAigMjUi4d+19rmB/El8Gk+X/jQ3DwmVBoUBhnQPzUUsw//kPTLqOQmGdMUNmenPtPuLXJb+wf99RPDwytufxYzLL9UuJNDZ06PLlyzx8+FD19bp166hRowa5c+fG1dWVTZs2fbSO2NhYIiIi1LZEZfp+JFFR0Vy+eI2Bo/pg72CLgYEBrTo0pVzF0tg52OJ77xH+T54zasJgrKwtMTIypO+Q7jjldMTOwTbd37emtW7dBBsbK9au/UfXUWjcvD5W1pZs/Xtnsvu/6daWe3ce4HXxWrL7taFdh+aULVeSaZPnJtnnYG9HbOwbIsJfqZUHBYXgkAE/2zz5cvFdr2946OtHl/b9WPvXZqbNGkfHzq2TPb7jt62JjIxm/27dDKHY2mbH0NCQoMAQtfKgoGAcHewyLIdhycpgakH8/xsbCstsKOPjICZa7ThlZBgKy2xvj8nuAAoFxnXbEbv7L2LWz0VhlhWzPpMhi/ZHlPXl2qWmU6dWlC9fivET3HQdJYnMcP1EynTa2OjZsycPHjwA4M8//6R///5UrFiR8ePHU6lSJfr27ctff/2Vah1ubm5YW1urbS+jA9KdZcygSSgUcPr6Aa77n+X7vp3Z8+9BlImJxMcnMLjHGPIXzMOl+8e49vg0VVxdOHHkDMpE3bc1e/bozIGDx3j+PFDXUfimW1uOHzlNYEBwkn0mpia07tCMTRn4V3nOnE64zZlIv14jiY198/EXZDADAwOue99k1vTF3PC5zYY1W9i4divf9eyU7PGdu7Zl+5Y9evm9ZCTDSvVJuHMZ5auXaX+RwgCFoRGxu1aRcPcqiY/vEvP3QhS2TmQpWEp7YTOJXLmcWTB/Gt93H6LWiys0Q6mhLbPS6QTRe/fuUbhwYQCWLVvG4sWL6dv3/czwSpUqMXPmTHr16pViHePGjWPkyJFqZRUK1El3lieP/OnWuj9m5qZktbQgOPAFi/74hSd+/gDc8L5N67pdyWppgZGxES9fhLHlgDvXr+l2AlWePDmpX78mHTv10WkOgJy5nHCtXZV+349Idn/zVg0xMzNj26bdGZapbPmS2NvbcvzM+54WQ0NDqteoRN/+39G+TU9MTIyxsrZU692wt7cl8IO/oLQhKDCYu7cfqJXdv+tLs5YNkxxbuVoFChUpwIDeo7WeKyUhIaHEx8dj/0Gvj729HQGBSRuY2qCwsSNLoTLErJujKlO+eonC0AhMzdV6NxRZbVQNknf/nxj0n7kuUREoo16hsNF+L5Y+XLvUVKhQGgcHOy5eOKAqMzQ0pGbNqgwc2AOLrPlJ1OEfV/p+/T4mM99Jogk67dkwNzcnJOTtP+j+/v5UrlxZbX+VKlXUhlmSY2JigpWVldpmoPj0b+t1dAzBgS+wsrbEtW41ju4/obY/8lUUL1+EkbdAbkqVK86RD/ZntO7dvyEoKIR9+47qNAdAp65teBEcisehk8nu/6ZbO44cOEboi3T8NfqZTh4/R/XKTalVvaVqu+zlzZbNu6hVvSVXL/vw5s0bateprnpNocL5yZ0nJ54Xr2g9n+eFKxQsnF+trEDBfPg/fZbk2G+7tefalevcvH5H67lSEhcXx+XL3tSr66oqUygU1KvryvnzXhmSwahiPZSRESTcfn++BH9flPFxZClU5n0uW2cMstmR8Pju22Mevb0LycDW+X1lZllRWFiifKn9X1b6cO1S4+FxmnLl61GxUiPVdunSVf7+ezsVKzXSaUMD9P/6idTptGejadOm/P777/z555/Url2brVu3UrZsWdX+f/75h0KFCmVIFte6VVEoFDy870ee/Ln5acpQfO89YtvfuwBo0qo+oSFhPPcPoEjxQoyfOYoj+09w5viFDMmXHIVCQffvv2Hd+i0kJCToLMe7LB27tGHrpl3JZsmbPzdVqrvQ/ZuBGZorMjKKWzfvqZVFR78mNPSlqnz92i3MdPuZly/DeRXxijnzJnPx/GUueV7Ver4/lq1l58H1DBnZl93bD1LOpTRdu3fgxxFT1I7LamlBi9aNmDYx6byTjLZw8R+sXrUQr8veeHpeYeiQvlhYmGXMBGWFAsOK9Yj3Ogb//eUXE02851FMWvQkNjoSZWw0Jq37kOB3m8T/NzaUIc+Jv3EB41a9if33d4h5jXGTriQG+5Pw4Lr2s6Pja/cRkZFR3Lih3pCNiormxYuXScp1RZ+v38fofsBdt3Ta2Jg9ezY1atSgdu3aVKxYkfnz53P8+HGKFy/OnTt3OH/+PNu3J39Xg6ZZWmVl1PjBODrbExYWwaE9HiyYuZT4+Le/OO0cbBk3bQQ57HIQHBjCjn/2smz+nxmSLSX169ckb95cuLvr/oPmWqcquXI7p3gXyjdd2/L8WSAnPc4mu1+Xfv5pJomJStau/w1jE2M8jp5i9PDJGXLua1eu0+e7YYydNJzhYwbwxO8pk3+ezfYte9WOa92uGQqFgh3b9mVIrtRs2bILO9vsTJk0GkdHO65du0HzFt0ICtL+sFOWQmUwyGZH3KWkPXmxe1ZjrFRi+t0YMDQi4e5VYrevVDsmZvMSTFr2xKzHeFAqSXh4g5hV0yExYxrrurx2X4LMfP0y83wLTVAolUqdXoOwsDBmzZrF7t278fX1JTExEScnJ2rUqMGIESOoWLFiuussYpf+12Skh+HPdR0hVU5Zs+s6Qopevcm4tTk+hbmRia4jpCowKkzXEVIUNqKKriOkymah7nox00KfpwTo+y/a+Df+Wj+HW95uGqlnnN96jdST0XS+gqiNjQ2zZs1i1qxZuo4ihBBCaEWi3je5tEvnjQ0hhBDiSydzNoQQQgihVV93v4YsVy6EEEIILZOeDSGEEELLZBhFCCGEEFolK4gKIYQQQmiR9GwIIYQQWia3vgohhBBCq77upoYMowghhBBCy6RnQwghhNAyuRtFCCGEEFr1tc/ZkGEUIYQQQmjVF9mz4avnT1XVd88iQ3UdIdPS96fSWhqb6TpCivT9qaqRJ+bpOkKqrOqM0XWEFFmbmOs6gs593f0aX2hjQwghhNAnMmdDCCGEEFolczaEEEIIIbRIejaEEEIILfu6+zWksSGEEEJo3dc+Z0OGUYQQQgihVdKzIYQQQmiZ8isfSJHGhhBCCKFlMowihBBCCKFF0rMhhBBCaNnXvs6GNDaEEEIILfu6mxoyjCKEEEIILZPGxkcM+KE79++eJzLiAWdP76ZSxXK6jqSir9n69/uey16HCQ25TWjIbU6f3EWTxnV1HSsJfb1+NV2rsGO7O48feRH/xp9WrRrrOhIAw0f252XkfX6ZPV5V1r3nN+zevwG/Z1d5GXkfK2tLHSbU7XsvITGR37Z50HT0Iir3nUHzMYtZsfMESuX7v2lfhEcy8Y/tNBg+jyr9ZjBg3jr8Al6o1TPNfTfNxyymct8Z1Bkyh2GL/+bhs2CN53V1rcL2f1fz6OEl3sQ+TfZ9NnnSaPweeREedp/9+/+mUKH8Gs+RFkNH9CMk4i4zZv2sKsuXPzdrNizltu95Hj69zJ/ui7Czy6GTfGmRiFIjW2YljY1UdOzYinlzJzN9xgIqVWnCNe+b7Nu7QS/e0Pqczd//OePHu1G5alOqVGvGseNn+HfbX5QoUUTX0VT0+fpZWJjj7X2TIcPGf/zgDFK+Qml69OrMdZ9bauVmZmYcPXyShfN+11Eydbp8763ee5otxzwZ160Z238ZxPBODXDff4aNR94+zVapVDJ8ySaeBr9k0dBv2Tz1B5xsbeg/dy3RsW9U9ZTI58S0Pq3Z/ssgfh/VDaVSyQ/z1pGQqNn7Gd69z4YNm5Ds/tGjBjJoUE8GDxmHq2tLoqOi2bNnPSYmJhrN8THlK5Sme89vuO5zW1Vmbm7Glh2rUSqVtG3xPc0adcbY2JgN/6xAoVBkaL60StTQlllJYyMVI4b15c9VG1mz9h9u3brHwEFjiY5+Tc8enXUdTa+z7dl7mP0HPLh//yH37vkycdJsIiOjqFK5gq6jqejz9Ttw8BiTJs9h584Duo4CvP2ltHLVAoYNHk9YWITavuXL3Fm0YAWenld1E+4DunzvXb3/hDrli1GrXBFy2mWjYaWSVCtZkOu+/gD4Bb7A+8FTxndvQakCOcnnZMuE75sT8yaOA+d9VPV0qFMRl6L5yGmXjeL5nBncvh4BoRE8CwnTaN6DB48xecpcdu5K/n02ZEhv3GYtYffuQ/hcv0XPXsNxdnKgdQb2tFlYmLP8z3mMGDqR8LBwVXnlqhXIkycngwf8xK2bd7l18y6DfviRcuVLUbN2tQzLlx5KDf0vs5LGRgqMjIyoUKEMRz1OqcqUSiVHPU5TtaqLDpPpd7YPGRgY0KlTKywszDl/wUvXcYDMdf30wdwFUzh08Dgnjp/VdZR0yej3XrlCubl405dHASEA3HkcwJV7j3EtXRiAuLgEAEyM3s/LNzAwwNjIkCt3HydbZ3TsG3aeukpOOxscs1tp+Tt4L3/+PDg5OeBx9P1nJCLiFRcvXqVKBn5GZs+fzOGDxzn5wXvPxNgYpVLJm//0CMXGxJKYmEjVavIZ1kc6vRtlyJAhdOrUiZo1a35yHbGxscTGxqqVKZXKz+5Ks7XNjqGhIUGBIWrlQUHBFCta8LPq/lz6nO2dUqWKcfrkLkxNTYiMjKJDxz7cunVP17GAzHH99EW7Ds0pW64k9Wq11XWUNNPVe69Xc1ciX8fSZtxvZDEwICExkSHt69O8ehkA8jnZ4pTDmiVbjjCxR0vMTIxYd/A8gaERBIdHqtW1+ehFFv5zmNexceRzzMGKMd9jZJhx/1w7ONgBEBiU9DPi+P992ta2fXPKlC1Bwzrtk+y75HmV6KjXTJo2hplTF6BQKJg4dTSGhoaq7PomMw+BaIJOezaWLl1KnTp1KFKkCLNnzyYgICDddbi5uWFtba22KRNfaSGtSI87dx7gUqkR1Wu0YMXKtfy1ahHFixfWdSyRDjlzOuE2ZyL9eo0k9j9/Qeo7Xb33Dl68wb7zPrj1b8+mKf2Z3qcta/afZdfpqwAYGWZhwZBv8At4Qc1Bs6nSbyaetx7iWqYQBh/8cdSsWhk2T/2Bv8b1IK9jDsYs3ULsmzitfw/6wjmnIzNnj+eHPqOTfe+9ePGSXt2H0rhpPfyeX8X3qRfW1lZcu3KdRA3PbdGUr30YRefrbBw6dIjdu3czb948Jk6cSNOmTenbty/NmjXDwODjbaFx48YxcuRItbJsOYp9dq6QkFDi4+Oxd7BVK7e3tyMgUPMzw9NDn7O9ExcXx4MHjwC4fMWHii7lGDK4DwMH/aTbYGSO66cPypYvib29LcfP7FSVGRoaUr1GJfr2/w6H7CX08h92Xb33Fv5zmF7NXGlatTQAhXM78PxFGKv2nKKVazkASuRz5p/pA3gVHUNcfALZrSzoOu0PSuZzVqvL0twUS3NT8jrmoEzBXLgOnI3H5duqurUt8P+fAwd7WwICglTl9vZ2XPO+ofXzly1XCnt7WzxObVeVGRoaUq1GJfr064azbSmOe5yhUtkGZM+ejfiEeCLCX3Hj3hn8tj3Rej6Rfjqfs1G6dGkWLVrEs2fPWL9+PbGxsbRp04bcuXMzfvx47t+/n+rrTUxMsLKyUts0MRs5Li6Oy5e9qVfXVVWmUCioV9eV8+d1O/dAn7OlxMDAABMTY13HADLn9dOFk8fPUb1yU2pVb6naLnt5s2XzLmpVb6mXDY3kZNR7LyY2DgMD9X97shgYkKhM+teopbkp2a0s8At4wc2Hz6hToWiK9b59uZI3cfEaTpyyhw8f8/x5IHXrvf+MWFpmpXLlclzIgM/IqRPncK3SnDo1Wqu2K5d92PrPburUaK323gsNfUlE+Ctq1qqKnV0ODuzz0Hq+T/G1342i856Nd4yMjOjUqROdOnXi8ePH/PXXX7i7uzNr1iwSEhJ0kmnh4j9YvWohXpe98fS8wtAhfbGwMMN9zWad5Mks2WbOGMuBA8d4/MQfS8usfNu5DbVrV6NZ8y66jqaiz9fPwsJcbT2D/PnyULZsSUJDX/LkybMMyxEZGcWtm+pzHaKjXxMa+lJVbm9vi72DHQUK5AWgZMmivHoVxdOnzwh7GZ6kTm3T5Xuvdrki/LH7JI7ZrSmY047bjwNYd/AcrWuWVx1z6OINslma45TDmntPg5izYT91KxSjeqlCADwNCuXgxRtUK1WQbJbmBIZG8Nfe05gYGeFaVrNDQRYW5hQqmE/1db58uSlbpgShL8N48uQZv/66inFjh3L//kMePXzClCmjefY8kJ27Dmo0R3IiI6O4/cE8m+ioaEJDX6rKv+3ajrt3H/AiJJRKlcszc/Z4li915/79h1rP9ymSa3R+TfSmsfFfefLkYcqUKUyePJkjR47oLMeWLbuws83OlEmjcXS049q1GzRv0Y2gDyZNSTZ1dna2rP5rMU5O9oSHv8LH5xbNmnfhyH9mtuuaPl+/ii5lOXpkq+rr+fOmALBm7T/07jNCR6mS17NPF8b+PFT19b5DmwAY2P9H/t7wb4bn0eV7b2y3Ziz914Nf1u0lNCIKOxtLOtRxoX/r2qpjgsNfMW/TQV6ER2JnY0mL6mXp37qWar+xkSGX7/qx/tB5IqJek8M6Ky5F8rJ2Qm9yWGXVaF4Xl7IcObxF9fW8uVMAWLv2H/r0Hcm8+cuwsDBn2dLZ2NhYceasJy1bdksyIV9XChUuwIQpo8iWzZonj/1ZOHc5vy9dretYIgUKpVJ3za38+fNz6dIlcuTQ7EJKhsY5NVqfEF8KS2MzXUdI0as3r3UdIVWRJ+bpOkKqrOqM0XWEFFmZmOs6QqpCIu5q/Rzd8rbTSD3r/TK+Ea8JOu3ZePhQP7u7hBBCCE3KzEuNa4LOJ4gKIYQQ4suml3M2hBBCiC9JZl4jQxOksSGEEEJoWWa+bVUTpLEhhBBCaJnM2RBCCCGE0CLp2RBCCCG0TOZsCCGEEEKrvvY5GzKMIoQQQnyB3NzcqFSpEpaWltjb29OmTRvu3LmjdkxMTAyDBg0iR44cZM2alfbt2xMYGKh2zOPHj2nevDnm5ubY29szZswY4uPT96weaWwIIYQQWqZUKjWypceJEycYNGgQ58+f5/Dhw8TFxdGoUSOioqJUx4wYMYLdu3ezZcsWTpw4wbNnz2jX7v1qpwkJCTRv3pw3b95w9uxZ1qxZg7u7O5MmTUpXFp0uV64tsly5EMmT5co/nSxX/ulkuXJonaeFRurZ+XjPJ782ODgYe3t7Tpw4Qa1atQgPD8fOzo6NGzfSoUMHAG7fvk3x4sU5d+4cVatWZf/+/bRo0YJnz57h4OAAwPLly/npp58IDg7G2DhtT1SWng0hhBAik4iNjSUiIkJtS+vD8cLD3z6JOXv27AB4eXkRFxdHgwYNVMcUK1aMPHnycO7cOQDOnTtH6dKlVQ0NgMaNGxMREcGNGzfSnFsaG0IIIYSWJWpoc3Nzw9raWm1zc3P7+PkTExk+fDg1atSgVKlSAAQEBGBsbIyNjY3asQ4ODgQEBKiO+W9D493+d/vSSu5GEUkodB0gFfo+5qfP1w70e6jCQKHfVy9r7dG6jpCq6EeHdB0hReb5Guk6gs5p6tbXcePGMXLkSLUyExOTj75u0KBBXL9+ndOnT2skR3pJY0MIIYTIJExMTNLUuPivwYMHs2fPHk6ePEmuXLlU5Y6Ojrx584awsDC13o3AwEAcHR1Vx1y8eFGtvnd3q7w7Ji1kGEUIIYTQskSUGtnSQ6lUMnjwYLZv346Hhwf58+dX2+/i4oKRkRFHjx5Vld25c4fHjx9TrVo1AKpVq4aPjw9BQUGqYw4fPoyVlRUlSpRIcxbp2RBCCCG0TBc3fg4aNIiNGzeyc+dOLC0tVXMsrK2tMTMzw9ramt69ezNy5EiyZ8+OlZUVQ4YMoVq1alStWhWARo0aUaJECb777jvmzJlDQEAAEyZMYNCgQenqYZHGhhBCCKFlulhB9PfffwegTp06auWrV6+mR48eACxcuBADAwPat29PbGwsjRs3ZtmyZapjs2TJwp49exgwYADVqlXDwsKC7t27M23atHRlkXU2RBL6PE1P39+s+nztQL+vn75PEE3U838qZYLop4t/46/1czTO3VQj9Rx8sl8j9WQ06dkQQgghtEwexCaEEEIIrUrv5M4vjdyNIoQQQgitkp4NIYQQQsu+wOmR6SKNDSGEEELLZBhFCCGEEEKLpLGRgv79vuey12FCQ24TGnKb0yd30aRxXV3HUjPgh+7cv3ueyIgHnD29m0oVy+k6EgATJ44k7o2/2ubjc0LXsZKQ6/dparpWYcd2dx4/8iL+jT+tWjXWdSQ1WbNaMG/eFO7dPU942H1OHN+Bi0tZXccCdHvtoqJfM/u3VTTq3I+KjTvTbfA4rt++p9pfum67ZLfVm3ao1XPy3CW6DPiJio07U73ldwydMCvDvgfQ38/txyg19L/MSoZRUuDv/5zx4924d/8hCoWC77/ryL/b/qJi5cbcvHlX1/Ho2LEV8+ZOZuCgsVz0vMLQIX3Yt3cDJUrVIjj4ha7jcf3GbZo06az6Oj4+XodpkpLr9+ksLMzx9r7JavdNbNuyStdxklixfC4lSxalZ69hPH8eSJdv23Fg/9+ULVePZ8/S/pRKbdDltZs8dyn3Hz7hl3HDsLfNzp7DJ+g7eio7Vi/GwS4Hx7ap5zl14TKT5y6jQa2qqrLDJ84xZf7vDOvTlcrlS5OQkMC9h48z7HvQ989tavR9nRZtk0W90iEo4Do/jZ3BavdNWqk/Pc6e3o3npWsMGz4BAIVCwSNfT5YuW82cuUs/q+7PXVpp4sSRtG7VhIqVNL+Qj6berNq6fppYliozXL934t/4065DL3btOvjZdWliUS9TU1NCX9ymfYde7N/voSo/f24fBw8eY/KUuZ9ct6Z/WWjy2kHqi3rFxMZStVlXlswYS61qFVXlnfqNxrVKBYb27pLkNUMnzCI6+jV/Lpj6Nm9CAo0792dQj860a94gXdk0taiXtj63GbGoV62c9TVSz0n/ox8/SA/JMEoaGBgY0KlTKywszDl/wUvXcTAyMqJChTIc9TilKlMqlRz1OE3Vqi46TPZeoUL58XvkxZ3bZ1m75ldy53bWdSQVuX5fLkPDLBgaGhITE6tW/vp1DNWrV9ZRKt1LSEgkITERY2NjtXJTE2Ou+NxKcnxIaBinznvRttn7X5C37voSFBKKwkBBx76jqNu+Fz/8NJ17D/20nh8yx+c2NUoNbZmVNDZSUapUMcJC7xId+ZBlv82iQ8c+3Lp17+Mv1DJb2+wYGhoSFBiiVh4UFIyjg52OUr138eIVevcZQYuW3Rg8ZBz58uXhmMd2sma10HU0QK7flywyMopz5y7x87jhODk5YGBgQJdv21G1qgtOTva6jqczFuZmlC1ZlBXrthAUEkpCQgK7D5/g2s27hIS+THL8roPHMDc3UxtCefr87WPFf1+zmX7dOvDbL+OxypqVXsMnER7xSuvfg75/bj9GF0991Sc6b2z89ttvfP/992za9HZoYt26dZQoUYJixYrx888/f3SsOjY2loiICLVNUyNDd+48wKVSI6rXaMGKlWv5a9UiihcvrJG6v2QHDx5j27Y9+Pjc4vDhE7Rs9R02NlZ07NBS19EyBbl+n6dnr2EoFAr8HnkR+cqXQYN6sXnzThITdfEoLP3hNm4YSqWS+h374NLoGzb+u5em9VxRJDN8tX2/B80b1MTkPz0h765f364daFi7GiWLFmTGT4NRKBQcPH42w76PzOprb2zodILojBkzmDNnDo0aNWLEiBH4+fkxd+5cRowYgYGBAQsXLsTIyIipU6emWIebm1uS/QqDrCiyWH12vri4OB48eATA5Ss+VHQpx5DBfRg46KfPrvtzhISEEh8fj72DrVq5vb0dAYHBOkqVsvDwCO7d86VgoXy6jgLI9fvS+fr60aBhB8zNzbCysiQgIIgN65fhm4ETGfVR7pyOuC+eQfTrGKKio7HLkZ3RU+eRy8lB7Tgv75s8euLPvEkj1crtcmQDoGC+3KoyY2Mjcjk5EBCk3tugDZntcyvU6bRnw93dHXd3d7Zu3cqBAwcYP348ixcvZvz48YwbN44VK1awcePGVOsYN24c4eHhapvCwFIreQ0MDDAxMf74gVoWFxfH5cve1KvrqipTKBTUq+vK+fO6n1PyIQsLcwoUyEvA8yBdRwHk+n0toqNfExAQhI2NNQ0b1mb3bv19KmpGMjczxS5HdsJfRXLW8yp1a6jPZfl331FKFClI0UL51cpLFCmIsZERjx6/n0wZFx+Pf2AQThkwjJHZPrcfUiqVGtkyK532bDx79oyKFd/OjC5btiwGBgaUK1dOtb9ChQo8e/Ys1TpMTEwwMTFRK0uuWzC9Zs4Yy4EDx3j8xB9Ly6x827kNtWtXo1nzpLO2dWHh4j9YvWohXpe98fS8wtAhfbGwMMN9zWZdR2P2rIns2XuYx4+f4uzkyKRJo0hISGTT5h26jqYi1+/TWViYU+g/v4jy58tD2bIlCQ19yZMnqX9eM0LDhrVRKBTcvfuAggXzMcttAnfuPGCNHvxsdXntzly8ghIl+XLn5LH/cxYsX0v+PDlp07Se6pjIqGgOnzjL6AE9krw+q4U5nVo1Yqn7JhztbXFysMP9/+/JRnWqazX7O/r8uf2YzDwEogk6bWw4Ojpy8+ZN8uTJw71790hISODmzZuULFkSgBs3bmBvr5tJXXZ2tqz+azFOTvaEh7/Cx+cWzZp34cjRUx9/cQbYsmUXdrbZmTJpNI6Odly7doPmLboRlAHdmR+TM5cT69ctJUeObAQHh3Lm7EVca7YkJCRU19FU5Pp9uoouZTl6ZKvq6/nzpgCwZu0/9O4zQkep3rO2smT6jLHkyulEaGgY23fsZ9Kk2XqxVokur92rqGgW/7mewOAXWFtmpUGtagzt3QUjw/e/BvZ7nEapVNK0nmuydYz8oTtZsmRhnNtiYmPfULp4YVbNn4q1ZVatZn9Hnz+3InU6XWdj4sSJrFixgtatW3P06FG++eYbNm7cyLhx41AoFMycOZMOHTqwYMGCdNWrrXU2vhaaWCtCW/T9bwN9vnag39dPE+tsaJO+L8qU2jobuqapdTa0JSPW2ajkXEsj9Xg+O6mRejKaTns2pk6dipmZGefOnaNv376MHTuWsmXL8uOPPxIdHU3Lli2ZPn26LiMKIYQQny0zz7fQBFlBVCShz39f6vubVZ+vHej39ZOejc8jPRufLiN6Nio61dRIPZee68dQfnrJs1GEEEIILZMJokIIIYTQqi9wECFddL6CqBBCCCG+bNKzIYQQQmiZDKMIIYQQQquU0tgQQgghhDbp+91M2iZzNoQQQgihVdKzIYQQQmiZDKMIIYQQQqtkGEUIIYQQQoukZ0MIIYTQMhlGEUIIIYRWfe3DKNLYEEl83R+Jz6Pv187QIIuuI6QoPjFB1xFSpc/XDsBCjx92Fn13p64jCB2TxoYQQgihZTKMIoQQQgit+tqHUeRuFCGEEEJolfRsCCGEEFomwyhCCCGE0CqlMlHXEXRKGhtCCCGEln3tj5iXORtCCCGE0Crp2RBCCCG0TPmV340ijQ0hhBBCy2QYRQghhBBCi6RnQwghhNCyr30YRXo2PmLAD93/1969x0VV538cfw0gI5fxLggqNy+INxIQQivTWJV1XV03NS+Fom4XSJE0NTMsU0xX09Sfl1IszdT1QmqlspgoqSkQ3sVraoYQKSgQt5nz+8Pd2SZQLBjOqJ/n43H+4DuH73nPEeEz3/M958v5s4fIv3WBA8nb6RzwmNqRjCw5G0i+qrCUbE88Ecjmzau4ePEIRUVX6Nu3/Pob3t4t2bRpJVlZJ/j55zMkJ2+neXNXFdLCi/94gbTUBG7knOFGzhmS922jd6/uqmSBys9fUdGVCrfx419UKfH/TJwYQWnJNeb98+0aOV5B4S+8t3QNPZ8fS0DfEQyPms6JjAsm+1y8co1XY+YR/LfRBP41nOdenUZmdo7x9eKSEt5dHMcTz75IYL9wxr+zgJybeTWSvzIGRamW7UElxcY9DBz4V/45N4YZ786nc1Bvjh47xZdffErjxg3VjmbR2UDyPSzZ7O3tOX78FFFRb1b4upeXO3v2bCYj4wI9ew6mc+dexMZ+QFFRcQ0nvePatUymTo0l8PFQgoL/zNd7v2HL5lW0bdtalTyVnT93d3+T7R//eA2DwUB8/Fc1nNRUgL8vY0YP59ixUzV2zJj3P+Rg2nFmvf4yW5bNpot/B8ZMjiUr5wYAV3/M4oXod/Bs7sKquW+yeVksLw7tj61tLWMfc5atJenQd8x7cyxx/5xG9o2bjH/n/Rp7D+LuNMpDOLZjY9u0Wvo5kLydIylHGfefXxQajYbvLx5hyf/FMWfukmo5xsOYDSSfpWarysqlRUVXGDhwNNu37za2ffLJYsrKyggPj6pSLjDfqq/Z108wafK7xK1eX6V+qrrqa0Xn77c2bvwQnc6R0NAhv7t/fTWdPwcHew4f3sWrr77BG1PGcvToKV6bEFOlPgsqWfW1qLiEx/uP4oPp0TwV1MnYPihiKk909mXsiEFMnLUIGxtrYl9/pcI+bhcU8tSgl3hvcgQ9nwwC4OKVH+k3ZiJrF0zH16fVXY9v6xHwB97V79Oknk+19HM993S19FPTZGTjLmrVqoWfX0cS9+w3timKQuKeZB5/3F/FZJadDSRfVVhytt/SaDSEhvbg3LmLbN++hitX0ti37/MKL7WowcrKikGD/oqDgz2Hvk1VO06lnJwaERrag9VVLIqqatEHs/jqy0T2/Opn0Nz0ej16g8FklAKgttaW706exWAwsO9wOu5NXXjxjdl0G/QyQ8e+ReKBFOO+p85doqxMz+Od2hvbvNxccXFqyNHT52vsvdyNoijVsj2oVC02MjMzeeutt+jRowc+Pj60a9eOvn37snLlSvR683zKuV+NGjXAxsaG7Kwck/bs7J9o4txYpVR3WHI2kHxVYcnZfsvJqRE6nSMTJrzC7t17+ctfhrNt2y42bFjBk//5ZKmG9u3bkHvjLIX5l/i/xbN5duBoTp8+p1qe+zV8+LPcvl1AfPxO1TIMGvRXOnVqz9Q3Y2v0uA72dvj6tGL5uniyf76JXm9ge2IyR0+fI+dGLjdyb1H4SxGrNmyna4Avy2Mn0aNrAOPfWcCRY3c+6efcyKVWLRvqODqY9N2wXl1ybuTW6PsR5al2N0pKSgohISG0bNkSOzs7zp07x9ChQykpKWHChAmsWrWKnTt3otPp7tlPcXExxcWm14cVRUGj0ZgzvhCPPCurO59VduzYzaJFKwE4duwUjz/uz5gxw9m//1tVcmVkXMC/c0/q1tHx97/3YdXKBfQI+bvFFxxhYYNYv35rud9nNaVZM1fmz3uH0D8PUSVD7OsvM23+Cp4ZGom1lRU+LT0IfboLp85dMk6MfDrYjxcGhALQpoUHR0+d419fJNK5Y/VcojAnec6GSqKiohg/fjwpKSns37+f1atXc/bsWdavX8/FixcpLCzkzTcrnlT1a7GxsdStW9dkUwy3q5wvJ+cGZWVlODk3Mml3cmrM9ayfqtx/VVhyNpB8VWHJ2X4rJ+cGpaWl5f6InzlznubNq2fe1B9RWlrKhQvfk/bdcaa+OZtjx07xauRo1fLcj65dA/H2bklcnHqXUPz8OuDs3JjD3+7kl8LL/FJ4mW7duhAZGc4vhZeNxaW5NHd1ZvU/p/Ht5ytJWPsBny2aQVlZGc1cnKhfR4eNtTUt3E1/rjybuxrvRmnUoB6lpWXcyi8w2efn3DwaNahn1uz3Qy6jqCQtLY3nn3/e+PXQoUNJS0sjKyuL+vXrM2fOHDZt2lRpP1OmTCEvL89k01jdezTkfpSWlpKWdowe3Z8wtmk0Gnp0f4JDh9S9/mvJ2UDyVYUlZ/ut0tJSUlKO0rp1C5P2Vq08uXLlB5VSlWdlZYVWa6t2jHsaMWIwqanHOH5cvcl/e/Yk81inHgR07mncUlLS+eyzrQR07onBUDOrltrXrk3jhvXJu13AgdTjdA/2p1YtG9q19uL7HzJN9r187TouTncK87atPLGxsebb704aX7909Ucys3/G16dljWS/l0f91lfVLqM4OTmRmZmJl5cXAFlZWZSVlVGnTh0AWrVqxY0bNyrtR6vVotVqTdqq6xLK+ws/JG7l+6SmHePIke8Y++oYHBzsWP3xhmrp/2HNBpLvYcnm4GBPixYexq89PJrTsWNbbt7M5erVH3n//eWsXbuE5ORv2bv3AD17Pk2fPiH07Dm4xrMCzHx3Mjt3fs2Vq9fQ6RwZ8lx/unUL5s99hqqSp7LzB6DTOTJgQB8mTXpXlYz/lZ9fwMmTGSZtBQWF/PzzzXLt5vBNyjEURcGjuQtXrmUx/6N1eDZ3oX/PpwAYObAPE2Ytwr99GwJ925KccoykQ2msmntnBFznYM+AXk8zd8Va6uoccHCwJ3bJx/j6tLrnnSiiZqhWbPTv35+XXnqJuXPnotVqmTFjBt26dcPOzg6AjIwMmjZVbygW4F//2kbjRg2Y/tYEmjRpzNGjJ+nzl+FkZ+dU/s2PcDaQfA9LNn//juzevdH49dy5d26BXLPmX4wZ8xrbtt25RXLixAjmzXubs2cv8NxzL3LgwJEazwrQuHEj4lYtxMXFiby82xw/fpo/9xnKvxNr7s6KX6vs/MGdSZkajYaNG+99e+jD7nZBIQvjNpCVc4O6OkdCunZm7MhB1LK582fqma6deWtsOB+t38bspZ/g0cyF+dPG4dfe29jH6y8NR2OlYfyMhZSWltEloANvRo5U6y2ZeJAvgVQH1Z6zkZ+fz6hRo9iyZQt6vZ7g4GDWrl2Lp6cnALt37yYvL4+BAwf+7r6r6zkbQjxsqvqsCHMy13M2qoslnzuovudsmENlz9lQW008Z6OuY4vKd7oPefkXKt/JAqn+UK+ioiLKyspwdHSstj6l2BCiYpb8B1OKjaqRYuOPk2LD/FRfiK127dpqRxBCCCHM6lG/jKJ6sSGEEEI87B7kO0mqgzyuXAghhBBmJSMbQgghhJkpj/gTRKXYEEIIIcxMLqMIIYQQQpiRjGwIIYQQZiZ3owghhBDCrB71ORtyGUUIIYQwMzVXfV2yZAkeHh7Url2boKAgDh8+XM3vrnJSbAghhBAPqQ0bNhAdHU1MTAxpaWn4+vrSq1cvsrOzazSHFBtCCCGEmak1sjF//nzGjBnDyJEjadu2LcuWLcPe3p5Vq1aZ4V3enRQbQgghhJkp1bT9HiUlJaSmphISEmJss7KyIiQkhIMHD1bp/fxeMkFUCCGEeEAUFxdTXFxs0qbVatFqteX2zcnJQa/X4+zsbNLu7OzMmTNnzJqzHEXcU1FRkRITE6MUFRWpHaVClpzPkrMpiuSrKkvOZ8nZFEXyVYUlZ6sJMTEx5QY8YmJiKtz32rVrCqAcOHDApH3ixIlKYGBgDaT9H9WXmLd0t27dom7duuTl5VGnTh2145RjyfksORtIvqqy5HyWnA0kX1VYcraa8HtGNkpKSrC3t2fTpk3079/f2B4WFkZubi6ff/65ueMayZwNIYQQ4gGh1WqpU6eOyVZRoQFga2uLv78/iYmJxjaDwUBiYiLBwcE1FRmQORtCCCHEQys6OpqwsDACAgIIDAxkwYIFFBQUMHLkyBrNIcWGEEII8ZAaPHgwP/30E2+99RbXr1/nscceY+fOneUmjZqbFBuV0Gq1xMTE3HWYSm2WnM+Ss4HkqypLzmfJ2UDyVYUlZ7NUkZGRREZGqppBJogKIYQQwqxkgqgQQgghzEqKDSGEEEKYlRQbQgghhDArKTaEEEIIYVZSbFRiyZIleHh4ULt2bYKCgjh8+LDakQDYt28fffv2xdXVFY1GQ3x8vNqRjGJjY+ncuTM6nQ4nJyf69+9PRkaG2rGMli5dSseOHY0PxAkODuarr75SO1aFZs+ejUajISoqSu0oAEyfPh2NRmOytWnTRu1YJq5du8bw4cNp2LAhdnZ2dOjQgZSUFLVjAeDh4VHu/Gk0GiIiItSOhl6vZ9q0aXh6emJnZ0eLFi2YMWPGH1pp1Fxu375NVFQU7u7u2NnZ0aVLF44cOaJ2LHEfpNi4hw0bNhAdHU1MTAxpaWn4+vrSq1cvsrOz1Y5GQUEBvr6+LFmyRO0o5SQlJREREcGhQ4dISEigtLSUnj17UlBQoHY0AJo1a8bs2bNJTU0lJSWFHj160K9fP06ePKl2NBNHjhxh+fLldOzYUe0oJtq1a0dmZqZxS05OVjuS0c2bN+natSu1atXiq6++4tSpU8ybN4/69eurHQ2482/663OXkJAAwMCBA1VOBu+99x5Lly5l8eLFnD59mvfee485c+awaNEitaMZjR49moSEBNasWcPx48fp2bMnISEhXLt2Te1oojI1uhLLAyYwMFCJiIgwfq3X6xVXV1clNjZWxVTlAcrWrVvVjnFX2dnZCqAkJSWpHeWu6tevr3z00UdqxzC6ffu20qpVKyUhIUHp1q2bMm7cOLUjKYpyZxEoX19ftWPc1aRJk5QnnnhC7Rj3bdy4cUqLFi0Ug8GgdhSlT58+Snh4uEnbgAEDlGHDhqmUyFRhYaFibW2t7Nixw6Tdz89PmTp1qkqpxP2SkY27KCkpITU1lZCQEGOblZUVISEhHDx4UMVkD568vDwAGjRooHKS8vR6PevXr6egoKDG1wq4l4iICPr06WPy82cpzp07h6urK15eXgwbNowrV66oHclo27ZtBAQEMHDgQJycnOjUqRMffvih2rEqVFJSwtq1awkPD0ej0agdhy5dupCYmMjZs2cBOHr0KMnJyYSGhqqc7I6ysjL0ej21a9c2abezs7Oo0TVRMXmC6F3k5OSg1+vLPdLV2dmZM2fOqJTqwWMwGIiKiqJr1660b99e7ThGx48fJzg4mKKiIhwdHdm6dStt27ZVOxYA69evJy0tzSKvRQcFBbF69Wq8vb3JzMzk7bff5sknn+TEiRPodDq143Hx4kWWLl1KdHQ0b7zxBkeOHGHs2LHY2toSFhamdjwT8fHx5ObmMmLECLWjADB58mRu3bpFmzZtsLa2Rq/XM3PmTIYNG6Z2NAB0Oh3BwcHMmDEDHx8fnJ2d+eyzzzh48CAtW7ZUO56ohBQbwqwiIiI4ceKExX3y8Pb2Jj09nby8PDZt2kRYWBhJSUmqFxxXr15l3LhxJCQklPsEZwl+/Sm3Y8eOBAUF4e7uzsaNGxk1apSKye4wGAwEBAQwa9YsADp16sSJEydYtmyZxRUbK1euJDQ0FFdXV7WjALBx40Y+/fRT1q1bR7t27UhPTycqKgpXV1eLOXdr1qwhPDycpk2bYm1tjZ+fH0OGDCE1NVXtaKISUmzcRaNGjbC2tiYrK8ukPSsriyZNmqiU6sESGRnJjh072LdvH82aNVM7jglbW1vjpyF/f3+OHDnCwoULWb58uaq5UlNTyc7Oxs/Pz9im1+vZt28fixcvpri4GGtraxUTmqpXrx6tW7fm/PnzakcBwMXFpVzB6OPjw+bNm1VKVLHLly/z73//my1btqgdxWjixIlMnjyZ5557DoAOHTpw+fJlYmNjLabYaNGiBUlJSRQUFHDr1i1cXFwYPHgwXl5eakcTlZA5G3dha2uLv78/iYmJxjaDwUBiYqJFXdu3RIqiEBkZydatW9mzZw+enp5qR6qUwWCguLhY7Rg888wzHD9+nPT0dOMWEBDAsGHDSE9Pt6hCAyA/P58LFy7g4uKidhQAunbtWu4267Nnz+Lu7q5SoorFxcXh5OREnz591I5iVFhYiJWV6Z8Ea2trDAaDSonuzsHBARcXF27evMmuXbvo16+f2pFEJWRk4x6io6MJCwsjICCAwMBAFixYQEFBASNHjlQ7Gvn5+SafJi9dukR6ejoNGjTAzc1NxWR3Lp2sW7eOzz//HJ1Ox/Xr1wGoW7cudnZ2qmYDmDJlCqGhobi5uXH79m3WrVvH3r172bVrl9rR0Ol05ea2ODg40LBhQ4uY8zJhwgT69u2Lu7s7P/74IzExMVhbWzNkyBC1owEwfvx4unTpwqxZsxg0aBCHDx9mxYoVrFixQu1oRgaDgbi4OMLCwrCxsZxfwX379mXmzJm4ubnRrl07vvvuO+bPn094eLja0Yx27dqFoih4e3tz/vx5Jk6cSJs2bSzid7KohNq3w1i6RYsWKW5uboqtra0SGBioHDp0SO1IiqIoytdff60A5bawsDC1o1WYC1Di4uLUjqYoiqKEh4cr7u7uiq2trdK4cWPlmWeeUXbv3q12rLuypFtfBw8erLi4uCi2trZK06ZNlcGDByvnz59XO5aJ7du3K+3bt1e0Wq3Spk0bZcWKFWpHMrFr1y4FUDIyMtSOYuLWrVvKuHHjFDc3N6V27dqKl5eXMnXqVKW4uFjtaEYbNmxQvLy8FFtbW6VJkyZKRESEkpubq3YscR9kiXkhhBBCmJXM2RBCCCGEWUmxIYQQQgizkmJDCCGEEGYlxYYQQgghzEqKDSGEEEKYlRQbQgghhDArKTaEEEIIYVZSbAjxEBoxYgT9+/c3fv30008TFRVV4zn27t2LRqMhNze3xo8thLAcUmwIUYNGjBiBRqNBo9EYF4N75513KCsrM+txt2zZwowZM+5rXykQhBDVzXIezC/EI6J3797ExcVRXFzMl19+SUREBLVq1WLKlCkm+5WUlGBra1stx2zQoEG19COEEH+EjGwIUcO0Wi1NmjTB3d2dl19+mZCQELZt22a89DFz5kxcXV3x9vYG4OrVqwwaNIh69erRoEED+vXrx/fff2/sT6/XEx0dTb169WjYsCGvv/46v12F4LeXUYqLi5k0aRLNmzdHq9XSsmVLVq5cyffff0/37t0BqF+/PhqNhhEjRgB3FhCLjY3F09MTOzs7fH192bRpk8lxvvzyS1q3bo2dnR3du3c3ySmEeHRJsSGEyuzs7CgpKQEgMTGRjIwMEhIS2LFjB6WlpfTq1QudTsf+/fv55ptvcHR0pHfv3sbvmTdvHqtXr2bVqlUkJydz48YNtm7des9jvvDCC3z22Wd88MEHnD59muXLl+Po6Ejz5s3ZvHkzABkZGWRmZrJw4UIAYmNj+eSTT1i2bBknT55k/PjxDB8+nKSkJOBOUTRgwAD69u1Leno6o0ePZvLkyeY6bUKIB4nKC8EJ8UgJCwtT+vXrpyiKohgMBiUhIUHRarXKhAkTlLCwMMXZ2dlklc01a9Yo3t7eisFgMLYVFxcrdnZ2yq5duxRFURQXFxdlzpw5xtdLS0uVZs2aGY+jKKYrx2ZkZCiAkpCQUGHG/64ofPPmTWNbUVGRYm9vrxw4cMBk31GjRilDhgxRFEVRpkyZorRt29bk9UmTJpXrSwjx6JE5G0LUsB07duDo6EhpaSkGg4GhQ4cyffp0IiIi6NChg8k8jaNHj3L+/Hl0Op1JH0VFRVy4cIG8vDwyMzMJCgoyvmZjY0NAQEC5Syn/lZ6ejrW1Nd26dbvvzOfPn6ewsJA//elPJu0lJSV06tQJgNOnT5vkAAgODr7vYwghHl5SbAhRw7p3787SpUuxtbXF1dUVG5v//Td0cHAw2Tc/Px9/f38+/fTTcv00btz4Dx3fzs7ud39Pfn4+AF988QVNmzY1eU2r1f6hHEKIR4cUG0LUMAcHB1q2bHlf+/r5+bFhwwacnJyoU6dOhfu4uLjw7bff8tRTTwFQVlZGamoqfn5+Fe7foUMHDAYDSUlJhISElHv9vyMrer3e2Na2bVu0Wi1Xrly564iIj48P27ZtM2k7dOhQ5W9SCPHQkwmiQliwYcOG0ahRI/r168f+/fu5dOkSe/fuZezYsfzwww8AjBs3jtmzZxMfH8+ZM2d45ZVX7vmMDA8PD8LCwggPDyc+Pt7Y58aNGwFwd3dHo9GwY8cOfvrpJ/Lz89HpdEyYMIHx48fz8ccfc+HCBdLS0li0aBEff/wxAC+99BLnzp1j4sSJZGRksG7dOlavXm3uUySEeABIsSGEBbO3t2ffvn24ubkxYMAAfHx8GDVqFEVFRcaRjtdee43nn3+esLAwgoOD0el0/O1vf7tnv0uXLuXZZ5/llVdeoU2bNowZM4aCggIAmjZtyttvv83kyZNxdnYmMjISgBkzZjBt2jRiY2Px8fGhd+/efPHFF3h6egLg5ubG5s2biY+Px9fXl2XLljFr1iwznh0hxINCo9xtFpkQQgghRDWQkQ0hhBBCmJUUG0IIIYQwKyk2hBBCCGFWUmwIIYQQwqyk2BBCCCGEWUmxIYQQQgizkmJDCCGEEGYlxYYQQgghzEqKDSGEEEKYlRQbQgghhDArKTaEEEIIYVZSbAghhBDCrP4f/lCYsyElwq8AAAAASUVORK5CYII=",
+ "text/plain": [
+ "<Figure size 640x480 with 2 Axes>"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "import matplotlib.pyplot as plt\n",
+ "import seaborn as sns\n",
+ "\n",
+ "sns.heatmap(cm, annot=True, fmt=\"d\")\n",
+ "plt.xlabel('Predicted')\n",
+ "plt.ylabel('Actual')\n",
+ "plt.title('Confusion Matrix')\n",
+ "plt.show()"
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": ".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": "3.11.2"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}
