{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.11.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"gpu","dataSources":[{"sourceId":34478,"databundleVersionId":3437841,"sourceType":"competition"}],"dockerImageVersionId":31192,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# Two-Stage Snake Classification","metadata":{}},{"cell_type":"markdown","source":"## Problem Statement\n* Most people can't tell if a snake is venomous or not.\n* This is a problem because if someone gets bitten, medical staff need to know what kind of snake it was to give the right antivenom.","metadata":{}},{"cell_type":"markdown","source":"## Goal\n\nBuild a two-stage image classifier using transfer learning:\n1. **Stage 1 (Binary):** Is the snake venomous or non-venomous? (Basic cnn, MobileNetV2, EfficientNet)\n2. **Stage 2 (Multi-class):** What species of venomous snake is it?\n\nThis mirrors how humans think about the problem. \"Is it venomous?\" comes before \"Which species does it belong to?\"","metadata":{}},{"cell_type":"markdown","source":"## AI Use Disclosure\n\nI used ChatGPT to assist with:\n- Understanding data augmentation and class weighting\n- Explain book explanations\n- Debugging errors (EfficientNet preprocessing expecting 0-255 range, label encoding for multi-class)\n- Researching pretrained models and EfficientNetB0 architecture\n- Herpetology research (venomous group, species identification)\n- Assist with visualizations\n","metadata":{}},{"cell_type":"markdown","source":"## Approach\n\n* Start with a baseline CNN from scratch\n* Compare against transfer learning models (MobileNetV2, EfficientNetB0)\n* Use class weighting to handle the imbalanced dataset\n* Fine-tune the best model for species classification","metadata":{}},{"cell_type":"code","source":"# Setup\nimport os\nos.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'  # suppress TF warnings\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nimport tensorflow as tf\nfrom tensorflow import keras\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nfrom sklearn.model_selection import train_test_split","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"_kg_hide-output":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Stage 1: Binary Classification (Venomous vs. Non-Venomous)","metadata":{}},{"cell_type":"markdown","source":"## Load and Explore Data","metadata":{}},{"cell_type":"code","source":"# Load metadata\ndf = pd.read_csv('/kaggle/input/snakeclef2022/SnakeCLEF2022-TrainMetadata.csv')\n\nprint(f\"\\nMetadata: {len(df):,} records\")\ndf.head()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Extract genus\ndf['genus'] = df['binomial_name'].str.split().str[0]\n\n# Filter for North and Central American countries\nnorth_central_american = ['US', 'MX', 'CA', 'CR', 'PA', 'GT', 'HN', 'NI', 'SV', 'BZ']\nmask = df['code'].isin(north_central_american)\ndf = df[mask].reset_index(drop=True)\n\n# ALL North/Central American venomous Species list\nvenomous_list = [\n    'Crotalus',      # Rattlesnakes\n    'Sistrurus',     # Pygmy/Massasauga rattlesnakes\n    'Agkistrodon',   # Copperheads/Cottonmouths\n    'Micrurus',      # Coral snakes\n    'Bothrops',      # Fer-de-lance\n    'Lachesis',      # Bushmaster\n    'Porthidium',    # Hognosed pit vipers\n    'Cerrophidion',  # Montane pit vipers\n    'Ophryacus',     # Mexican horned pit vipers\n]\n\ndf['is_venomous'] = df['genus'].isin(venomous_list).astype(int)\n\n# Build full image paths\nIMAGE_DIR = '/kaggle/input/snakeclef2022/SnakeCLEF2022-medium_size/SnakeCLEF2022-medium_size'\ndf['image_path'] = IMAGE_DIR + '/' + df['file_path']","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"3% country code is 'unknown' total of 8,487 images\n- seems insignificant number, no need to add them","metadata":{}},{"cell_type":"code","source":"# Preview venomous snake records\ndf[df['is_venomous'] == 1].head(5)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Quick stats\nprint(\"NORTH AND CENTRAL AMERICAN SNAKE DATASET (ALL)\")\nprint(f\"Total: {len(df):,} images\")\nprint(f\"Total Species: {df['binomial_name'].nunique()}\")\ndf['is_venomous'].value_counts()","metadata":{"trusted":true,"_kg_hide-input":false},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Venomous Species distribution\ndf[df.is_venomous == 1].genus.value_counts()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"most_common_names = {\n    # Venomous\n    'Crotalus': 'Rattlesnake',\n    'Micrurus': 'Coral Snake',\n    'Agkistrodon': 'Copperhead',\n    'Sistrurus': 'Pygmy Rattlesnake',\n    'Bothrops': 'Fer-de-lance',\n    'Porthidium': 'Hognosed Pit Viper',\n    'Cerrophidion': 'Montane Pit Viper',\n    'Ophryacus': 'Horned Pit Viper',\n    'Lachesis': 'Bushmaster',\n    # Non-venomous (common ones)\n    'Thamnophis': 'Garter Snake',\n    'Lampropeltis': 'Kingsnake',\n    'Pantherophis': 'Rat Snake',\n    'Nerodia': 'Water Snake',\n    'Pituophis': 'Gopher Snake',\n}","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Visualize Venomous Group Distribution\nvenomous_counts = df[df.is_venomous == 1].genus.value_counts()\nlabels = [most_common_names.get(g, g) for g in venomous_counts.index]\n\nplt.figure(figsize=(10, 5))\nplt.bar(labels, venomous_counts.values)\nplt.xlabel('Snake Group')\nplt.ylabel('Number of Images')\nplt.title('Venomous Snake Images by Group')\nplt.xticks(rotation=45, ha='right')\nplt.tight_layout()\nplt.show()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Class distribution\ncounts = df['is_venomous'].value_counts()\nlabels = ['Non-Venomous', 'Venomous']\n\nplt.figure(figsize=(6, 4))\nplt.bar(labels, [counts[0], counts[1]])\nplt.ylabel('Number of Images')\nplt.title('Dataset Class Distribution')\nplt.show()\n\ndf['is_venomous'].value_counts()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"Shows clear imbalanced distribution. Will need handle this bec models will just predict \"non-venomous\" for most or all snakes.","metadata":{}},{"cell_type":"code","source":"# Sample images - 1 venomous, 1 non-venomous\nfrom PIL import Image\n\nvenomous_sample = df[df.is_venomous == 1].iloc[0]\nnonvenomous_sample = df[df.is_venomous == 0].iloc[0]\n\nfig, axes = plt.subplots(1, 2, figsize=(10, 4))\n\nv_name = most_common_names.get(venomous_sample['genus'], venomous_sample['genus'])\naxes[0].imshow(Image.open(venomous_sample['image_path']))\naxes[0].set_title(f\"Venomous: {v_name}\")\naxes[0].axis('off')\n\nnv_name = most_common_names.get(nonvenomous_sample['genus'], nonvenomous_sample['genus'])\naxes[1].imshow(Image.open(nonvenomous_sample['image_path']))\naxes[1].set_title(f\"Non-Venomous: {nv_name}\")\naxes[1].axis('off')\n\nplt.tight_layout()\nplt.show()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"### Dataset Cleaning","metadata":{}},{"cell_type":"code","source":"# Data Cleaning - Remove corrupted images\n# source: https://www.kaggle.com/code/parkjohnychae/check-and-remove-corrupted-files-from-the-metadata\nvalid_indices = []\ncorrupt_count = 0\nfor i, path in enumerate(df['image_path'].values):\n    try:\n        img = tf.io.read_file(path)\n        tf.image.decode_jpeg(img, channels=3)\n        valid_indices.append(i)\n    except:\n        corrupt_count += 1\n\n    if (i + 1) % 20000 == 0:\n        print(f\"  Checked {i+1:,}/{len(df):,}...\")\n        \n# Keep only valid images\ndf = df.iloc[valid_indices].reset_index(drop=True)\nprint(f\"\\nRemoved {corrupt_count} corrupted images\")\nprint(f\"Clean dataset: {len(df):,} images\")","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Data Preprocessing","metadata":{}},{"cell_type":"code","source":"# 80/20 Split\n# Train/val split (stratified to maintain class balance)\ntrain_df, val_df = train_test_split(\n    df, \n    test_size=0.2, \n    stratify=df['is_venomous'], \n    random_state=42\n)\n\nprint(f\"Train: {len(train_df)}\")\nprint(f\"Validation: {len(val_df)}\")\ntrain_df['is_venomous'].value_counts()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Visualize train/val split\nfig, ax = plt.subplots(figsize=(8, 4))\n\ntrain_counts = train_df['is_venomous'].value_counts().sort_index()\nval_counts = val_df['is_venomous'].value_counts().sort_index()\n\nx = np.arange(2)\nwidth = 0.35\n\nbars1 = ax.bar(x - width/2, train_counts.values, width, label='Train')\nbars2 = ax.bar(x + width/2, val_counts.values, width, label='Validation')\n\nax.set_xticks(x)\nax.set_xticklabels(['Non-Venomous', 'Venomous'])\nax.set_ylabel('Number of Images')\nax.set_title('Train/Validation Split (80/20)')\nax.legend()\n\n# Add count labels on bars\nfor bar in bars1:\n    ax.text(bar.get_x() + bar.get_width()/2, bar.get_height() + 500, \n            f'{int(bar.get_height()):,}', ha='center', va='bottom', fontsize=9)\nfor bar in bars2:\n    ax.text(bar.get_x() + bar.get_width()/2, bar.get_height() + 500, \n            f'{int(bar.get_height()):,}', ha='center', va='bottom', fontsize=9)\n\nplt.tight_layout()\nplt.show()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"### Handle Class Distribution Imbalance\n\n* using compute_class_weight from SciKit Learn","metadata":{}},{"cell_type":"code","source":"from sklearn.utils.class_weight import compute_class_weight\n\nclass_weights = compute_class_weight(\n    class_weight='balanced',\n    classes=np.array([0, 1]),\n    y=train_df['is_venomous'].values\n)\n\n# convert to dictionary format that Keras exopects\nclass_weight_dict = {0: class_weights[0], # non-venomous weight\n                     1: class_weights[1]} # venomous weight\n\nprint(f\"{class_weight_dict}\")","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"* class_weight_dict = {0: ~0.6, 1: ~2.6} (roughly)\n* When model misclassifies a venomous snake, loss is multiplied by ~2.6\n* When model misclassifies a non-venomous snake, loss is multiplied by ~0.6\n* this should force the models to actually learn venomous features\n* penalizes the model more for missing venomous snakes","metadata":{},"attachments":{"c8199293-1f33-4861-8252-c14259010bbb.png":{"image/png":"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"},"4f2b688a-85d1-4912-a49f-40af20ff0fdd.png":{"image/png":"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"},"f5b4c75d-8930-4087-a067-40348749e38c.png":{"image/png":"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"},"0158fcc9-d507-4b86-b7eb-3d0fa3a03f35.png":{"image/png":"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"},"d0139b4a-1dc5-4b8c-97c4-a42ecbaddaa5.png":{"image/png":"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"},"340763e5-f508-49ec-95c7-a4652a8a734c.png":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAdQAAAAZCAYAAACIJt+yAAAAAXNSR0IArs4c6QAAADhlWElmTU0AKgAAAAgAAYdpAAQAAAABAAAAGgAAAAAAAqACAAQAAAABAAAB1KADAAQAAAABAAAAGQAAAABhHQ6LAAAP3ElEQVR4Ae1bCXhUxR2fecfuJiEBJFyeSBUhCFYUQa1UAoJKSQLKBolaxQMqllKtVYutUUTFT21FwQ8VsZxhEZKAIpgDQa2fVCl41bPiAQhymHs3+96b/v5v88Jms29zeHyf+WY+yJs3M//rNzP/mf+8WcZkkghIBCQCEgGJgERAIiARkAhIBCQCEgGJgERAIiARkAhIBCQCEgGJQNsQmD9/o3fF2tK+DVR8xdqyk9rGoX2tSc6iRRuS20f9w1OtXLkhPRDYdEwizlQfCLzYK1Gbn7IuEAh4qO8WLVqk/5hyVxW/fCzJaqsM0HQKBDZ2byudbN9+BPLzt2iB4k2ntJ+DpJQIdDwElJ/KpO4n6Od6dPUjkreqsOTXHl3ZHQiUnPhDyw8UlQyKnuiQU9K1d7L/h5bTXn56Ssp84fPMSUTPvZ57mS/pkURtfqq6pes292C+7ns8mrozrdfJZ/yYclWu7+Le7r9vswxv+kPM53uhzXSSoN0I9B8sBnLu/WRxcXFqu5lIQolAB0OgyYJKUSTs47E2NkQNzcqpHerUeDTEyy3aUMNHXhVC9Pf7L/oyVlbMO4/Hg2S6RZ2Ca3/hzDMrmg+3Ija5RFhxZUTTUz6WlnbosW0i7eJHw0uWLPFRPVRphqNT5/CzePM+cOqcZ7T8hnwzvkuXbk5x2sc+G/rNKXbFAAvpOUyItLqK+t5XZI95yyFww5/qY3g7JPYzkd6mGRx6+JvqJx0Ct34mvKL7I1xdO5sF67IdOucZ3cYpa3jyQGHZOauLt6yyNwwxlfFeoQtFzs0wprYN9jarI5poe2P5Rte59WFLPGJ50nssPrFtEtW72ZJoPsfyd3SIVw7+neKVyzKJQEdAgBes23y6ouo74Q2KBePjORNHLC5m5WaNWmUPfm96AQy9CMvA10zwhZOyRz5ylIavE1xcCtr9Qlgz/NmjNy1fu/F4j54UgBMexpEEEwEWPJhn6l0v0FStZFLWSD1C79kerKjv7uus34Q15l4bTCF08HprUnbm8NXry7Ew8lnQJ4UJViBCwTv8/ku/XVNU9qDg/I9g7QHvXfXh4G+uvOzSr4l+dVH5fEXhkQhHsO+ga9c167d8DF0OC86Oh/6dubAWTMoZdYfd3kWGrQv+rCkuuws+dDL+10PWYLjTzTDoNbzfhOrOFrMezc0elU/tA8Xlt0D322FLD7TdAZ2n+LMzP1qx4oWuempyId4vgD17gZclGHshNytzxqqisvNUzp+GLRlCsC2CWfngtw22LwCvLv6szDzi7aTA+i3vwpYD4DMAMpLBZyHqTkTf5Ni8mTXdnzWqvGBt2RBV40uhy0C02wN97/ZnjVxMuKua5x3QFcOOYcDnWDecSWZgfVkmZ0opsmjOgqYlLlQ5Oz+enQ5W6BuFC1GOPryZeFBqrd52O0s84M/JXBmvnz2mUovofQPsOg8bMhOsFwPj6ZA9G2N3KPLZpIfgSi5jIgTMfwlsXglX10yeMmX8wdXF5VOBUz7o0lGexCxrXl2lMefqq8fWBNaXF8BIC3pPQT1bU1Q+lymQE/x2PGvjHMC8SUJUvxJj9xLgZnDBCw9/8/HUadOmhVuNhQuPtJ6nDFUV/jLGhr0wwd5JsP0+2H4afSqIhw/MgRq08Mevx7gYGPEBrZ/PlpaOscR3VorKtBQz+XhFVUuB7VOGsErijumi0tEYGssx1ntiTB4G9tP8OaOfJ71kkgh0FAQQoeoY/1zFjDsYNsz+MAzOVlkMp+CBI7kHzqtv2LAGGCa7ERPhIVowHRq42X2WYQ2xBN/OmTqDQNH1pDw4/ZOCZri3YRpwyDzX9HY5sylgJJMlhTsHlZrDXzyO6OJELMhjI3rwVQXrS4YrjM+1mDm1PiwysBiexX3ePNrFw2H/Cf7hVhE0umCBTPNq3msc3tW8ajYc7XrIf1aE6vo55aD3MVNMxEIwnSnKrXRM5SbDobGfgifBWZ2EDcaDBrOGQN4I4HG/yc2JTFi3wEHcTrt9fPs7G7o/LASfKYJ1PdDuCyy+q4mHlpJ8D977At8RAvZAh56ODJUrtFnZWmfU98TWYydsfsCpi/cErRd4pmFJHo8F+G8K53eirDJcb5wJ/u/DcU8jOkVTnrcYe1sEQ92AB+SLp7Dgn0b9hmqsG2KvaZoXtohB8NBWrDFTbQeIPqIPqG52wsYk2HwyM60H6sxwZINEyiC1Vm9qR33l1s84Kh8NO4cZInycaTEaLzcWrCs91ZbNeOTbK+khWB8cBTxqhMOnIz9IS0q+hBY5hbOFsGdOnWmeAJv2MIXtocXU1lGItRhbE5Yv35hGKiN/lWWJwvbMAeFNz8cYHVIftjIM6Iv8mC69T73NltPKPnTjARs0jIHGkweMOWxCuX1HwBUfEozkXh/xAe2Zz6lW0i8iiynbhA3avW5jGi4GY5N/sG93nQ8DcBXmzl8jWsm/EoGOgwAW1EiqrwrelnfZ6P9h4/84LXaW1rkfHNxwOOPOuq48pKrsd9TSq/qyGkhYpai6a/LE0R/jILMcXhMOGzkRXga6J7yq9mdNUa+BQ7e4pbpeGLn22muDjFVXY4ItgLMs8mePfExlynDipTDlGl3nj8Bh4hKPkp2fPxL+iZcgj29m2nNwiveJ0KGHqC2l67Kzq7AA1iJbR9GsXYg/WHzm+yeM2l5tffciXrUUM7WvmwyHJur5GqLJwJTs0e+g7FM42VsmZ43+t8HNjbDT50k97gSNq+dSnT9n5Gpbrmk8Cic3mBZuOMCzEJVuyM3OfB0RfAnZSLzpAg5wPgH2nOnT9IVw/MMo8mrpCNJi4kn/hMy3DcN4ifjUG8H78y4f8wnoX8FihCjlxV7gezLit0f9/osPI+J9GrrsR9R4HrWnVC2q7qB+awkDv9+PoJRXwgEaZJdQxDkgj2sn8YVz3wacl109cewBeo9OLekd3datny0ztB0YHdGY/qaisonCEBfCjk+iaRvyr1GUO+XysR9iPOzCwtu/yuvVMA6wIRP7kozD36HvQtDXPoYnGowjbCRFjSfNO5G+8aOkp1lbgyiz7XMA8oC1WEbzaUpW5vuYGwXYLP3K0bM1WLTEw+EV/WwJn5bq2zOfuaKVok+2v/+fbTdgTPd2HdOcbULdhb36+LZhs7dfhMzx0brLvESgIyDQuKCGUsPw9XQ+hWEfnTh7A0ebz9F/7GCvC3MLR56xidPxm51Upt0Nx34HE8rniAZWYLLVO3VuT+5NfwLHY6nhqpqpThssxPsb5Qp2m7Cs+fn5+QoLWddblpUDZQ9C1QXYyd/u0LT0TAmHsSBjOeaGbWM8GQl5cPuY0W4imNJoMwpU2jg4tKai2PnUULICh64gWmpsi8XJrlNDODyNpAKyE/bMQ810n1FBGwIKI536hmZNH7pH2LYcLbVsGaZXt/tUURDDNSRsPIRQlMa+dsrp2UYMXO2M5pko76Z3NI1bP1ceMA6Ew+YwnA78DSHkAK7xUjqujKaNk7dxoc0W6g6qTN2Ai09HaOMVMuqXO+2xeaBj/eUYt3mqol4J8F+gY2K7vo1zAAuzQv8c3tiUoC8EnQ40SYmwaImHjVETbjiqaAGfluqPsmv9fMZc3w2sBvYdfG56wjEdDBUawjgLY/xFbPCmYkNMm1uZJAIdCoFGJ9uJpc5atnZTb40pMzH5ayu+3f0RFsN/IfIZhHXobTPM34Wbf0wX/KyECHB2NuqL939Z9wzan4LJ5qX2cGTk0DQ6ZqR3J+FyyFWQMxWL5PW6XhumizSGxd9AfTq3FKtiXx2iXzGFK8oNGRnnH899ytdY8r3/3bF1GlaD7TBggMOLngoTiFLZoH+uK+0WXR6bd5MR264178QLi3u/QFHZhCWFhV00zv8AHd7HpasKRES7sJSNW7m+fODq4rIR4DeGePr9o/bAgX+J+n51h74qRb4n8gsY66oC82o4/AH0E5vWyI9ug4tDe22+ijaTji9x8ea3qO/FLJMwbZLaikEiO5sw/p4vbv3cpWfyTbquvl4tKgtDRoi+L2uKop3aGnH0sxr0UU9EqTglEHOYIW42KvcdiaY1mHgWY2sUyq41ubXErmvHHEAE+gbGfR7dYseRdD+M0Vx8OI+zEY2W3jTvxoOH2GfUsv+QEePpWzl2h/bJEZW1hE9L9cSjWXKZz047UxhZwLPSp3o243wK49ZlTPs8L6lCm8nqD87FJu7vmLv94m0KHL7yKRH4OSLQuKDiSGqCT/fuhSPIwWC/hi5Q1H1Xn4846VOf7vlc1/hn2Fm+Xsmrn3eiH73CZ0dB+P5nHY3QOI4v+eTefZJwlKtMgAPDk7Fvvgq+h4X6ZUWoWxUe1qiM6LnKr6S8oqgl3Ne9CpeUDl+Rk/kmdLgLx6VLj+ndCcfB7AxhWvl0Kxj8noBjLMw4c0QdyE7DJYgnid5JcETLsUBnJKnKfirDwtYYyR05kkT6CkPVwm4yHD7OE3g00jtl9AybkXKKMogXGN8J5/ZMJ7ULRT8DsIDlUjtTWHfjaLFGZ/w94BeAPoeonBJiSD8i0VGduvWpxffQf+C73WxahBHQFgD3XlpKyuJIy8jfaFtAa+vlgUeLbkN5RLyXI8IZ4U1LqkDkPw9ObgYugHwQ22+txcCRm8hOkuuGlUNPbRLp7bRz62deH0REKfamKJ0PejXPV+jcV6t4dWms7Hh6BINqJPLn7CIoOg5fItekdOuDzwdHEx3rA6sdGD8HPtyhbqKa9syBUIVxN/T8gPu03aqmvION0jYeOjiP+Dk2Uj4RFm487I0YvoljvhYoOj5/4GQE/CPzMAE+JM8Nv9hx0Zr57NCEK0SFUV1zMeaoD/YWuY1pGD4XfOmCVy3m78PogwexoNp6k24ySQQ6BAIF68rPwE1YQZEhRVcwCv6oaaIr83T5pmmp+1tb27tzYrzhkkiTJqQL3Z5FYTNdmzRs3UtcGa0jbdaKfn6S1KwUBYl+r0d1P8ZuPdFPW2J0bCsGrnbG8P1er279THjFGxeJhK0pLp0cKN5SZV+2Q0Nc0hpnj3v6nW0rUnvGNP1sJ8FPd1ohNfJzrXg8GuyIO/5bwqel+ljF2mN7vDFNY5xuGhO/WBnyXSLQERDAz2bKz6Dr79iJd3JuPHYEw6QNEoFoBOgSGL7vb8EWrBdOUA4hck9D1LRoUk7m7Oh2Mi8RkAhIBNqLAKddY0bG0B5+/7hv2stE0kkEfg4IIKpTLa1rP1xdNnAz+lPojBNYmSQCEgGJgERAIiARkAhIBCQCEgGJgERAIiARkAhIBCQCEgGJgERAIiARkAhIBCQCEgGJgERAIiARkAhIBCQCEgGJgERAIiARkAhIBCQCEgGJgERAIiARkAhIBCQCEgGJgERAIiARkAhIBCQCEgGJgERAIiARkAhIBCQCEgGJgERAIiARkAhIBCQCEgGJQEdH4P/7gjQIDJD/CQAAAABJRU5ErkJggg=="},"8dc8c821-efc1-46f1-83a0-fd28a6bf325b.png":{"image/png":"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"},"24d8a2ef-972d-42f0-a9b3-0d646d92d561.png":{"image/png":"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"},"13a57cc4-e9f3-4973-aef4-2c27fa2b9905.png":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAdQAAABUCAYAAAA74MYAAAAAAXNSR0IArs4c6QAAADhlWElmTU0AKgAAAAgAAYdpAAQAAAABAAAAGgAAAAAAAqACAAQAAAABAAAB1KADAAQAAAABAAAAVAAAAADBfpJzAAA44UlEQVR4Ae1dCZgUxdmu6u45luVYBAQ8gjERBA8Eo6iJBwt4hl0WZLkMAZJ4RGPMYeIVReIdNcaIiQeiIrAsyrKIKLC7KF7RX/GI4h0RlPtmYY+Z7vrfr2eqt6ene2YWliOx+nlmuruOr756666uqpcxdSkEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYXAtxmB8vIXDiovf67LAY4Bn/5MdbdMOpaXL243s3LRIZncNMduxoxnOxI25IfklpeXh5vjv7lup1ZUFMyYV3NMc/19293PrHj+iGnPvNA1CIeyOQuP3dtpR3nzoYeebeXV4f77F0SmP1N1pNfc/T6jsup46Ke7zdSzQkAhsP8RMHZHBR4JTxKctYffMbvjf1/4mVmx+ExD15ag0exWWjpopW+YUe1qg2l9YXe+17587uLjGDfrSovP/dxrF/Qeys+/3+JsC+wv13noPR7pdAee7wlybze+Rvi40qGFLwW5CTIvL68+lOl8ORP4MXZqeeUL32csMo0xgfjw5ZYpxo0cWvhekP9M5rPn1owVnN/MuOjEBXtONGwcV1paWuf1Uz635kamsZ8hvLAQrGxEcf/feN1438vnVl3IufZPS7AxI4YULvTa5/JOHSEWNR5HXAsR9i4hxNQRxYXXw69w+y+rqD5b07Un3Gb18cbehhb+o66Ji2dW1Jw9qqSw2m1fNq/qJE3oL8XqWh0K80a3XUs+h0Pa4vZdW90GmY+75XY6PHSqoeuLYRZym7ufDa6XsWgnyldT3ObqWSGgENi/CGi5BI/ecJ7bHRoN7n6nZ+pZ+/Xqp06dGn3ooYfSKge4be2VsSfvybAdvfTYlpdR0R7taUz5lMrKNuWVS/5eXllzAbe4496ro+DGdZyFr8pFJ4ojubNYkzzTrD9p89raf7j9p+EYCZ/FdVbudpPzc4SfwNHqsYYNP7L98MiTSBXBTHMA52K9tpty0Vh9B52lqVxYs5iwShDEIBbtQI1VyoXGdDTT+LVciEeYED/XuLiYGssUR54XYP4K41oZYOoAzSMe65xfedS4AWGeLEw2QVjiLo3za2dVVp/uFaBzhkZRrDYta5j8RePbtg4v7n8ZdHhN00SaHzSm1womZo8Z82PqGMmLT5w40SkrTz65MF9aNOOekvfIH0eGobs375FZxktYf0dSX5fRjbJUCCgE9jkCGUeoZc9U99UMPo1z3mt2Zc0mVGJXDx8yYKpby6eeWXBYOJRXDrt+nOfx8nk15ax+4xjG2rbDSPZZVFynoWEzUZlOKS0uvHTW3CpU+NpTkNkZbjczy7qkdMjAp90y6ZkqsF59z1zHhfmr4cUDy2bNrf4917Q/xxsb+4y+8JyPZ1XWPIHaaCdr2PgHFulYBi+DZs+r+RojtgdRYd5jGe16aoy/icqvk9GKHR4Khadh+NIbFXkI+mBkZz7MmH4aKtwe0OMtNEUnwP+LdfHY6IgWukHjbCSqPDa7cskYyKPReNo1ffr89qE2rSrg93S4Wy3QCiGM+eRQ16PPtu8sbsfjDD8chca+Qr+E4g3MltTB5zCM2BaQX2A0UOP6MHqWF2SvLy0qvMl5F5Cgse0YOZpPzll4MLA41TStc0YOHfQKRlk36Fx/c8bTC482jHBvjDT/hsHWjzyjbQ69+sTrGj6/6KLzt1MjEW0TOhoN6CnAaNuH7yy9DmlgId0eRSNYjHBvkGHbd85LkObThxcPuAXvHCPk3qzRrE1x43lBfBcLEb9MMONttxXy1gzBeJfS4v4YcTZd1PlpxVp9/5N3jH9PnNg/To29GYprliVWIV9dN6Kk/zMTJy4xevVhdzKmfRc+lzb5xhPnnfD/pRU3N+iaxkcOHfgZ3pFEsBJsK/RxGkkyS07BnstNdh69I13+Deeb4PZYi1nDgdc23eDouPBjkGe+gZSbSov6T5k5t/o0XeOLkD52J3F2ZfVwxOcW5PcelAa+eY+j6HF+CeT8GXm23ey51ZNRtq6hcOmizhdmOFZbXFwyoqjQ7nRBn+WcWf8Q9ZumikjHB2g0PbJo4P8lfKh/hYBCYH8jkFKheJXRDO1pFPovzLjZA9XQQ4Jr9ydHgo7TUChvDCrWbvVmrGvcjBeikhphRgr6iGh4IGqufnERO9S02DloQC8um1N1FOf6JahJlq9ZURdFRTUTjeufHGGuB6rMMfpZijAHkzEa02I0GlEjFBoMHXRU+j+2GKtCY3ozdDwyFrd6xk12MUYXd1EjjxkzcpMXa1evGeHQLRh1rcFo7iA0plOg46bSIYNQWaK6FPxgDC3vxeMxmLY8Nk8LnVvLd1wPd/Ng+ZhoqOtO7vwuI7/VzagMj0Q8z8DQcAL06yzd4TmCMO2Rqx+OO9torwlmTYC+m1l93XfWrayvln4R2TrBrU3uH0Yk7hET2lLWEWHbU5JhYSC+jMWtBpr+ZY1bTfuuG/phJjNXw+iNODN2kJ280AF4SQ9pb4Tb5H2D0d2lkYLwFRD6JJoYkvUx4U9ukQYfIZ1s+dKvfefsaODbHw3uN/jVI7p/j8VEQ4obzws6JjdL3N1WmPH4ALJSGllqiNqytisNbrzZqy97Z1ZlVX8WMR43tOhP0fG4H9PLT5CMY04Qo5GKmjDN190y7Wc0qEjrYeGw8S6w+BiN1zvU+SA7dFAakG4d3X7adY0eT3mmjlkfkDmlIWYB2sYxUt+69otXKB0BytuivqED8gfSXjyMuPdA58tAOM6oFXkqhHf7+2jGvEf5wxRDEc6lTNN+Rx0IqQ9NsUPHBcB+ApmVV1SfjPB6NMQaKmCHjgv/ilv6ydK9uisEFAL7HwFtVkXNMIwQHqZf+bzqcVIlWnSEyuW7mMi8Dz37T9esrJ+EUV0/FGa7EpfuLBGbhgrhgYhu/MHQ9HGoaCwU9E6W2fAmGrotBgu9genHoSIuzrJHCJy9ALlndTkiuhSV0zrRYNoNppTnvmNUUIH3c8vLF6BiZKdgZIKpLj4YvfNTUCPmbVu78wUyR0XTLhTS7tJ1dhn5j+jRIrccdAai0GsLVURo2LfIhi7p5pXSIYUzMJr4BJXU+zDr8bPi4h0YYezCc11p6fkbUmS5XlCRnohK71l8v3u1tHjgYlTQc13W9mMQjuP796+3BN+OCjNOYVx55flNjZFg7bjQurl/GIFh+jJxUeOCkSQ6CeJFMhGGaXeMdMOwG8G6ukaoAnP0RkYNGfgy4lY8unjgOjKjC98Wv4fW4tTGHTsPhsMrMNS9UcPsA7PEHcAJs7foWiQvi2tQE6NhzwXcjwVeR8LppXBdBHk/COe3+r3HWU6vGIHdNryo/9Vux4ZhDIES/xL1G9qj0/GqxrX56FR0EKxhunRXXlk1CJ2Ah2E+kfKoNJd3dPCew5TBuA+XLW0TN+uPRJ75XlQP/ZTsMd2/GHnhF+hM/EK61yytK+JS+9OhAzdJM4uJBwjD9u0Po2nq72Ka+d7S0nM3jyge8AjyImZQxGnSre89Q95DWPeXlgx4s9ba+hz8GvlmmyNTZMTZw3gfRN/LhaaNQvhVFw07/+ukm8+RNF1T3KsXhYBCYL8igFlRVosGZwP9qIKX2piRUKISNc3NjhnjBTQ6pHc0BKhTMbXJjJvQQF2DQcKXGHFNRyNqN7jb18fXx2JmP3yHuxG9/J7c4FVYPXkMq2+oiIv4iRgVPocKaQLGnFSZ+F67rB3zUQm2YZHozah8PjDjsQdRuZ6GgMci9IWXXDKYGj0ow15H4/s4/aDYz2LcesUtEH6r8Z1tLKbMNuP5UrQ0NE3pd5l+hkFmgEBDPBw/wMRu0Nzus+HodiufIWQThucfpfyE9YW0x5RkPRqDnZpg9qhIb9DX2HYWt0fT7TrlH0XvGrcS5tJj8j6yZMAXHy5jefSdkEZ6GDkegineg1G5T0NrugaNte2fnHNuHYV0olFuyoWGtAY4PjFiyIBn7cVFQrwGfz1THO3By/J3X75r+TsvXUCdIDS4lw7HdOrw4sLectp61ryaUsRwPhrMW5PTzmmh6ZqeTyNXGm2PKjlvBRre96GjjZGlsQLk+VrkHYz2mi64aerYNBkzmY6ahvmW5IW8iO6H5nQ2EI7z7LjJIe/lx2Jxcq/xuF2mpN/kYrXPeISNh+ARaNwfl3a4t0E/J+Z6V48KAYXAfkbASK60XOjVY1Tx2auxeGeV0I0rsT3jt/k873KMBm6Au4OoIkJjchptE0GF9AOYVa5bWfdol24RmvaMkKyCzq1+iWna39WaW3saZt4L0VBktaYZR7Eon6JbWJnauOHnItwR37HYX6giklOM5Fde40tKtmJKrQaNx2Wou/5M306h05foqV8sTGus7Q6LS2BXFIvH39ZYKKyF+Pshwa5CK7dMysHnsyJUfsug8zLBzUoMoh076cZ715jAKJUd98Scqg7uEYvbHUZu72GUegG2rjykC9EBdmfjN8vtJhOOaI0RhtYBi3l6LX/3FWeaddSQwjcgg36+V+nQga9jJe4tSI9J5KC0dMA3mGF4Hw3INfiW+TUwvQ4dh1WYXv2Avt9iAvKK+I5dv3MvtKFvkm7hEn+Map8Hvn+dNbfmNya3qlFpT0Anhb5RM5oaRg+sPRq227Hw6Vl0oiaVz1n8qNB1QEvfyq3J9EmA4/seGrInS/E91x1G0LNbrnQj9ZHvuDujZsT9l2jqJyPdbzNFwzTaBmPtMreFQrt2Uthxy5wxqmTQi3BzlabrrZGHxmF2oxsE/AAdPKyEho1g/SwhpmGkOZPe6TLRcTCwYIpWX9MoNGGa+LfTcV7NSqEZVz711ILfhdrklUClLswyX+eNfBtN7h/d94zBwHsVWkbk14S/5uY9nek0NW+UzVt8ysiiQf9CZ+YRYHs3mu5tuzatpBkb+0Iad0NcvpHv6q4QUAjsfwTSetRulUwhhsPBma31gi2oZP+ICvM3GDHUYVCGlZqsi5GfPwVV04MYRY7sekQevqNpJajIcYdpY/1TqHBW52vtNkaM8CpUMC/v4LVVuN+Kyngwvn3uQkN5N+Tc4VN5OmpgdGtXIqjcn7XloiJHGLGdbLv9Xre1cSJkfB4Nhb8MGfwLVECvbue1T8uRRGhbFIMYLCxhvC/i8iNUWH/BqPpLuUcVFVNKwyIDxmjgKVRkvfJ0bZ00895NYd2EBmhniPEPIL8cdagzVYhnR24Qjqxh8+sI/y2u6R/0OuH0Yq/8TO/QbwvCMKQbMy7GA/UenEew8Iafjm9zP4EdFivzExDv0aH8aE7TgzT1bVnW1UibSSGuv4dOw6e15rZJFA6mXYsgezg9Dy8agIVOYjbTjUWIw0tI17d4w6a76412BcDtF0zXLiR3fhfS38J3AWek55br595rhknpm20zjV9n6NEv6Rdu3eqvFDZmS0ajEbWnYS2z8TcYcbZCXD7WdT6P0gg6PkZ+oXMInaGtbtn1W756D3mr1gzpx5O5Ow3pHTMgF2KK94xI27xtkHsn8tXlWFC3nDo0aPAeRmejDB26xWjoaZrcjl9Q3nPL3rIlj9yKuG7E1q6q/wCPi2jrDq0ej+2sfQKy0T9ks8aPH2+XLXsPLcofRtsvk17qUggoBA4MBFCvZL9osQS+K9LUGOqB9Iu2zLRps47LAu92QX4jO0KcVpJKczSgWq9epxSsXWvtpG+H5RU1+BbJfijt5Z03mnM9216kVdo9SAcyR2NfazJx4ciiwkqqpPI7dPsGX3pvHDGk/+Q0QR4DjLhaYyQ9wWOMNsP6iL6bknkSHxpZZLyy4ZjRs8cS30HP03Q+GyuqO1EnR1pj9NTWjTWZ0zQ93FClnPNFq2e7dq0NO9PqCZ8yvzj5gEaka9e25vIbMK3ExeKhFzF6+l3pkAHOiCpLwGlys7gPtPaLK63ebd++Dt+qne//tLL6DXwzno/v53ZnQQrEYi17BItpcOqQ+F4kz4OL7Y6wQBg0DWvjs6d5j4TSd2Is5FuExWUnyxW9WEWMmSI+nKbA7YDVn0JAIXBAICArsv2qTNmcxT2xreEUrxKi0arOtUH1+nW/Yzp0ClrAsejpfwXzjhixvCPq40Mge5vbnd8zVZ4HdW41wmsXZ+JzWqziNd9X77RQi0Xz3kcCbsWCnp77Ktxs4VBaYjVsv9KiAY9nc7u/7LHadzpwG4ZG6iyaVnXrkdDfWBavrT189OjBG912u/O8J3kPW8P+hG//kzDKnYYtOWNl+NjetQJj/N/7bTeTbtRdIaAQUAjsNQRopSSdfpTca7jXwtmXgmkUaTes+zLQ/4GwaP8wRpN5QVGhtQE0ixJk31zz3c17lFftU6E8ASbT/IDoDHtUU68KAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAL7HYHAaSNaSYhVhObYoeesT2rJpz+96PuN+s4NtJ2FzGgRRkwr6Fa76csV7bt06yb3CO73WGVQgKb6zFCbzrQvMcgZLSbpcGj40DHDBv4nyE0mc4SBI+ha52U6FCLIf4KtxsTilgvWBrlprrlXH2KpyY+1brt8+atf9+r1w8Na4jt1c3VS7oMRoGleMyLyaatOsKvMNnuahzNLt8v+bufxbLL97DOVC2/+9vNPZsTi07h91WpaPElsUHrD1o2uhWpB3hxz+hRgGCHLvaWqJdLKCUA9/Ncj4Gy78MYkakT+hsUQdJTemWSHlbh9uc7fMljBo3i1T5fhkQ7FYa7NaNv5+2fiCMFXcR5s67Fjz9lJ7g/YK9zhXCyAohNoOgXpmAvjR5Bf2zzS8S7syzgRz/0yuvOxxPnHLc/k49Jn1rzqUdiK9BQOC/igV58z/oQFWpX4pniQe4+qj1rKaF8iABYkXWiUf87b3WD3OA9nC9iVp7I5bQl7d7lIY2nKUZdQSP+X3qEb1V2VubBBefV2s0k5di2QVo6s5IP93XwP2JRcgyCvaPW+lxEIXHiBc2QXYGn+qYneH540YRdu7L9zCjmOpTsbje6rOMbNaUThXofOaSNfmIf9zCl+QX7ITl60AEc+0z2IoYMOeXe7k88UhleGyy5QN+nGe/eG75Ydq911Pc7nTdlXivDRo0+/YJ6yOGZvMPmk6CP4uVjxPI+2XGxZ+9nzjY2x7p7GlEMnwiPlIvyyLehKYp+W9imCfF6S4fn6o3DhJc2O/Lgxd4v1mnvTSroNyivS3u+OcH3TUboNYlci+yx6OHF0syBJue67ZDdym+X6HBRnv3gR9kFpnpKnkoH7yfDq5cXAm1Ze9/LdXS4sD0uTny7kzw8nbkISLj82KD/3bjluNikyp8udVkH+Ey5z/99TNqXcQ1IuWxoBpxB7BYPEuDM29q/B1oIi7H+bj+X/r2Jz3XsY0VwWE2ZvnA37vr18X7DJlskWgYXjXdjjLFsxCELXmaa4AkfcPW8XsiQbDA6HcNhgiMRZ00PvYrP9HGyCP5/8YD/85djb+YLURbpBdToXRxScD/9fYeP9P+H2IjTkfeCuaqe1bSRNQdOJQJIJBHY4QSbBBEKysP3gNmy6vwLH4+iQ8wn2vB4OlpBOQbqBS/UsQzcWYztKSOpC9+T+v5GQjcNyxfHQayFO3HkF7zi5h7XDxv97cfLORHKHo4NOojN0iTlGsuvAj8OuQ/pKJh8cBOAw+eBIvcmIXwG2SYxJZfLBcQhM+DL5IOw0Jh93WFIfYPeVpvFfUVywhegLHDgxFqP1amBhN+oIG5R1/Cp0mvKxk7JMNNRfQ9PWYEK5A9P/v0HahyH3vcZY/Y9dZ8qysnk1xbrg/wAeXeFvBw6duArnGz9G4dCFbSpldCwjGvHR9I6Tjm7FKRGn4ZxeOuCDTmHCXsvc8gbSLA9coDOoY4f8FqeDPzav/XRCu05H9KD8tCd5JRNrTFA6UnzoolETRlEOuxKM7DQB9tfjSGVsu8KRhmA0QjxfjNXuHElbcoJwI3yAd1+w75xXXll9GfLX7SaLD2RCD4Px4RGkQy+k5RJAOhH5bamtgOfPm4eDykdQvLKlucxTmfK4VIncIg5Zy01G/JPlAjiC3EKrgmxkKVZPLE04LrOvLG8yLLwTOeNxlF9NKzZ41JBzVmHv8Xqc4HXxiJKBc/H8b+xDvp3O8bbD9cGVVoNLNimUC4dNCkdhXu7ELZFWo5Afd6GjeowMT+Phu1BiOcox4k1loPoW6P1D1Cn9pV/al+7HpqTxCM7G5tvoWFDqcIBNaRfi+gtJCCH9o07+A+qyk2ON8Wu9bErSjbrvWwQCR6j2YeqCva0zDYdzv3AQErQfGtdHKMOAGuw8MMd0R57uhg38z0mVkblXWXGrLyqON1BxJzJdRjYYsIHi7Fjyg6Nl3uRMdzJqQiYxxmCimbGVsVisN+0jRSG5D/e/Ilw0aPzkfN5mILkNYgKZObfqdOLLxHFJ15pxqw8yYL3UF5X5zZDhw1TjuEh9wIH8KKjdcMbuHXFG8eRnwP9tJjeHQp/fouH8o91LhTtgY4/ygth1SF/43TdMPkl93Cw6dabVT7csHBaUYMSho+7Q6bgVp2BNaIyJXsDpRB6NjKECjcr996jIfoe9uzj/lreNGJFxbmCQiW5EBb9w85rafOSTJ1D5/Mltj9OFnoGMEjp0AuaoZvhPcBRgRTD+iXT3yxsi0nEidOnbGLN6xe0pdXF2Qdejrk6wC+1ZXgEYgawxQeko4xnErgS88lAejsAxS/fGY7Fj8XyckdfqPPKXDbcEWQW/D07H4nvqWzpxyeJUqrp4Y2dU1u8ivW4nOblcQeXDL165pLkdryx53NErx3KTCX9HVv2ml9CRSGVpcpU30gtltDtn5m3U8QfeOEAtdKfjP/mAfO+wQQXhmolNyiPvMIx7JwkrfrwTnmWCT5iV0NGlcIv2VhuDMoIBR+La22xKMhx137cIBLLNkBoYcT2HzDmIhSMYdfL1o4oHvYtMAUopcT5O6Tkbzyvo6DWpMlXYxPqBEcoSNIRocCknZWaD2S523JBgCrFq4LiHlOW+w81NYy48+zNU8ksR5r9sdhgKl3hNmd5DMrr4MYEYGj8RlfwGOhXJ1k2Iu6XsbLpJd577K8RPSSN0mH+OhuG3dIJNnJsLIC8abnPo4Snufdh1pL77msnHzaLjPZ8YHadTSG8wz4wLhfg9qJgOwltx4sxfHKfHtLtAZPA4OlS3iIZNd6XEMW7+GmnzdUHn/DvQ4TkWDWHK92m4r0Se2RluGxmKkdOZaBA7m7t2YpTZ/LyBfIVjBcU0WjA2uqjwQ+SZMjQsP5L67ElekTJ87z7p6HYXyK6UcGQzGtFZ1ChPNMtztG2cETdRiBHNVPCf/hSjwHm0iAbYIm/xPuAwxHGf+D4PLCQdnVsX77PMb37lA/ilsT/llObuQHxkuK2Tz80rNz4CyIhO/CISDwxP01makn7Q2VhCHMpURpHHngFONJvle2XCFQ18VjYpEkrhUZ1A1IRI3zkU3kfv6WBH4puimjYGI+BT4ezQ+M7a6VKJvc2mJMNR932LAOoif7YZUgNTsWCEYT1RQV6MUdkLtpGJhsOu1Phw2OM7q98FPhT3lYUNJuHU48ft33lOc2OHk4kJBIfGIgpNjDCgKUP76rpy0s3l3v0I9nP5ikPunWdpZt992HWkvviQs1m6xcHs+5TJR4brvgOadcTYQz9U2FfjNJ777QMOGqyf43xfolPbiLSfjFHiH6U/e0Su67MA8kCmidfR4KZNQdJKSpg/Rb10HOB/EdzOd04hygn/pnRHw4wkBM9O8kLlinNwMZWfdjX5SVplzStShO+hDj7pKN3TPZBdye0o8WzrkQNuGxHJLRbXB5A3MApRPqarjNIHdneix3tpNL5tV8I4+F/mN3m+NblEeiSYcnzilS3N00LykZHmxm2Qpdz44g//aEQlBm5pac84Xxt5InF5v3si/6bIyIQrgsvKJkWhuMNLBsuoUwI1HkNw4zECHgXNn3PyPBztbTYlqYe671sEAtlmSA2MvN4CU8c6ZIpCnPVdaqsW3/K6MDrtgNkZmOJNm0pJU59nZ4NJ89NMg4xMIBrvgp5iF5zROiYuGpeijND3zsQVoBsaRxp9Oowf0vlu3aPh573sOp+889raXn3O3GdMPrnoHbf466iyO3JLs7au2/li+y75RP4ewraat3lUA5m1VfTRspcuwcrg7piq7CllhtsejpETPwQN7qX1pngtz9CQT5z2TjoDU4B4LMQZzXAIEB0kFmwF4I8WZ5nj0fOA2Q8QCvAxWAn5iGVAM6I1E+wfHmeBrxnzSgbWGMw9pKUjVfz42ZV3ILtSgCbZcQM3b9ychKnaKqwB2IxR6h/xLXolZgK6121a9VCrg77zE1TSkxlrX0ZsPYC8Naj0nNkXL2uN7deHKccvXscdd/rhVkhbEZTmaVHKgk2a+wAD3sC+CGLtQfwchisvS1OaOC5OAyZnCDO+Gc3nMGS5RH4iGZyfgintZ6UfIjYIwhUDhqxsUrYchFc+p+ZMi8U2ucMzzcZHdS1yLcrHCei+pCxSlPlG6oG7U2hagk3JJVc97kMEUB9lvJAX+fNwERcN1iJySVMuqBQX4lfHGzcuITPZ8yVmF3rHFAiYsRJsG7mwwXj90DtdaXIhEx8jnN4nMirGn4l39Nr9mUCKBtRAl4eRW6diqfwK+AePZGbdvIwfCW0S/+je4tNd+hXDqhEyDYWFfXfcBbDrBDHQ7C0mH6kPEtzR3zLsZ/vdpowT4gZMcz15UNfWIGJnvUGRN5H2qCKtHwDWFWhM6xDFHnFhOQ2YvVdXsPmgBZ3XytBWAt9W+EhKB8SnXDT9ZldsnK//eJlOsx1sd/JGw7b4Tah7lmMl5Ard0MBvKpbyho13tkheycAag+ruVuS1QJakQHYlxFNi7wYkG27kh2j6wMkzFB2I39DiJDCxlmKUNqB1hyN2YV3AfWgcrqfzqGF/Ed6vccv35uGg8uEXr2HDBnyVKc1lOE68smBD7h230nPy7i43mVh7UsqFD0uTWz7yH+oF7WndCP8bZT2GpUt/oODQGXkI+fjKXr2tUagPnHIQhGsmNilPNLbi67sdHvCMy/ASe915DYaw65e/wxd4/AS+tgSbUqBwZfG/gQBtNG+pZeWZEAla5k/hY4Vo2nYQkrW3daORDC3sonC8utNKP5ihPvC/MulGfpMLfRzPmcJyHGV+4F6Z5JzSjlY94tFXV1oxjZ/P1GvmwMg2UxyDfNP2C+8WjCC3QeZBeSWZT9LimQu2fmkSFD6Z7w5uFAbpQv5tnXDYP1Z8zqD3bJdfnIPilS3N3WEFyXC7yfU5CP9c/NMK6dnzllDjxZP5NRdvths3rm5PZO5+D3jmXncUPvib/wN9/hLgp9nGfmWM0jSJWbPlKQ8ti0BapdGy4pU0hYBCYC8jwLHd6drY9voHvLR9ezncA1J8YssR640p8h/vTwVpFa8RCn2Idn0la9h1akuefLY/46XCzoxAymEJmZ0qW4WAQuAAREBgheltB6Be+0Ul0bgRW4na79ZMSUsq/OkH4c+PPL6xK04t2gC5mGFWl0JAIaAQUAgoBBQCCgGFgEJAIaAQUAgoBBQCCgGFgEJAIaAQUAgoBBQC/0UIBC5KUvRtir7tvygfK1WbgUBLUI7Ryuw9oTjMpi5WrYKEYPcoELPJ9rO3GWyYP21irrooejg/ZL9dZoH7UIm+LaqHZks4iL4tHA592kovcJaA2/Rt4dBy0LedxHnksyAmCynjgLgTfZse/b9MuhD1VTikf5LJTUY7opOKRudndBNgiQPWJ7Fo3j0B1rtn7NKH6NvytXabeER/FvtKf4z9nF81d3vB7imhfB0wCBDlGAtN2RN99riMZAvclWezOW0Je3e5o8aVDmpw5OaoC9HD5XXodg75S9LD2UQUjpwsD0QPhzOh/5zirAXSKkUeXuisZhzY8yh+W3Goxdez51X76mnjMG9JBbb+7MDBOMtmVdQM88pS76kIBDao2BuNvVyKvi0VrqY37x5IyqTS1o9OKtHLlS6a7jDPa3rD6b7Ype9+p2caDcBd2h5a2ifo1YPce8NK0ecAoG/z05n09tsjSeZ0BflJ2Gb+T2KXhiv5gh2tCE2zy9Q5dKc1yQjSLZMM8ud3edPO6yYozcldFj2cOLopx7zy6Z3C8DPPxSwozn7xIuyD0jwlzyLgTG7denkx8KaV26372V3u/tfp4XqeIH6Fg1tG4cCbn+CgEJy2pd1fNqemtxsP+zkamYL1ycfimJ5xOJzieVCZPDy1oqIgzZ0ycBBwCpljknxQ9G2Kvq2l6NsoSxFdFk5RWof9A0fjyJp2ODXp76B4u862q6z5LTLiH3HE1sH2aUqCjcY+wk9sirAA6rNkNrVv6GW3GD1cEM1ZS1AJZqQnC6D5k/Gk0QJGUd9KerhcqOTQJxqJX0ZaxYz4f4vo4XQ98jjOKf5sRFF/IltnGIF+iQZzKkhHJsn8Rh2Tg7p232Wa1gWg4VxEHRoz0v74DV/Fll955fkN0p26pyIQOEJV9G2pQNlvOdJQocHAqFPRt7kRRIMZQWPanpliqGDmZRiI/55OlgHbxw9wHNzdOGD8SlFfdzCw+wrjxVkSb/SQfanPUmS3ID1cEM2ZoodzI57+HITbntLD0QgTRyxmpA+k8oYGIiutIo7VDKTnc2L0LaCHQ5k6DFywy2WcMVr9CGXuUPlO93adjvw+bgaOFL2Gpnx5pOM6MFEVq8bUjVL6s6JvE6JdKKTdpevsMoInokeL0mFKMWkeDZUPtZWk0/q20LdJ9HA+7OTSkgFv7rJ2zEchDuWbbY40uE7UVp+XDuk/i8jMmRm/F43t8a5j3NKoz+hbDo7ae5h+xBnaUvRwMl18ac6SkVD0cDI1m+4ZcfPJ/82hh2uG2+aVyyb1U56+HfRwHHzkKI3Ji85dR6fFM7jSk5+YxA6caXwOpsTv1bh2U3nFohOkP3VPR0DRt+VEH5YOnG2ShYbKduNDbSXptL4t9G1e9PJjMftgco3H6ZODLokUyJ2paTb5QZuGVp4Cbkuxqc/Qm65FL3sD/Ygbs6Xo4WS6yIP2KURMQSdozuzg3X+KHk6ikRE3n/xP5/6yDJSAUi7dm+PW8ZelXNoyHcdND98Wejh8elkNWrujZMwxQ9Qd7esa+U53Lb7xU7qDxOPOUUMGvLZuRT0tlBRC13uSubr8EVD0bUIUxeLxtzXwxGgh/n5IsKsUfVvL07f5Zz9QfoA2ztD4veVzq0t2iu1LDM5/ja7zh8SignNZfb2NGFK4EBb0c66WoIfLRO2GKibkBJbhIZMMrujhtqIH9Zfm0MOBPvCwTPSBGZIizUrRwyUgweeVBeit/r68ovoJzBQdA9OjwCz1PL6ThjG1+0DcMmeUlgx6ETNAr2HO99byeQsncBEeCnfoRYJWUF2BCPiNAtyO0UFX9G1uQFAhOLRPbnM3DRWZO+4CqK2+bfRtaCQd3LZsyaNRqIjrRoxo4/ByLdP4o631gi34FtaTWeYIia2DozQIuLcUPVwQzZkctToUhbtDJajo4e5GwbijOfRw2egDZXYIyifucqno4RJobVlX+1fQHr6Gid83BNcmC2bdSlSB9Ua7AlDcjdZ0/TRyKRpM0NyJtpyF/wP3N4KL+PrRRYU48F9d+x2B3aHo2h2lg5bhB209oTD2tm40xUSrNCkcb5yS3wpRH/hfmXQjv16qtUxh+YeQZrrP6duSGnD0kLGYq+WvTBgGhRaUj4Lc+5kHyaCRANynpXkuaeeX5n5hSzOE1WxaPQpDTovaOh0A9HDNoZKTcQ+6B+Ef5N5t/r9ED+e3BQvYpBEL0FYomR/cWKjndATSCnW6E2WiEFAIfIsRUPRwrsRX9HAuMNRjGgLOYQRpNspAIaAQUAhg5k/RwzVlA0UP14SFelIIKAQUAgoBhYBCQCGgEFAIKAQUAgoBhYBCQCGgEFAIKAQUAgoBhcD/MAIZFyU99cyCw7gwUlZehiy+nvYIHiiYtAQVVUvGBavk8sxQm86jSs5bESSXVp3uTeqroHCVuT8CtBJ2X1KF+WuhTFsSAVpVj2MJwqWlF6xtSblKlkIgEwIZ96GGQ3kzibItFDI+lj8R0fcrhY9NKeSmVgK9kcFDj2aK5D612wN6uLS47VPFD9zAcLxgIXWcpIbed2me6z0N5xzpuXKRv6e65RKGcpMdATcdW3bXyoVCoGUQyNigUhCWsO4B84cuf2AIeUwGHUTVRPbe/Uw0cpP+3PcgqqjEqMHtMvHspVZyU1F5qZuk76AwpD3dvTRPQbKC4kzx9cqQ8mHnu+9Q2su7N27SPChMaZ+8c799rrDjOfr3iEu8uuOUfE6b1aD4ud25BREurnfaa5pGQ+ey96eqE3yRiGhN9FLedwjIhHFSB0dvL85eqjDSJxNmXnlu/XEQSqquScugPakpfn1eMuXd5uZREu9NpyAZzdWXMMnmB24wE5B+JbF20sftwqufV3/pNpn3HRluOjbpBuH7lkO//ZjSj7orBJqDgJMB/TyBcutlHAjzxojiAWB7aLoyUVzpRvh9nIpTiXMx+w0v7n8ISGzd1FzvWXFz1Mihgz6yqZQ4fwTnSPbCMc1LcFrHRISzdFYGGivq/XOmVUETiGf1ONJ5GHoEP2QaG4n3LTh49QQcpfViXTw2euzQc9YHhSFj0hKUXCRrVmXNbTgU+QpopaNC/QSnjRxeWtS/k12BRDqWwckg6PU17B4EJvfMrFh8lqGn0sP5xU2YYq1u8CdBa3YMTiz5BtG+CXJTiKGJFg0nbK5HQvbGUQGtcerQDGxz+BnCFOWV1fcBX1A08VZ4/6yxMXaBEQpN4lzw0qJC0F0RrVr1LcD0h8OL+vend7qaZPKeCLcVsH0Qxt9BHIbg4PrVSKtLS4sG1CB+eSzaaQbOBj0PbuIQW7F57acT2nU6ooc3H8yaV3Md/F4Ft/lIpzJJDWcHiD/6vIAZkXLEpR90RvqKcla/cQyPdFqFeHXFibomcH1Ss/lcm955/cYrmQ/GMm0R5hyc8nI+8FmHvHw5sGxMy0Nc9MUJTSeh01icKW9reuhdr7zS4oEvyDiAhHm1W1ekwwRP/l+GuNvUdOTHxi/ScY2w2M9GlBQ+Q2az5y35CHpONgVbpvuUj0yUdtl0h25zkQfPRzp+hTL3T2ByEXDug2CrdlrbRo4vKdkapC/M70IaH4IO9UWkJyjzpnOLrRg+pPD6bPRqQWW6bF5NsS74Pwgz4LIDJ/FcRR12iiPybFY6tqY0ZpVIv8HIW1ssDhlFA2Yiv01G/AqQz8cElUOaqfCjxEP0EFV1KQSaj0DWESpICMagkC+RP+oJBlE1EcUVVKDKcLVpmmclqbnuQe68szFmfg92Dbpu2ByYOteooXmpLt7YGdXnu2iQbif1/eieyNy+/KiVYIGzKQ9mFrsXj8egojg2TwudS+6DwiC7xBVCncV1FLyVsVisNyrtr1DZ3If7X5mwjocyJ+fzNgPJbVCcZ86tOl3j/FpLsGvNuNUHlX59UjioazreDBlHxuJWz7jJLgYud1HD4di7H3ziFhSm2xt0jyCMrjFmFaISHg0cx+Fc3CHTn6k6Emn3a8u0xoj6eAEd8Y5p+1/gWL9ZqC5KnphT1QFyUAdpY4DZXK9MNBxt0WEZDLsbKX4IZ3usMd4HafkhGs5LyL2IdJyI/76NMatXnLF+eD67oOtRV3vzQdm8xadAr1stZk5ojIlewOhEHo2McYcZCuWNQUJ2qzdjXeNmHB0nPsKMFPQRDQ3HUlDCsi6kxtP7HoxxIm2RG9dYcasvDtF/kzP9cuaDMxoZzJ4k6PaCMQ+Q54qEV7eM1HTwh0P96xDPRZrGfkJiqAOIdOreGG+YG5h3SVfBfCntsumeLZ9n0hd+wVaQwIh0RSHPR0c2j0aM2ejVgso0Kp8bkb8Wbl5Tm4989QSw+BPJpvTIhY5NUurB78ZY3DwaPtGR16agAU2dBQkohyIaHgi//eIidqhpsXNQF1xcNqfKOTTe1kX9KQSagUBG+jZbjmDvorF4nH44QPmxLl1EezQ6381EcVUrdlwzcujAT21qLiE2oef/1zHDBv6H1df/OC4ar0PBPQQyDkex7BM1Qg+iguiHRuG0J+csPBhV/AuwO6vLEdGlGG2tw3mSg2V8MlAr2RRfCOcTyKTDm3tkDEMKTN73hJILB7ufiEK5YcSQ/pMpzpYQd0vxqIROQVOQEz2cN2654CzDwf0eOst2ZHHhHDx/gnNxTwiZm79Gj38EzuvsI6L6ncAFgzvW8aP39PmouDZFNW0MKnCiTjs0vrN2ukuW/Qi//ygtKXw7Ho8/TwaN8frbxlx49mdIqxchpweZoQLCmZ9iGqWtfcYnZ2WodH9EdnTJfKAz7RR6B5/iuFCI34MKEwtGtGIyk5clYtOA1wMR3fiDoenj0LBY3NI7lZaeu5meEafNwKjW+54NY6TtDZQu+HhRg7zVw4uzm98xIw1ZUlGvPKk/3b265UBNBzjNx5B/zps+fX579O4ugpjFumFYgeUjEWAapV2Out9EaYj0W4qG7F8glJ5ROmTgcuRRcGPqPXLRNxF8039O9GpBZTpu/hq6fF3QOf8OxPdYtNKdmiSznOnYGnfUX015EJMYf4ecPMto190lB8nuXw4ts+FNZOItBgu9oelsqIiLs5BXPnP7Vc8KgeYgYCC3Jaiw4AuUPtu9njG99yGmYp+Q5tRQ0fFK8rBwMsfICwQt6Gej1XVfxHuF9y22E9siEsacaBvWwLexqG1QhsPIv0BFq2H6qGs0vm0Xi7etiEf0ZbrQB6PCnsCiRilcHme7zu3PJGd6AwaeQWEEygmm5AqKc5JJ0A6TxIIGCZ1rFGF5Jenh6BUjvfkxJpahd48GJfNFlFhBYXp9IkgHeDyCRxmsENH2fXXGZ6FhfBDp8hx0ohkCfEPrH8c072OY9hwP+6OQ/s+NHj14o1emfA+FBQaf7stCXBPx4wgLLZ0TNkaBlkbT3j4X9FqHFvJxstKEPhttZKPbmc6Mm1C5jmSWdp3g5kLGdRyQ74hGhNygUgRd7z4YI0+5EoFCSktbd/D2c0bMPXk7kzyXbmnUdJSmbmq65e+8suiYPmeuN/LzR2O6AHnduiJj3g139Opt57091F3m32B9cQo1ps7TZrTsM16Jii3EjkHnbTjK7GTMXBBZ9Z8dRYnCzVOm8d3yJKbrs5BIK4Um7kdKUz1xkuPH/UB0bMmsgIpG6uq4aGgTs23jCDzkmCKFUBKdV588smNdfH3rjma/sM7PRHn4CTdYFaaR+4wces4Hjj/1oBBoBgIaUWHRdxD6JUc4Gb0TPRUaUBQC40o6mH1W5ZKfwkMXTCW+7vUIRpV/odr7XvncqgvLyxd04hFtJirKm23WB5LBRfe6TauqIK8znicz1l5n0fDzujCuZI0bb0UljKlX0d0utEnhaLV3oPR0gMxebnNv2BnD8DrO8p4pzhj3vYdS2wXfzzDiW3g4BoK/dMRx/hoqguNA//m2GeP/RhH/W0jwE3Wm74Abg6ZCHbd4cMftk3deW5srzvY0b3n1ofj2RBVad5Ozt9EhOQM4bdm6ZufVFjfXQEeaErMv02x8FNXj8UibK4DvVGne3Dsa69chF9+oFn8HU2XdUduOQLvzilcOUbTBrCOaW2vbmjqMFMVormn4tuu6OPsB3irXrayDbvz7kIsq3Llq0Qqf6lp01fQegLHj0+fBjbM7D2VKZx8xQUaObhRvSg+ipptaUVHgpqaTnhG+hZmYx/HZ42+It167edXc3cm7LaF7Rn25+Bwt6onTn150VFllzVBM855BcUjSq32NvBT5aNlLlyA/vYl80FPGz777lOm8gu8cCmwOQcfqtvq4WIS4H5/ipxkvrVmbq6Y980JXg2lXIl/v2rZhxScod7VoT3vOmPFsR3QmfcthQedWvwyF9FdrxfaKhngDfYIAW5mhpnybgb1ymopAWo/TbY1KzTM6SdjmSnGVoOYSN6CReZRH89ajgTgYpF2TSAq+WZSiBzmgdYcjdmHkdh8K6PX2/tYAujNHr4bNVIm/xTX9g14nnG5PGwbpGRhGUpgcZe8RJRcW56DhfxiN11Sdh1ZglNEGIzfUkYzVbW2ciD7y59FQ+MuQwb/AVOur23nt02tX1aMHLBZhpPZSyipOT9yCcHawSD4QruCM/BoVyFPQZcrIosJ5jbF4BfDdeVDX1jsNbjyNDst6vNt6JfbI8hpUfuuXv8MXpMtrSndgaOeBsOkdqeKD+Lb4TYjHch41VuiG9j7CWMobNt7pxZXyAcK6AWPGJ6FPLSrf3vh8MDE1XI6pfz6y6xF5+AatlWDM3fQtmosH8D34uq7doonG3/UehLFXBzRYmI1PpAvz4Ex6yDwUhHlGee6IuHTLRk0nvcXj5lQ0Lhjqi5njx4+3450p70pdpX95z1l34IB8aucF8ouwMcUsrEz6inprDvyso210mJq+lfIO+c2JXs2nTI8sGfAFGr35mNia18rQVkJ2K3ybj5FMuoLi6KZjS7ikjigviYYiq+FnCPQad8kll8Twvb4MQrpg5D8lKI/wxvqnkH9X52vtNkaM8Cro8/IOXlsl5aq7QmCfIpBtmbxUhkYB9ko7aeC6u6mipHHCvT/dmXTTnLtfGM3x73YbFGcaPaUthkh6JLuUhtMtMIfnoDDJKxaLfYpVluMTS/+fpdW87suXio2+15VXLvkP/P7F7Xh3n2lrg3d7Q4AsX32k2z3BaU/8yvDd90yYu93l+Lzb1HS7k3dbQPdAfYNkU/6jfAU80K6lX0FlmuoF/Hw/E6RLSTUpm1PTG3lY0LYbmgGArW/Y0ldQHiGMvTSI0o+6KwSag0DGDNgcQcrt/kEA2xk+xoj4zuFDBuQ0dTvj6YVHY+vMh6h7VrKGXaeqk2T2T7qpUPccAWpQsa3sXYxAW48de87OPZeoJCgE9gwBWiOhrv9mBBoaTxNse12uUfj0g/DnRx7f2BX7dDfAD2aq1aUQ+O9E4OP3l/67V6+Tuo4de4FqTP87k1BprRBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKAYWAQkAhoBBQCCgEFAIKgZZF4P8B9sQxeKd5npYAAAAASUVORK5CYII="},"23d09b78-4bbb-426b-9a6d-7f66fa52676e.png":{"image/png":"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"},"acd622c8-846d-4ed2-8a3c-56c55978b6ee.png":{"image/png":"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"},"2c2b1375-f63f-4440-9963-d9c6b6578cae.png":{"image/png":"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"}}},{"cell_type":"markdown","source":"### Data Pipeline","metadata":{}},{"cell_type":"code","source":"# Chapter 13\n#IMG_SIZE = 64\nIMG_SIZE = 224\nBATCH_SIZE = 32\n\ndef load_image(path, label):\n    img = tf.io.read_file(path)\n    img = tf.image.decode_jpeg(img, channels=3)\n    img = tf.image.resize(img, [IMG_SIZE, IMG_SIZE])\n    return img / 255.0, label","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"from tensorflow.data import Dataset\n\n# paths and labels\ntrain_paths = train_df['image_path'].values\ntrain_labels = train_df['is_venomous'].values\nval_paths = val_df['image_path'].values\nval_labels = val_df['is_venomous'].values\n\n# Create datasets\n# Training set\ntrain_ds = Dataset.from_tensor_slices((train_paths, train_labels))\ntrain_ds = train_ds.shuffle(1000)\ntrain_ds = train_ds.map(load_image)\ntrain_ds = train_ds.batch(BATCH_SIZE)\ntrain_ds = train_ds.prefetch(1)\n\n# Validation set\nval_ds = Dataset.from_tensor_slices((val_paths, val_labels))\nval_ds = val_ds.map(load_image)\nval_ds = val_ds.batch(BATCH_SIZE)\nval_ds = val_ds.prefetch(1)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Baseline CNN","metadata":{}},{"cell_type":"markdown","source":"Simple Architecture: 3 Conv blocks -> Flatten -> Dense -> Output\n\nSigmoid for binary classification tasks","metadata":{}},{"cell_type":"code","source":"# Baseline CNN (from scratch)\nfrom tensorflow import keras\n\nbaseline_model = keras.Sequential([\n    # First conv block\n    keras.layers.Conv2D(filters=32, kernel_size=3, activation='relu', \n                       input_shape=(IMG_SIZE, IMG_SIZE, 3)),\n    keras.layers.MaxPooling2D(pool_size=2),\n    \n    # Second conv block\n    keras.layers.Conv2D(filters=64, kernel_size=3, activation='relu'),\n    keras.layers.MaxPooling2D(pool_size=2),\n    \n    # Third conv block\n    keras.layers.Conv2D(filters=128, kernel_size=3, activation='relu'),\n    keras.layers.MaxPooling2D(pool_size=2),\n    \n    # Classifier head\n    keras.layers.Flatten(),\n    keras.layers.Dropout(0.5),\n    keras.layers.Dense(units=128, activation='relu'),\n    keras.layers.Dense(units=1, activation='sigmoid')  # Binary output\n])\n\nbaseline_model.compile(\n    optimizer='adam',\n    loss='binary_crossentropy',\n    metrics=['accuracy']\n)\n\nbaseline_model.summary()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Train baseline CNN\nhistory = baseline_model.fit(\n    train_ds,\n    validation_data=val_ds,\n    epochs=5,\n    verbose=1\n)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"~81% accuracy seemed suspicious. Almost the same ratio with the non-venomus class.","metadata":{}},{"cell_type":"markdown","source":"### Baseline CNN Confusion Matrix","metadata":{}},{"cell_type":"code","source":"# Get predictions\ny_pred_probs = baseline_model.predict(val_ds)\ny_pred = (y_pred_probs > 0.5).astype(int).flatten()\ny_true = val_df['is_venomous'].values\n\n# Confusion matrix\nfrom sklearn.metrics import confusion_matrix, classification_report\n\ncm = confusion_matrix(y_true, y_pred)\nplt.figure(figsize=(6, 5))\nsns.heatmap(cm, annot=True, fmt='d', cmap='Blues',\n            xticklabels=['Non-venomous', 'Venomous'],\n            yticklabels=['Non-venomous', 'Venomous'])\nplt.xlabel('Predicted')\nplt.ylabel('Actual')\nplt.title('Baseline CNN - Confusion Matrix')\nplt.tight_layout()\nplt.show()\n\nprint(classification_report(y_true, y_pred, \n                           target_names=['Non-venomous', 'Venomous']))","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"- this means almost all images are classified as non-venomous\n- venomous recall: 0.02 only catches 2% of venomous snakes\n\n**This is bad, 98% of venomous snakes are missed**","metadata":{}},{"cell_type":"markdown","source":"## Transfer Learning w/ MobileNetV2","metadata":{}},{"cell_type":"markdown","source":"### Load Pretrained Model","metadata":{}},{"cell_type":"code","source":"# Transfer Learning with MobileNetV2\nfrom tensorflow.keras.applications import MobileNetV2\n\n# Load pretrained model (without top classification layer)\nbase_model = MobileNetV2(\n    weights='imagenet',\n    include_top=False,\n    input_shape=(IMG_SIZE, IMG_SIZE, 3)\n)\n\n# Freeze the base model\nbase_model.trainable = False","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"### Build MobileNetV2 Model","metadata":{}},{"cell_type":"code","source":"mobilenet_model = keras.Sequential([\n    base_model,\n    keras.layers.GlobalAveragePooling2D(),\n    keras.layers.Dropout(0.5),\n    keras.layers.Dense(units=1, activation='sigmoid')\n])\n\nmobilenet_model.compile(\n    optimizer='adam',\n    loss='binary_crossentropy',\n    metrics=['accuracy']\n)\n\nmobilenet_model.summary()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Train MobileNetV2\nhistory_mobilenet = mobilenet_model.fit(\n    train_ds,\n    validation_data=val_ds,\n    epochs=5,\n    class_weight=class_weight_dict,\n    verbose=1\n)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"- validation accuracy fluctuates, same with loss","metadata":{}},{"cell_type":"markdown","source":"### MobileNetV2 Confusion Matrix","metadata":{}},{"cell_type":"code","source":"y_pred_probs = mobilenet_model.predict(val_ds)\ny_pred = (y_pred_probs > 0.5).astype(int).flatten()\ny_true = val_df['is_venomous'].values\n\ncm = confusion_matrix(y_true, y_pred)\nplt.figure(figsize=(6, 5))\nsns.heatmap(cm, annot=True, fmt='d', cmap='Blues',\n            xticklabels=['Non-venomous', 'Venomous'],\n            yticklabels=['Non-venomous', 'Venomous'])\nplt.xlabel('Predicted')\nplt.ylabel('Actual')\nplt.title('MobileNetV2 - Confusion Matrix')\nplt.tight_layout()\nplt.show()\n\nprint(classification_report(y_true, y_pred, \n                           target_names=['Non-venomous', 'Venomous']))","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"- without handling imbalance (class weights), venomous recall was 54%\n- Venomous recall showed significant improvement\n- Shows more false positives where non venomous were classified as venomous\n\n**For real-world application, its better to flag a harmless snake as dangerous than miss a venomous one.**","metadata":{}},{"cell_type":"markdown","source":"## Transfer Learning w/ EfficientNetB0","metadata":{}},{"cell_type":"markdown","source":"### EfficientNet Data Preprocessing\n\n* EfficientNet needs a separate data preprocessing because it expects pixel values in 0-255 range. MobilNet divides by 255 which scales the pixels 0-1 range.\n* Model has its own internal rescaling layer","metadata":{}},{"cell_type":"code","source":"# EfficientNet expects pixels in [0-255] range, not [0-1]\ndef load_image_efficientnet(path, label):\n    img = tf.io.read_file(path)\n    img = tf.image.decode_jpeg(img, channels=3)\n    img = tf.image.resize(img, [IMG_SIZE, IMG_SIZE])\n    return img, label\n\n# Create datasets for EfficientNet\n# Training set\ntrain_ds_eff = Dataset.from_tensor_slices((train_paths, train_labels))\ntrain_ds_eff = train_ds_eff.shuffle(1000)\ntrain_ds_eff = train_ds_eff.map(load_image_efficientnet)\ntrain_ds_eff = train_ds_eff.batch(BATCH_SIZE)\ntrain_ds_eff = train_ds_eff.prefetch(1)\n\n# Validation set\nval_ds_eff = Dataset.from_tensor_slices((val_paths, val_labels))\nval_ds_eff = val_ds_eff.map(load_image_efficientnet)\nval_ds_eff = val_ds_eff.batch(BATCH_SIZE)\nval_ds_eff = val_ds_eff.prefetch(1)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"### Load pretrained model","metadata":{}},{"cell_type":"code","source":"from tensorflow.keras.applications import EfficientNetB0\n\n# Load pretrained model (without top classification layer)\nefficient_base = EfficientNetB0(\n    weights='imagenet',\n    include_top=False,\n    input_shape=(IMG_SIZE, IMG_SIZE, 3)\n)\n\n# Freeze the base model\nefficient_base.trainable = False","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"### Build EfficientNet Model","metadata":{}},{"cell_type":"code","source":"efficientnet_model = keras.Sequential([\n    efficient_base,\n    keras.layers.GlobalAveragePooling2D(),\n    keras.layers.Dropout(0.5),\n    keras.layers.Dense(units=1, activation='sigmoid')\n])\n\nefficientnet_model.compile(\n    optimizer='adam',\n    loss='binary_crossentropy',\n    metrics=['accuracy']\n)\n\nefficientnet_model.summary()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Train EfficientNet\nhistory_efficient = efficientnet_model.fit(\n    train_ds_eff,\n    validation_data=val_ds_eff,\n    epochs=5,\n    class_weight=class_weight_dict,\n    verbose=1\n)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"### EfficientNetB0 Confusion Matrix","metadata":{}},{"cell_type":"code","source":"y_pred_probs = efficientnet_model.predict(val_ds_eff)\ny_pred = (y_pred_probs > 0.5).astype(int).flatten()\ny_true = val_df['is_venomous'].values\n\ncm = confusion_matrix(y_true, y_pred)\nplt.figure(figsize=(6, 5))\nsns.heatmap(cm, annot=True, fmt='d', cmap='Blues',\n            xticklabels=['Non-venomous', 'Venomous'],\n            yticklabels=['Non-venomous', 'Venomous'])\nplt.xlabel('Predicted')\nplt.ylabel('Actual')\nplt.title('EfficientNetB0 - Confusion Matrix')\nplt.tight_layout()\nplt.show()\n\nprint(classification_report(y_true, y_pred, target_names=['Non-venomous', 'Venomous']))","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Model Comparison","metadata":{}},{"cell_type":"code","source":"from sklearn.metrics import roc_curve, auc\n\n# Get predictions from each model\n# Note: baseline and mobilenet use val_ds, efficientnet uses val_ds_eff\ny_true = val_df['is_venomous'].values\n\npred_baseline = baseline_model.predict(val_ds).flatten()\npred_mobilenet = mobilenet_model.predict(val_ds).flatten()\npred_efficient = efficientnet_model.predict(val_ds_eff).flatten()\n\n# Calculate ROC curves\nfpr_base, tpr_base, _ = roc_curve(y_true, pred_baseline)\nfpr_mobile, tpr_mobile, _ = roc_curve(y_true, pred_mobilenet)\nfpr_efficient, tpr_efficient, _ = roc_curve(y_true, pred_efficient)\n\n# Calculate AUC scores\nauc_base = auc(fpr_base, tpr_base)\nauc_mobile = auc(fpr_mobile, tpr_mobile)\nauc_efficient = auc(fpr_efficient, tpr_efficient)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Plot\nplt.figure(figsize=(8, 6))\nplt.plot(fpr_base, tpr_base, label=f'Baseline CNN (AUC = {auc_base:.3f})', linewidth=2)\nplt.plot(fpr_mobile, tpr_mobile, label=f'MobileNetV2 (AUC = {auc_mobile:.3f})', linewidth=2)\nplt.plot(fpr_efficient, tpr_efficient, label=f'EfficientNetB0 (AUC = {auc_efficient:.3f})', linewidth=2)\nplt.plot([0, 1], [0, 1], 'k--', label='Random Classifier')\n\nplt.xlabel('False Positive Rate')\nplt.ylabel('True Positive Rate')\nplt.title('ROC Curve Comparison')\nplt.legend(loc='lower right')\nplt.grid(True, alpha=0.3)\nplt.tight_layout()\nplt.show()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Sample Predictions on Actual Images","metadata":{}},{"cell_type":"code","source":"# Sample predictions\nimport random\n\nfig, axes = plt.subplots(2, 3, figsize=(12, 8))\n\nfor i, idx in enumerate(random.sample(range(len(val_df)), 6)):\n    img_path = val_df.iloc[idx]['image_path']\n    true_label = val_df.iloc[idx]['is_venomous']\n    \n    # Load and predict\n    img = tf.io.read_file(img_path)\n    img = tf.image.decode_jpeg(img, channels=3)\n    img = tf.image.resize(img, [IMG_SIZE, IMG_SIZE])\n    pred_prob = efficientnet_model.predict(tf.expand_dims(img, 0), verbose=0)[0][0]\n    \n    # Display\n    true_text = 'Venomous' if true_label == 1 else 'Non-venomous'\n    pred_text = 'Venomous' if pred_prob > 0.5 else 'Non-venomous'\n    color = 'green' if (pred_prob > 0.5) == true_label else 'red'\n    \n    axes.flat[i].imshow(img.numpy().astype('uint8'))\n    axes.flat[i].axis('off')\n    axes.flat[i].set_title(f'Actual: {true_text}\\nPredicted: {pred_text} ({pred_prob:.1%})', \n                           color=color, fontsize=10)\n\nplt.suptitle('EfficientNetB0 Sample Predictions', fontweight='bold')\nplt.tight_layout()\nplt.show()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Stage 2: Venomous Species Classification","metadata":{}},{"cell_type":"markdown","source":"## Prepare the data","metadata":{}},{"cell_type":"code","source":"# Filter to venomous snakes only\nvenomous_df = df[df['is_venomous'] == 1].copy()\n\nprint(f\"Total images: {len(venomous_df):,}\")\nprint(f\"Total species: {venomous_df['binomial_name'].nunique()}\")\nprint(f\"Groups: {venomous_df['genus'].nunique()}\")","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Quick stats\nvenomous_df['binomial_name'].value_counts().describe()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Species distribution (top 20 most common)\nvenomous_df['binomial_name'].value_counts().head(20).plot(kind='barh', figsize=(10, 8))\nplt.xlabel('Number of Images')\nplt.ylabel('Species')\nplt.title('Top 20 Venomous Species by Image Count')\nplt.gca().invert_yaxis()\nplt.tight_layout()\nplt.show()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Data Preprocessing","metadata":{}},{"cell_type":"markdown","source":"### Encode Labels\n\n* Encode class_ids to sequential labels (0-82)\n* Why is it needed? class_ids range from 21 to 1345, but only concerned with 83 species.\n* If we used the original class_ids, the model would create 1346 output neurons\n* (one for each possible value 0-1345) when we only need 83, this is a hig overhead \n* So we map them: 21->0, 22->1, 23->2, ... 1345->82","metadata":{}},{"cell_type":"code","source":"# Step 1: Get all unique class_ids and sort them\nunique_classes = sorted(venomous_df['class_id'].unique())\n\n# Step 2: Create a mapping dictionary {original_id: new_index}\n# Example: {21: 0, 22: 1, 23: 2, ...}\nclass_to_idx = {}\nfor i, class_id in enumerate(unique_classes):\n    class_to_idx[class_id] = i\n\n# Step 3: Create reverse mapping for decoding predictions later\n# Example: {0: 21, 1: 22, 2: 23, ...}\nidx_to_class = {}\nfor class_id, idx in class_to_idx.items():\n    idx_to_class[idx] = class_id\n\n# Step 4: Apply the mapping to create encoded labels\nvenomous_df['encoded_label'] = venomous_df['class_id'].map(class_to_idx)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Compute class weights to handle imbalanced species\nfrom sklearn.utils.class_weight import compute_class_weight\n\ny_all = venomous_df['encoded_label'].values\nclass_weights = compute_class_weight('balanced', classes=np.unique(y_all), y=y_all)\nclass_weight_dict = dict(enumerate(class_weights))\n\nprint(f\"Weight range: {min(class_weights):.2f} - {max(class_weights):.2f}\")","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Train/val split\n# 80/20\ntrain_df, val_df = train_test_split(\n    venomous_df, \n    test_size=0.2,\n    random_state=42\n)\n\nprint(f\"Training samples: {len(train_df):,}\")\nprint(f\"Validation samples: {len(val_df):,}\")","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Visualize train/val split\nfig, ax = plt.subplots(figsize=(6, 4))\n\ncounts = [len(train_df), len(val_df)]\nlabels = ['Train', 'Validation']\n\nbars = ax.bar(labels, counts, color=['steelblue', 'darkorange'])\n\nax.set_ylabel('Number of Images')\nax.set_title('Stage 2: Train/Validation Split (80/20)')\n\n# Add count labels on bars\nfor bar in bars:\n    ax.text(bar.get_x() + bar.get_width()/2, bar.get_height() + 100, \n            f'{int(bar.get_height()):,}', ha='center', va='bottom', fontsize=10)\n\nplt.tight_layout()\nplt.show()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"### Data Pipeline","metadata":{}},{"cell_type":"code","source":"# Augmentation function\ndef augment_image(image, label):\n    image = tf.image.random_flip_left_right(image)\n    image = tf.image.random_brightness(image, max_delta=0.1)\n    return image, label","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# EfficientNet preprocessing\n# paths and labels\ntrain_paths_s2 = train_df['image_path'].values\ntrain_labels_s2 = train_df['encoded_label'].values\n\nval_paths_s2 = val_df['image_path'].values\nval_labels_s2 = val_df['encoded_label'].values\n\n# Create Datasets\n# Training set\ntrain_ds = Dataset.from_tensor_slices((train_paths_s2, train_labels_s2))\ntrain_ds = train_ds.shuffle(1000)\ntrain_ds = train_ds.map(load_image_efficientnet)\ntrain_ds = train_ds.map(augment_image)\ntrain_ds = train_ds.batch(BATCH_SIZE)\ntrain_ds = train_ds.prefetch(1)\n\n# Validation set\nval_ds = Dataset.from_tensor_slices((val_paths_s2, val_labels_s2))\nval_ds = val_ds.map(load_image_efficientnet)\nval_ds = val_ds.batch(BATCH_SIZE)\nval_ds = val_ds.prefetch(1)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Build EfficientNet Model for Species Classification","metadata":{}},{"cell_type":"code","source":"from tensorflow.keras.applications import EfficientNetB0\nfrom tensorflow.keras import layers, models\n\nNUM_CLASSES = venomous_df['encoded_label'].nunique()\n\neffnet_base = EfficientNetB0(weights='imagenet', \n                            include_top=False, \n                            input_shape=(IMG_SIZE, IMG_SIZE, 3))\n\n# Freeze the base model -all layers\neffnet_base.trainable = False","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"species_model = keras.Sequential([\n    effnet_base,\n    keras.layers.GlobalAveragePooling2D(),\n    keras.layers.Dense(units=256, activation='relu', kernel_initializer='he_normal'),\n    keras.layers.Dropout(0.5),\n    keras.layers.Dense(units=NUM_CLASSES, activation='softmax')\n])\n\nspecies_model.compile(\n    optimizer='adam',\n    loss='sparse_categorical_crossentropy',\n    metrics=['accuracy']\n)\n\nspecies_model.summary()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Train EfficientNet for Species Classification\nspecies_history = species_model.fit(\n    train_ds,\n    validation_data=val_ds,\n    epochs=20,\n    class_weight=class_weight_dict\n)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Plot training history\nfig, axes = plt.subplots(1, 2, figsize=(12, 4))\n\n# Accuracy\naxes[0].plot(species_history.history['accuracy'], label='Train')\naxes[0].plot(species_history.history['val_accuracy'], label='Validation')\naxes[0].set_title('Accuracy')\naxes[0].set_xlabel('Epoch')\naxes[0].set_ylabel('Accuracy')\naxes[0].legend()\n\n# Loss\naxes[1].plot(species_history.history['loss'], label='Train')\naxes[1].plot(species_history.history['val_loss'], label='Validation')\naxes[1].set_title('Loss')\naxes[1].set_xlabel('Epoch')\naxes[1].set_ylabel('Loss')\naxes[1].legend()\n\nplt.tight_layout()\nplt.show()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Fine Tuning","metadata":{}},{"cell_type":"code","source":"# Unfreeze the base model, just the last 10 layers\neffnet_base.trainable = True\nfor layer in effnet_base.layers[:-10]:\n    layer.trainable = False\n\n# Recompile with smaller learning rate\nspecies_model.compile(\n    optimizer=tf.keras.optimizers.Adam(learning_rate=0.00001),\n    loss='sparse_categorical_crossentropy',\n    metrics=['accuracy']\n)\n\nspecies_model.summary()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Train with fine-tuning\nfine_tune_history = species_model.fit(\n    train_ds,\n    validation_data=val_ds,\n    epochs=15,\n    class_weight=class_weight_dict\n)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# Combine histories\nfull_acc = species_history.history['val_accuracy'] + fine_tune_history.history['val_accuracy']\nfull_loss = species_history.history['val_loss'] + fine_tune_history.history['val_loss']\n\nplt.figure(figsize=(10, 4))\nplt.subplot(1, 2, 1)\nplt.plot(full_acc)\nplt.axvline(x=19, color='r', linestyle='--', label='Fine-tuning start')\nplt.title('Validation Accuracy')\nplt.xlabel('Epoch')\nplt.legend()\n\nplt.subplot(1, 2, 2)\nplt.plot(full_loss)\nplt.axvline(x=19, color='r', linestyle='--', label='Fine-tuning start')\nplt.title('Validation Loss')\nplt.xlabel('Epoch')\nplt.legend()\n\nplt.tight_layout()\nplt.show()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"- Frozen (20 epochs): val_accuracy went from 19% to 38.3%\n- Fine-tuned (15 epochs): val_accuracy improved to 40%\n- Fine-tuning gave about 2% boost over frozen\n- unfroze last 10 layers with lower learning rate\n- Spike in val_loss when fine-tuning starts, then it continues decreasing (still learning)","metadata":{}},{"cell_type":"markdown","source":"## Evaluation","metadata":{}},{"cell_type":"code","source":"# Collect predictions\nval_predictions = []\nval_labels = []\n\nfor images, labels in val_ds:\n    preds = species_model.predict(images, verbose=0)\n    val_predictions.extend(np.argmax(preds, axis=1))\n    val_labels.extend(labels.numpy())\n\nval_predictions = np.array(val_predictions)\nval_labels = np.array(val_labels)","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"### Confusion Matrix","metadata":{}},{"cell_type":"code","source":"# Top 10 Species Confusion Matrix\nfrom sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay\n\nunique, counts = np.unique(val_labels, return_counts=True)\ntop10_indices = unique[np.argsort(counts)[-10:]]\n\nlabel_to_species = dict(zip(venomous_df['encoded_label'], venomous_df['binomial_name']))\ntop10_names = [label_to_species[l].split()[-1][:12] for l in top10_indices]\n\nmask = np.isin(val_labels, top10_indices)\ncm = confusion_matrix(val_labels[mask], val_predictions[mask], labels=top10_indices)\n\nplt.figure(figsize=(10, 8))\ndisp = ConfusionMatrixDisplay(cm, display_labels=top10_names)\ndisp.plot(cmap='Blues', xticks_rotation=45, values_format='d')\nplt.title('Top 10 Species')\nplt.tight_layout()\nplt.show()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"**Species Legend:**\n- viridis = Prairie Rattlesnake\n- pyrrhus = Southwestern Speckled Rattlesnake  \n- lepidus = Rock Rattlesnake\n- scutulatus = Mojave Rattlesnake\n- cerastes = Sidewinder\n- atrox = Western Diamondback\n- oreganus = Western Rattlesnake\n- molossus = Black-tailed Rattlesnake\n- ruber = Red Diamond Rattlesnake\n- asper = Fer-de-lance (Central American pit viper)","metadata":{}},{"cell_type":"markdown","source":"- Best performers: asper (115), ruber (97), cerastes (105)\n- Common confusions are between rattlesnake species (all Crotalus genus):\n  - atrox confused with scutulatus (15) and ruber (12)\n  - viridis and scutulatus confused with each other (15 each way)\n  - oreganus confused with molossus (13) and viridis","metadata":{}},{"cell_type":"markdown","source":"## Sample Predictions on Actual Images","metadata":{}},{"cell_type":"code","source":"plt.figure(figsize=(12, 8))\n\nfor images, labels in val_ds.shuffle(1000).take(1):\n    preds = species_model.predict(images, verbose=0)\n    pred_labels = np.argmax(preds, axis=1)\n    confidences = np.max(preds, axis=1)\n    \n    for i in range(6):\n        plt.subplot(2, 3, i+1)\n        img = images[i].numpy()\n        img = (img - img.min()) / (img.max() - img.min())\n        plt.imshow(img)\n        \n        actual = label_to_species.get(labels[i].numpy(), \"Unknown\").split()[-1]\n        predicted = label_to_species.get(pred_labels[i], \"Unknown\").split()[-1]\n        correct = pred_labels[i] == labels[i].numpy()\n        color = 'green' if correct else 'red'\n        \n        plt.title(f\"Actual: {actual}\\nPredicted: {predicted} ({confidences[i]:.1%})\", \n                  fontsize=10, color=color)\n        plt.axis('off')\n\nplt.suptitle('Sample Predictions', fontsize=14)\nplt.tight_layout()\nplt.show()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Summary","metadata":{}},{"cell_type":"markdown","source":"### Stage 1: Binary Classification (Venomous vs Non-Venomous)\n\n| Model | Accuracy | Venomous Recall |\n|-------|----------|-----------------|\n| Baseline CNN | 81% | 2% |\n| MobileNetV2 | 77% | 79% |\n| EfficientNetB0 | 81% | 83% |\n\nEfficientNetB0 performed best overall. The high venomous recall (83%) is important for this use case since in this case, it's better to have a false alarm than to miss a venomous snake.\n\n### Stage 2: Species Classification\n\n- 83 venomous species from North/Central America\n- EfficientNetB0 with transfer learning + fine-tuning\n- Final validation accuracy: ~40%\n- Most confusion happens between similar rattlesnake species (Crotalus genus)\n\n40% accuracy seems low, but the loss and accuracy curves show the model was still learning steadily through all 20 epochs. It was able to pick up meaningful patterns, esp for visually distinct species.","metadata":{}},{"cell_type":"markdown","source":"### Lessons Learned\n\n1. **Class imbalance** - Both stages had imbalanced data. Class weighting helped the model pay more attention to underrepresented classes.\n\n2. **Label encoding matters** - For multi-class classification, labels need to start at 0 and be continuous (0 to 82 for 83 species). Getting this right fixed training errors.\n\n3. **Transfer learning works** - EfficientNetB0 beat the baseline CNN significantly with minimal extra code.\n\n4. **Similar species are hard** - Rattlesnakes within the same genus look very similar.","metadata":{}}]}