1103 lines
677 KiB
Text
1103 lines
677 KiB
Text
|
{
|
||
|
|
"cells": [
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 1,
|
||
|
|
"metadata": {
|
||
|
|
"ExecuteTime": {
|
||
|
|
"end_time": "2022-07-05T17:07:51.927811Z",
|
||
|
|
"start_time": "2022-07-05T17:07:48.781405Z"
|
||
|
|
},
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:28:44.434510Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:28:44.433750Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:28:47.737019Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:28:47.736442Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:28:44.434436Z"
|
||
|
|
},
|
||
|
|
"id": "8hbqOqrTKhVE",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [],
|
||
|
|
"source": [
|
||
|
|
"import os\n",
|
||
|
|
"import random\n",
|
||
|
|
"import numpy as np\n",
|
||
|
|
"import matplotlib.pyplot as plt\n",
|
||
|
|
"import seaborn as sns\n",
|
||
|
|
"from keras.models import Sequential\n",
|
||
|
|
"from keras.preprocessing.image import ImageDataGenerator\n",
|
||
|
|
"from keras.applications import EfficientNetV2L\n",
|
||
|
|
"from keras.optimizers import Adam\n",
|
||
|
|
"from sklearn.metrics import confusion_matrix, classification_report\n",
|
||
|
|
"from keras.utils import image_dataset_from_directory\n",
|
||
|
|
"from keras.layers import (Dense, Conv2D, MaxPool2D, Dropout, \n",
|
||
|
|
" Flatten, GlobalAveragePooling2D, Rescaling)"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 2,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:28:47.738287Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:28:47.737964Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:28:47.741595Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:28:47.741046Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:28:47.738271Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [],
|
||
|
|
"source": [
|
||
|
|
"sns.set_theme(style='darkgrid')"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 3,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:28:47.742341Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:28:47.742199Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:28:47.746494Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:28:47.745989Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:28:47.742327Z"
|
||
|
|
},
|
||
|
|
"id": "8xRpd7aTcGAK",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [],
|
||
|
|
"source": [
|
||
|
|
"np.random.seed(42)"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 4,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:28:47.747595Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:28:47.747405Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:28:47.751243Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:28:47.750746Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:28:47.747581Z"
|
||
|
|
},
|
||
|
|
"id": "Ik_Ed16RgzH1",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [],
|
||
|
|
"source": [
|
||
|
|
"# path to files on Adam's local machine\n",
|
||
|
|
"train_path = 'data/seg_train/'\n",
|
||
|
|
"test_path = 'data/seg_test/'"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 5,
|
||
|
|
"metadata": {
|
||
|
|
"colab": {
|
||
|
|
"base_uri": "https://localhost:8080/",
|
||
|
|
"height": 286
|
||
|
|
},
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:28:48.974528Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:28:48.974039Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:28:49.133624Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:28:49.132844Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:28:48.974512Z"
|
||
|
|
},
|
||
|
|
"id": "Cm4rla2kKhVG",
|
||
|
|
"outputId": "c59ecc9c-18c5-44b7-f139-c89a869c8640",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"image/png": "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
|
||
|
|
"text/plain": [
|
||
|
|
"<Figure size 432x288 with 1 Axes>"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"metadata": {
|
||
|
|
"needs_background": "light"
|
||
|
|
},
|
||
|
|
"output_type": "display_data"
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"plt.imshow(plt.imread(train_path + 'buildings/0.jpg'));"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"Here is a building from the buildings Collection"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {
|
||
|
|
"id": "6WqU_6rBKhVG"
|
||
|
|
},
|
||
|
|
"source": [
|
||
|
|
"## Modeling\n",
|
||
|
|
"\n",
|
||
|
|
"### ImageDataGenerator\n",
|
||
|
|
"\n",
|
||
|
|
"We'll use the `ImageDataGenerator` class to handle rescaling our data to be between 0 and 1 rather than between 0 and 255. You can read more about the `ImageDataGenerator` class in the documentation [here](https://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/image/ImageDataGenerator)."
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 6,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:28:56.963717Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:28:56.963065Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:28:56.967228Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:28:56.966537Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:28:56.963687Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [],
|
||
|
|
"source": [
|
||
|
|
"class_names = ['buildings', 'forest', 'glacier', 'mountain', 'sea', 'street']"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {
|
||
|
|
"id": "ip-0O2xwKhVH"
|
||
|
|
},
|
||
|
|
"source": [
|
||
|
|
"### Train and Test Data\n",
|
||
|
|
"\n",
|
||
|
|
"80% train/test split with shuffle"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 8,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:29:11.672947Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:29:11.672683Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:29:12.942926Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:29:12.942365Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:29:11.672931Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"name": "stdout",
|
||
|
|
"output_type": "stream",
|
||
|
|
"text": [
|
||
|
|
"Found 14034 files belonging to 6 classes.\n",
|
||
|
|
"Using 11228 files for training.\n",
|
||
|
|
"Found 14034 files belonging to 6 classes.\n",
|
||
|
|
"Using 2806 files for validation.\n"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"#Inferring labels goes in alphanumeric order, so will match our class names\n",
|
||
|
|
"train_data = image_dataset_from_directory(directory=train_path,\n",
|
||
|
|
" labels='inferred',\n",
|
||
|
|
" label_mode='categorical',\n",
|
||
|
|
" class_names=class_names,\n",
|
||
|
|
" color_mode='rgb',\n",
|
||
|
|
" batch_size=32,\n",
|
||
|
|
" image_size=(150, 150),\n",
|
||
|
|
" validation_split=.2,\n",
|
||
|
|
" subset='training',\n",
|
||
|
|
" shuffle=True,\n",
|
||
|
|
" seed=42)\n",
|
||
|
|
"\n",
|
||
|
|
"test_data = image_dataset_from_directory(directory=train_path,\n",
|
||
|
|
" labels='inferred',\n",
|
||
|
|
" label_mode='categorical',\n",
|
||
|
|
" class_names=class_names,\n",
|
||
|
|
" color_mode='rgb',\n",
|
||
|
|
" batch_size=32,\n",
|
||
|
|
" image_size=(150, 150),\n",
|
||
|
|
" validation_split=.2,\n",
|
||
|
|
" subset='validation',\n",
|
||
|
|
" shuffle=True,\n",
|
||
|
|
" seed=42)"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {
|
||
|
|
"id": "ezaEVseFKhVH"
|
||
|
|
},
|
||
|
|
"source": [
|
||
|
|
"### Basic Network\n",
|
||
|
|
"\n",
|
||
|
|
"We'll fit a basic convolutional neural network to start with."
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 9,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:29:19.918941Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:29:19.918311Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:29:19.949324Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:29:19.948563Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:29:19.918911Z"
|
||
|
|
},
|
||
|
|
"id": "mAlAbM3OcGAF",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [],
|
||
|
|
"source": [
|
||
|
|
"# instantiate the model\n",
|
||
|
|
"basic_model = Sequential()\n",
|
||
|
|
"\n",
|
||
|
|
"#Rescale our data so color values are between 0 and 1\n",
|
||
|
|
"basic_model.add(Rescaling(1./255))\n",
|
||
|
|
"\n",
|
||
|
|
"# this is the fancy stuff!\n",
|
||
|
|
"# 16 different 3-by-3 filters that pick up on patterns in the images\n",
|
||
|
|
"basic_model.add(Conv2D(filters=16, # number of filters\n",
|
||
|
|
" kernel_size=(3, 3), # height/width of filter\n",
|
||
|
|
" activation='relu', # activation function\n",
|
||
|
|
" ))\n",
|
||
|
|
"# reduce the dimensionality inside the CNN\n",
|
||
|
|
"basic_model.add(MaxPool2D(pool_size=(2, 2))) # dimensions of region of pooling\n",
|
||
|
|
"\n",
|
||
|
|
"# before this information inside the NN is 3-d -- to connec to a Dense\n",
|
||
|
|
"# layer, we need to squash it\n",
|
||
|
|
"basic_model.add(Flatten())\n",
|
||
|
|
"\n",
|
||
|
|
"# let it learn a little more...\n",
|
||
|
|
"basic_model.add(Dense(100, activation='relu'))\n",
|
||
|
|
"\n",
|
||
|
|
"# output layer\n",
|
||
|
|
"basic_model.add(Dense(6, activation='softmax'))"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 10,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:29:23.093969Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:29:23.093284Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:29:23.106381Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:29:23.105554Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:29:23.093940Z"
|
||
|
|
},
|
||
|
|
"id": "KD4bRtFAcGAF",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [],
|
||
|
|
"source": [
|
||
|
|
"# compile\n",
|
||
|
|
"basic_model.compile(loss='categorical_crossentropy',\n",
|
||
|
|
" optimizer='adam',\n",
|
||
|
|
" metrics=['accuracy'])"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 12,
|
||
|
|
"metadata": {
|
||
|
|
"colab": {
|
||
|
|
"base_uri": "https://localhost:8080/"
|
||
|
|
},
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:30:15.072057Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:30:15.071462Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:30:35.983523Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:30:35.982892Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:30:15.072028Z"
|
||
|
|
},
|
||
|
|
"id": "cc7o5JcQmGrK",
|
||
|
|
"outputId": "b6122964-6ea1-4783-b9d2-fe6bf56c111b",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"name": "stdout",
|
||
|
|
"output_type": "stream",
|
||
|
|
"text": [
|
||
|
|
"Epoch 1/9\n",
|
||
|
|
"351/351 [==============================] - 2s 6ms/step - loss: 0.0384 - accuracy: 0.9935 - val_loss: 1.3848 - val_accuracy: 0.7153\n",
|
||
|
|
"Epoch 2/9\n",
|
||
|
|
"351/351 [==============================] - 2s 7ms/step - loss: 0.0286 - accuracy: 0.9958 - val_loss: 1.5794 - val_accuracy: 0.7153\n",
|
||
|
|
"Epoch 3/9\n",
|
||
|
|
"351/351 [==============================] - 2s 7ms/step - loss: 0.0232 - accuracy: 0.9978 - val_loss: 1.5496 - val_accuracy: 0.7245\n",
|
||
|
|
"Epoch 4/9\n",
|
||
|
|
"351/351 [==============================] - 2s 7ms/step - loss: 0.0367 - accuracy: 0.9911 - val_loss: 2.0015 - val_accuracy: 0.6828\n",
|
||
|
|
"Epoch 5/9\n",
|
||
|
|
"351/351 [==============================] - 2s 7ms/step - loss: 0.0527 - accuracy: 0.9868 - val_loss: 1.8391 - val_accuracy: 0.6793\n",
|
||
|
|
"Epoch 6/9\n",
|
||
|
|
"351/351 [==============================] - 2s 7ms/step - loss: 0.0343 - accuracy: 0.9920 - val_loss: 1.8650 - val_accuracy: 0.6900\n",
|
||
|
|
"Epoch 7/9\n",
|
||
|
|
"351/351 [==============================] - 2s 7ms/step - loss: 0.0412 - accuracy: 0.9905 - val_loss: 1.8137 - val_accuracy: 0.6878\n",
|
||
|
|
"Epoch 8/9\n",
|
||
|
|
"351/351 [==============================] - 2s 6ms/step - loss: 0.0204 - accuracy: 0.9963 - val_loss: 2.1676 - val_accuracy: 0.6896\n",
|
||
|
|
"Epoch 9/9\n",
|
||
|
|
"351/351 [==============================] - 2s 7ms/step - loss: 0.0298 - accuracy: 0.9931 - val_loss: 2.2046 - val_accuracy: 0.6828\n"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"# fit\n",
|
||
|
|
"basic_history = basic_model.fit(train_data,\n",
|
||
|
|
" validation_data=test_data,\n",
|
||
|
|
" epochs=9,\n",
|
||
|
|
" verbose=1)"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 13,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:30:57.895175Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:30:57.894790Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:30:58.034694Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:30:58.034119Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:30:57.895146Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"image/png": "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
|
||
|
|
"text/plain": [
|
||
|
|
"<Figure size 432x288 with 1 Axes>"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"metadata": {
|
||
|
|
"needs_background": "light"
|
||
|
|
},
|
||
|
|
"output_type": "display_data"
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"plt.plot(basic_history.history['loss'], label='Train loss')\n",
|
||
|
|
"plt.plot(basic_history.history['val_loss'], label='Val loss')\n",
|
||
|
|
"plt.legend();\n",
|
||
|
|
"plt.show()"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 14,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:31:01.220234Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:31:01.219850Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:31:01.353825Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:31:01.353254Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:31:01.220205Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAX8AAAD7CAYAAACCEpQdAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAAAsfUlEQVR4nO3deZgcVb3/8XdVdffsk8lMJntCIpAjQgKEPSwqOwKK7CgEFVFkEVzu9QoaAggCCngjQRDEG0G5ymIECQguFyGIghKWn3JAtoSQkGGyztpb/f7omp6eJZPp2bon9Xk9zzzVfaqq+zvb55w+1V3l+L6PiIiEi1voAkREZOQp/EVEQkjhLyISQgp/EZEQUviLiIRQpNAF9EMJsA+wBkgVuBYRkdHCAyYBzwLt3VeOhvDfB3iy0EWIiIxSBwNPdW8cDeG/BmDDhmbS6fw/k1BXV0ljY9OQFzVYqis/qit/xVqb6srPQOtyXYexYysgyNDuRkP4pwDSaX9A4d+xbzFSXflRXfkr1tpUV34GWVev0+U64CsiEkIKfxGRENrmtI8x5vvAScAMYLa19uVetvGARcDRgA9ca629Y1vrRESkMPoz8l8KHAK83cc2nwZ2AnYGDgAWGmNm9GOdiIgUwDbD31r7lLV21TY2Ow243VqbttY2kOkwTunHOhERKYCherfPdLq+MlgJTOvHun6rq6sccHH19VUD3nc4FWNdvu8XZV1QnD8vKN66oHhrU135GY66RsNbPQFobGwa0Nud6uuraGjYMgwVDc7W6vJ9n1TaJ5FMd36l0sQTKRKpNMmc9njO+kxbqsd+Hdslc9encvbN3TaRBnzGVJYwtqrzq7aqtMv9msoSopGRfa/AaPs9FoNira3Y6kqm0mxpSTBhfBXNW9pG/G97Wwb683Jdp89B81CF/0pgBzIfI4auo/2+1o0Kad8nkUjTnkyRSKSJJ1PEc5eJFPFkznJr7TlLH2htSxLPDewgrAd7fZ2I5xKNBF+eSyyaWXa0VcSi1HgusajXpT0acSkvj7FmXRMbtrTx7vvNvPzmetrjPd8mXF0eZWxHp1BdQm1H51BZwtjqTHtJ1BvcNyIyQL7v09qeZGNTnE3NcTY1tQfLOJua29nYFGdzc5yNTe00tyW77Ou5DmUlEUpjXuYruF0W62iLUFaSWWbWB23ButKSzm2jERfHcQr0U+jbUIX/vcC5xpgHgDrgBDIHibe1blglkin+/Pw7vNfQlBO+HcHdLZQTmRBuD0I9EbS3J9IkU+kBPX/EcymJZkI1FvWIRTxiUZdYxKWyooQxFelsQHcGsNejLRbpGubZbbp/eS6RiIs7iD+23kYZre1J1m9pZ8OWNjZsbmfDlnY2NGWW729q47V3Nvb4BwKoKI0ErxZKg1cQJdRU5XQUVaWUlXhF+88hxSeVTrO5OdEjwLPB3nG7OU4i2fP/NuK51FTGGFMZY0JtObOm11BTEaOqPEZZeYyGxmba4ila40na2lO0xZO0xVM0tyZ4f2MbbfEkrfFUrwOi3niu06MT6d5RZO/ndDhlwbrKsuiwTUX1562ei4ATgYnA740xjdbaXY0xy4AF1trngLuA/YDXgt2utNa+Edzua92wevH19Sz+9Utd2hwHYlGPkiCQO4K5JOJSWhKhuqIjqLsto24mvLfWHiyjUZeSIJxdd+uhVmwvfftSVhJhSkmEKeMqtrpNeyLFxi3tnZ1EcLuj7e33trC5Od5jv5KY1+1VQ9fOYmxVCZVl0eH89ka9tO/T3Jpgc0uCLc1xtrQmaG5NUF1dSmtLnIjnEvGcYNn1tuc5RINl9/We64xYx9zanuwxQt/Y3M7mpjgbc0bsTS0JenthXFEaoaayhOqKGDtPHcOYihLGBCE/pqIkE/gVMcpKIlv9nvL5n0z7Pu3xFG3xzg6itb3rsqO9rT3oTIK2lrYk6ze3dW7Tnur1e+rwnfPmMbmmtF915cMZBdfwnQG8OdA5fycaYcOG5mxYj+QfdF+KNfyHs65kKp3TQbQHHUQbG7O329nY1N5j2iviudSNKaUs5lFRFqWiNBIso1SWRigvjVJRFqGiNEpFWWfbSMzdDsfPK+37tLQl2dISZ0tLgi0t8c5gb0mwuSXeZd2W1sSgpwq3JuI5eJ5LxHWIRFwibtA55N4OOgzPc7fekQSdjOs4JIG1DU050zBx2hM9R9Ke6wThHesM84oYYypLqAmWYypiVFfEhuR3Xaj/Sd/3aU+kunUcKdrak6R9n0P3m8GmjS15P27OnP9M4K3u60fNAd+BGldThp/oOSUhIy/iuYyrKWNcTdlWt+l4Wb++2xRTazxN46YWmluTNGxspaUtSXNb36FXEvUoL810CpXZzqGzkygvjVBZ2rUzqSiLUBId2qmozPxzKhvYHeG9uSO8OwK+ObNsak2Q2spAp6I0QmV5jOryKBNqy9l56hiqymNUlUeprohRVRalqiJGRWmU2toK3mvYQio4lpRK+ySDNw0kO26nOpY9b6dSmeNQqV63ySwz2/i0xVMkU8neHyvtkwyeH6C8NEJ1eYyayhgzJlVRE4R4xyi9I+QryqKDmsIcLRzHCY4fZF69dBcbpmNn2334y+jiuW72XUVM7mzvbVSW9n3a2pM0tSVpbk3Q3JbIdAqtiS5tza2ZjmLt+haa2jJTIsnU1nsNz3W6dgjdOofctglb4ry7dnOP0Xj3YN/a85WVeFSVxaiqiDJuTCkfmFwVhHkm4DuCvWMZ8fo/wi22gU/HO9kmTRxTlK96w0bhL6OW6ziUl0YpL41CH68muvN9n3gyHXQOSVraEjQFHURuZ9EcdCAbmtp5p6GZ5rYEbds40FcS9bJhPbayhOnjq6iqiFJVFqO6IhqEekegR4lGwvOOKMdxiHjb/0h+tFD4S+g4jkNJ1KMk6lFbnd++yVSalvZktuMoKY2RjCeyga+3t8poofAXyUPEc6kORu9QvAfuRbaluD7KJiIiI0LhLyISQgp/EZEQUviLiISQwl9EJIQU/iIiIaTwFxEJIYW/iEgIKfxFREJI4S8iEkIKfxGREFL4i4iEkMJfRCSEFP4iIiGk8BcRCSGFv4hICCn8RURCqF9X8jLGzAKWAHVAIzDfWvtat20mArcBM4EocLW19u5g3ULgfODdYPPl1toLhuIbEBGR/PV35H8rsNhaOwtYTCbku7sReM5aOwc4BLjGGDMtZ/3PrLV7BF8KfhGRAtpm+BtjxgNzgXuCpnuAucaY+m6b7g48CmCtbQBWAKcOWaUiIjJk+jPtMw1Yba1NAVhrU8aYd4P2hpzt/g6cbox5DpgBzAPeyll/ujHmSGAtcLm19i/5FFpXV5nP5l3U11cNeN/hpLryo7ryV6y1qa78DEdd/Zrz76evATeRGfGvBP4IJIJ1t5I5BpAwxhwB/MYYs4u1trG/D97Y2EQ67eddVH19FQ0NW/Leb7iprvyorvwVa22qKz8Drct1nT4Hzf0J/1XAFGOMF4z6PWBy0J4VTPWc2XHfGLMM+Fewbm3Odo8bY1YBuwFP5PG9iIjIENnmnL+1dh2Z0fwZQdMZwPNB2GcZY+qMMZHg9qHAbOAXwf0pOdvtQWZayA66ehERGZD+TvucBywxxiwANgDzITu6X2CtfQ7YF1hkjEkB7wPHW2tbgv2vMcbsBaSAOHBW7qsBEREZWf0Kf2vtK8B+vbR/LOf2I8DOW9n/7IEWKCIiQ0+f8BURCSGFv4hICCn8RURCSOEvIhJCCn8RkRBS+IuIhJDCX0QkhBT+IiIhpPAXEQkhhb+ISAgp/EVEQkjhLyISQgp/EZEQUviLiISQwl9EJIQU/iIiIaTwFxEJIYW/iEgIKfxFREJI4S8iEkIKfxGREFL4i4iEUKQ/GxljZgFLgDqgEZhvrX2t2zYTgduAmUAUuNpae3ewzgMWAUcDPnCttfaOofomREQkP/0d+d8KLLbWzgIWkwn57m4EnrPWzgEOAa4xxkwL1n0a2AnYGTgAWGiMmTGYwkVEZOC2Gf7GmPHAXOCeoOkeYK4xpr7bprsDjwJYaxuAFcCpwbrTgNuttelg3VLglMEWLyI
|
||
|
|
"text/plain": [
|
||
|
|
"<Figure size 432x288 with 1 Axes>"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"metadata": {
|
||
|
|
"needs_background": "light"
|
||
|
|
},
|
||
|
|
"output_type": "display_data"
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"plt.plot(basic_history.history['accuracy'], label='Train accuracy')\n",
|
||
|
|
"plt.plot(basic_history.history['val_accuracy'], label='Val accuracy')\n",
|
||
|
|
"plt.legend();\n",
|
||
|
|
"plt.show()"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"This model doesn't seem to be performing very well"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {
|
||
|
|
"id": "1sGS2tyjKhVJ"
|
||
|
|
},
|
||
|
|
"source": [
|
||
|
|
"### Prebuilt Networks\n",
|
||
|
|
"\n",
|
||
|
|
"Today we'll use `EfficientNet`. You can read more about the keras implementation of `EfficientNet` [here](https://keras.io/examples/vision/image_classification_efficientnet_fine_tuning/#:~:text=Keras%20implementation%20of%20EfficientNet)."
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {
|
||
|
|
"id": "XzBFnZUncGAJ"
|
||
|
|
},
|
||
|
|
"source": [
|
||
|
|
"#### Import EfficientNet and freeze"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 15,
|
||
|
|
"metadata": {
|
||
|
|
"colab": {
|
||
|
|
"base_uri": "https://localhost:8080/"
|
||
|
|
},
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:31:09.035471Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:31:09.035203Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:31:14.664011Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:31:14.663373Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:31:09.035455Z"
|
||
|
|
},
|
||
|
|
"id": "WIhFD_gKcGAK",
|
||
|
|
"outputId": "c400468e-c3fe-4497-9842-79f386fa63f7",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [],
|
||
|
|
"source": [
|
||
|
|
"#top = false so we don't get the output layer of effnet\n",
|
||
|
|
"effnet = EfficientNetV2L(include_top=False, weights='imagenet')"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 16,
|
||
|
|
"metadata": {
|
||
|
|
"colab": {
|
||
|
|
"base_uri": "https://localhost:8080/"
|
||
|
|
},
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:31:14.665176Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:31:14.664960Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:31:14.688847Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:31:14.688260Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:31:14.665161Z"
|
||
|
|
},
|
||
|
|
"id": "iH3UYtn6cGAK",
|
||
|
|
"outputId": "6d71c7a2-9efb-4b5c-a46f-6a3bb10321f1",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"text/plain": [
|
||
|
|
"False"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"execution_count": 16,
|
||
|
|
"metadata": {},
|
||
|
|
"output_type": "execute_result"
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"effnet.trainable = False\n",
|
||
|
|
"effnet.trainable"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {
|
||
|
|
"id": "KxV0EayjcGAK"
|
||
|
|
},
|
||
|
|
"source": [
|
||
|
|
"\n",
|
||
|
|
"#### Add EfficientNet to a Sequential model\n",
|
||
|
|
"\n",
|
||
|
|
"Between `EfficientNet` and our first Dense layer, we will need to use `GlobalAveragePooling2D()`. (This appears to be related to a bug inside of keras; you can read more [here](https://stackoverflow.com/questions/48851558/tensorflow-estimator-valueerror-logits-and-labels-must-have-the-same-shape).)"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 17,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:31:24.580846Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:31:24.580321Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:31:26.456824Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:31:26.456030Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:31:24.580817Z"
|
||
|
|
},
|
||
|
|
"id": "XsOqCzq_tJZ_",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [],
|
||
|
|
"source": [
|
||
|
|
"eff_model = Sequential()\n",
|
||
|
|
"eff_model.add(effnet)\n",
|
||
|
|
"eff_model.add(GlobalAveragePooling2D())\n",
|
||
|
|
"eff_model.add(Dense(6, activation='softmax'))"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 18,
|
||
|
|
"metadata": {
|
||
|
|
"colab": {
|
||
|
|
"base_uri": "https://localhost:8080/"
|
||
|
|
},
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:31:26.458355Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:31:26.458112Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:31:26.494031Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:31:26.493375Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:31:26.458334Z"
|
||
|
|
},
|
||
|
|
"id": "J-QUvl1fx3XG",
|
||
|
|
"outputId": "c146558d-157a-42c2-a956-132375724116",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"name": "stdout",
|
||
|
|
"output_type": "stream",
|
||
|
|
"text": [
|
||
|
|
"Model: \"sequential_1\"\n",
|
||
|
|
"_________________________________________________________________\n",
|
||
|
|
" Layer (type) Output Shape Param # \n",
|
||
|
|
"=================================================================\n",
|
||
|
|
" efficientnetv2-l (Functiona (None, None, None, 1280) 117746848\n",
|
||
|
|
" l) \n",
|
||
|
|
" \n",
|
||
|
|
" global_average_pooling2d (G (None, 1280) 0 \n",
|
||
|
|
" lobalAveragePooling2D) \n",
|
||
|
|
" \n",
|
||
|
|
" dense_2 (Dense) (None, 6) 7686 \n",
|
||
|
|
" \n",
|
||
|
|
"=================================================================\n",
|
||
|
|
"Total params: 117,754,534\n",
|
||
|
|
"Trainable params: 7,686\n",
|
||
|
|
"Non-trainable params: 117,746,848\n",
|
||
|
|
"_________________________________________________________________\n"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"eff_model.summary()"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {
|
||
|
|
"id": "Qkgpfj-9cGAK"
|
||
|
|
},
|
||
|
|
"source": [
|
||
|
|
"#### Compile and evaluate"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 19,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:31:54.783074Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:31:54.782685Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:31:54.811131Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:31:54.810534Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:31:54.783046Z"
|
||
|
|
},
|
||
|
|
"id": "viuBcMiEcGAK",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [],
|
||
|
|
"source": [
|
||
|
|
"eff_model.compile(optimizer=Adam(learning_rate=0.0001),\n",
|
||
|
|
" loss='categorical_crossentropy', \n",
|
||
|
|
" metrics=['accuracy'])"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 20,
|
||
|
|
"metadata": {
|
||
|
|
"colab": {
|
||
|
|
"base_uri": "https://localhost:8080/"
|
||
|
|
},
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-17T21:31:56.504469Z",
|
||
|
|
"iopub.status.busy": "2022-07-17T21:31:56.504198Z",
|
||
|
|
"iopub.status.idle": "2022-07-17T21:37:11.005343Z",
|
||
|
|
"shell.execute_reply": "2022-07-17T21:37:11.004527Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-17T21:31:56.504453Z"
|
||
|
|
},
|
||
|
|
"id": "-IDTcsTOoYZK",
|
||
|
|
"outputId": "3db8b63d-04b9-41fe-f14a-ec1b32f248a6",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"name": "stdout",
|
||
|
|
"output_type": "stream",
|
||
|
|
"text": [
|
||
|
|
"Epoch 1/10\n",
|
||
|
|
"351/351 [==============================] - 52s 100ms/step - loss: 1.1717 - accuracy: 0.6575 - val_loss: 0.7577 - val_accuracy: 0.8357\n",
|
||
|
|
"Epoch 2/10\n",
|
||
|
|
"351/351 [==============================] - 29s 83ms/step - loss: 0.6484 - accuracy: 0.8325 - val_loss: 0.5315 - val_accuracy: 0.8660\n",
|
||
|
|
"Epoch 3/10\n",
|
||
|
|
"351/351 [==============================] - 29s 83ms/step - loss: 0.5048 - accuracy: 0.8555 - val_loss: 0.4439 - val_accuracy: 0.8753\n",
|
||
|
|
"Epoch 4/10\n",
|
||
|
|
"351/351 [==============================] - 29s 83ms/step - loss: 0.4414 - accuracy: 0.8636 - val_loss: 0.3968 - val_accuracy: 0.8813\n",
|
||
|
|
"Epoch 5/10\n",
|
||
|
|
"351/351 [==============================] - 29s 83ms/step - loss: 0.3988 - accuracy: 0.8726 - val_loss: 0.3675 - val_accuracy: 0.8842\n",
|
||
|
|
"Epoch 6/10\n",
|
||
|
|
"351/351 [==============================] - 29s 83ms/step - loss: 0.3717 - accuracy: 0.8779 - val_loss: 0.3467 - val_accuracy: 0.8917\n",
|
||
|
|
"Epoch 7/10\n",
|
||
|
|
"351/351 [==============================] - 29s 83ms/step - loss: 0.3538 - accuracy: 0.8822 - val_loss: 0.3318 - val_accuracy: 0.8963\n",
|
||
|
|
"Epoch 8/10\n",
|
||
|
|
"351/351 [==============================] - 29s 83ms/step - loss: 0.3363 - accuracy: 0.8897 - val_loss: 0.3201 - val_accuracy: 0.8988\n",
|
||
|
|
"Epoch 9/10\n",
|
||
|
|
"351/351 [==============================] - 29s 83ms/step - loss: 0.3265 - accuracy: 0.8887 - val_loss: 0.3111 - val_accuracy: 0.8999\n",
|
||
|
|
"Epoch 10/10\n",
|
||
|
|
"351/351 [==============================] - 29s 83ms/step - loss: 0.3149 - accuracy: 0.8949 - val_loss: 0.3035 - val_accuracy: 0.9041\n"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"eff_history = eff_model.fit(train_data, validation_data=test_data, epochs=10)"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 33,
|
||
|
|
"metadata": {
|
||
|
|
"colab": {
|
||
|
|
"base_uri": "https://localhost:8080/",
|
||
|
|
"height": 265
|
||
|
|
},
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-06T22:39:29.986688Z",
|
||
|
|
"iopub.status.busy": "2022-07-06T22:39:29.986347Z",
|
||
|
|
"iopub.status.idle": "2022-07-06T22:39:30.103357Z",
|
||
|
|
"shell.execute_reply": "2022-07-06T22:39:30.102578Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-06T22:39:29.986670Z"
|
||
|
|
},
|
||
|
|
"id": "5RM9RSYRoYWs",
|
||
|
|
"outputId": "a43236c4-2b16-42da-f5cb-b9c08c4a87fe",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"image/png": "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
|
||
|
|
"text/plain": [
|
||
|
|
"<Figure size 432x288 with 1 Axes>"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"metadata": {
|
||
|
|
"needs_background": "light"
|
||
|
|
},
|
||
|
|
"output_type": "display_data"
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"plt.plot(eff_history.history['loss'], label='Train loss')\n",
|
||
|
|
"plt.plot(eff_history.history['val_loss'], label='Val loss')\n",
|
||
|
|
"plt.legend();\n",
|
||
|
|
"plt.show()"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 34,
|
||
|
|
"metadata": {
|
||
|
|
"colab": {
|
||
|
|
"base_uri": "https://localhost:8080/",
|
||
|
|
"height": 265
|
||
|
|
},
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-06T22:39:30.124496Z",
|
||
|
|
"iopub.status.busy": "2022-07-06T22:39:30.124010Z",
|
||
|
|
"iopub.status.idle": "2022-07-06T22:39:30.217731Z",
|
||
|
|
"shell.execute_reply": "2022-07-06T22:39:30.217203Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-06T22:39:30.124465Z"
|
||
|
|
},
|
||
|
|
"id": "5RM9RSYRoYWs",
|
||
|
|
"outputId": "a43236c4-2b16-42da-f5cb-b9c08c4a87fe",
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"image/png": "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
|
||
|
|
"text/plain": [
|
||
|
|
"<Figure size 432x288 with 1 Axes>"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"metadata": {
|
||
|
|
"needs_background": "light"
|
||
|
|
},
|
||
|
|
"output_type": "display_data"
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"plt.plot(eff_history.history['accuracy'], label='Train accuracy')\n",
|
||
|
|
"plt.plot(eff_history.history['val_accuracy'], label='Val accuracy')\n",
|
||
|
|
"plt.legend();\n",
|
||
|
|
"plt.show()"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"It looks like this could train for a little while longer but we'll call it good here for fear of overfitting"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"# Confusion matrix\n",
|
||
|
|
"First create some data for it:"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 18,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-06T22:23:05.431294Z",
|
||
|
|
"iopub.status.busy": "2022-07-06T22:23:05.431089Z",
|
||
|
|
"iopub.status.idle": "2022-07-06T22:23:06.142409Z",
|
||
|
|
"shell.execute_reply": "2022-07-06T22:23:06.141756Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-06T22:23:05.431278Z"
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"name": "stdout",
|
||
|
|
"output_type": "stream",
|
||
|
|
"text": [
|
||
|
|
"Found 14034 files belonging to 6 classes.\n"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"all_data = image_dataset_from_directory(directory=train_path,\n",
|
||
|
|
" labels='inferred',\n",
|
||
|
|
" label_mode='categorical',\n",
|
||
|
|
" class_names=class_names,\n",
|
||
|
|
" color_mode='rgb',\n",
|
||
|
|
" batch_size=32,\n",
|
||
|
|
" image_size=(150, 150),\n",
|
||
|
|
" shuffle=False,\n",
|
||
|
|
" seed=42)"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 40,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-06T22:47:05.275883Z",
|
||
|
|
"iopub.status.busy": "2022-07-06T22:47:05.275387Z",
|
||
|
|
"iopub.status.idle": "2022-07-06T22:48:03.127611Z",
|
||
|
|
"shell.execute_reply": "2022-07-06T22:48:03.127049Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-06T22:47:05.275853Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"name": "stdout",
|
||
|
|
"output_type": "stream",
|
||
|
|
"text": [
|
||
|
|
"439/439 [==============================] - 57s 129ms/step\n",
|
||
|
|
"Confusion Matrix\n",
|
||
|
|
"[[8.89548152e-01 4.56412597e-03 1.82565039e-03 1.04974897e-02\n",
|
||
|
|
" 9.12825194e-03 8.44363304e-02]\n",
|
||
|
|
" [1.32100396e-03 9.81946279e-01 8.80669309e-04 1.10083664e-02\n",
|
||
|
|
" 3.08234258e-03 1.76133862e-03]\n",
|
||
|
|
" [3.74376040e-03 9.56738769e-03 8.05324459e-01 1.33943428e-01\n",
|
||
|
|
" 4.40931780e-02 3.32778702e-03]\n",
|
||
|
|
" [1.19426752e-03 3.98089172e-03 1.39331210e-01 8.34394904e-01\n",
|
||
|
|
" 2.03025478e-02 7.96178344e-04]\n",
|
||
|
|
" [6.59630607e-03 3.95778364e-03 1.75901495e-02 1.05540897e-02\n",
|
||
|
|
" 9.58663149e-01 2.63852243e-03]\n",
|
||
|
|
" [4.99580185e-02 2.93870697e-03 4.19815281e-04 5.03778338e-03\n",
|
||
|
|
" 3.35852225e-03 9.38287154e-01]]\n",
|
||
|
|
"Classification Report\n",
|
||
|
|
" precision recall f1-score support\n",
|
||
|
|
"\n",
|
||
|
|
" buildings 0.93 0.89 0.91 2191\n",
|
||
|
|
" forest 0.97 0.98 0.98 2271\n",
|
||
|
|
" glacier 0.83 0.81 0.82 2404\n",
|
||
|
|
" mountain 0.84 0.83 0.84 2512\n",
|
||
|
|
" sea 0.92 0.96 0.94 2274\n",
|
||
|
|
" street 0.92 0.94 0.93 2382\n",
|
||
|
|
"\n",
|
||
|
|
" accuracy 0.90 14034\n",
|
||
|
|
" macro avg 0.90 0.90 0.90 14034\n",
|
||
|
|
"weighted avg 0.90 0.90 0.90 14034\n",
|
||
|
|
"\n"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"y = np.concatenate([y for x, y in all_data], axis=0)\n",
|
||
|
|
"\n",
|
||
|
|
"y_pred_array = eff_model.predict(all_data)\n",
|
||
|
|
"y_pred = np.argmax(y_pred_array, axis=1)\n",
|
||
|
|
"\n",
|
||
|
|
"#Create confusion matrix\n",
|
||
|
|
"print('Confusion Matrix')\n",
|
||
|
|
"matrix = confusion_matrix(np.where(y==1)[1], y_pred,normalize='true')\n",
|
||
|
|
"print(matrix)\n",
|
||
|
|
"print('Classification Report')\n",
|
||
|
|
"print(classification_report(np.where(y==1)[1], y_pred, \n",
|
||
|
|
" target_names=class_names))"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"## Create heatmap "
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 59,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-06T22:56:17.443695Z",
|
||
|
|
"iopub.status.busy": "2022-07-06T22:56:17.443185Z",
|
||
|
|
"iopub.status.idle": "2022-07-06T22:56:17.673770Z",
|
||
|
|
"shell.execute_reply": "2022-07-06T22:56:17.673084Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-06T22:56:17.443661Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"image/png": "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
|
||
|
|
"text/plain": [
|
||
|
|
"<Figure size 720x360 with 2 Axes>"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"metadata": {},
|
||
|
|
"output_type": "display_data"
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"sns.set(rc = {'figure.figsize':(10,5)})\n",
|
||
|
|
"sns.heatmap(matrix, annot=True,fmt='.2f', cmap=\"Blues\");"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "markdown",
|
||
|
|
"metadata": {},
|
||
|
|
"source": [
|
||
|
|
"### Can you tell the difference between these pictures? Our model can!"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 24,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-06T22:41:17.462231Z",
|
||
|
|
"iopub.status.busy": "2022-07-06T22:41:17.461836Z",
|
||
|
|
"iopub.status.idle": "2022-07-06T22:41:17.584227Z",
|
||
|
|
"shell.execute_reply": "2022-07-06T22:41:17.583586Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-06T22:41:17.462205Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"image/png": "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
|
||
|
|
"text/plain": [
|
||
|
|
"<Figure size 432x288 with 1 Axes>"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"metadata": {
|
||
|
|
"needs_background": "light"
|
||
|
|
},
|
||
|
|
"output_type": "display_data"
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"mountain = random.choice(os.listdir(\"data/seg_train/mountain\"))\n",
|
||
|
|
"plt.imshow(plt.imread(f'data/seg_train/mountain/{mountain}'));\n",
|
||
|
|
"plt.show()"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 13,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-06T22:40:21.212301Z",
|
||
|
|
"iopub.status.busy": "2022-07-06T22:40:21.211546Z",
|
||
|
|
"iopub.status.idle": "2022-07-06T22:40:21.343654Z",
|
||
|
|
"shell.execute_reply": "2022-07-06T22:40:21.343049Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-06T22:40:21.212265Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"image/png": "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
|
||
|
|
"text/plain": [
|
||
|
|
"<Figure size 432x288 with 1 Axes>"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"metadata": {
|
||
|
|
"needs_background": "light"
|
||
|
|
},
|
||
|
|
"output_type": "display_data"
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"glacier = random.choice(os.listdir(\"data/seg_train/glacier\"))\n",
|
||
|
|
"plt.imshow(plt.imread(f'data/seg_train/glacier/{glacier}'));\n",
|
||
|
|
"plt.show()"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 31,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-06T22:42:17.308200Z",
|
||
|
|
"iopub.status.busy": "2022-07-06T22:42:17.307758Z",
|
||
|
|
"iopub.status.idle": "2022-07-06T22:42:17.429258Z",
|
||
|
|
"shell.execute_reply": "2022-07-06T22:42:17.428820Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-06T22:42:17.308173Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"image/png": "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
|
||
|
|
"text/plain": [
|
||
|
|
"<Figure size 432x288 with 1 Axes>"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"metadata": {
|
||
|
|
"needs_background": "light"
|
||
|
|
},
|
||
|
|
"output_type": "display_data"
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"buildings = random.choice(os.listdir(\"data/seg_train/buildings\"))\n",
|
||
|
|
"plt.imshow(plt.imread(f'data/seg_train/buildings/{buildings}'));\n",
|
||
|
|
"plt.show()"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 29,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-06T22:42:02.906854Z",
|
||
|
|
"iopub.status.busy": "2022-07-06T22:42:02.906295Z",
|
||
|
|
"iopub.status.idle": "2022-07-06T22:42:03.045861Z",
|
||
|
|
"shell.execute_reply": "2022-07-06T22:42:03.045228Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-06T22:42:02.906820Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"image/png": "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
|
||
|
|
"text/plain": [
|
||
|
|
"<Figure size 432x288 with 1 Axes>"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"metadata": {
|
||
|
|
"needs_background": "light"
|
||
|
|
},
|
||
|
|
"output_type": "display_data"
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"street = random.choice(os.listdir(\"data/seg_train/street\"))\n",
|
||
|
|
"plt.imshow(plt.imread(f'data/seg_train/street/{street}'));\n",
|
||
|
|
"plt.show()"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"cell_type": "code",
|
||
|
|
"execution_count": 60,
|
||
|
|
"metadata": {
|
||
|
|
"execution": {
|
||
|
|
"iopub.execute_input": "2022-07-06T23:54:11.547266Z",
|
||
|
|
"iopub.status.busy": "2022-07-06T23:54:11.546755Z",
|
||
|
|
"iopub.status.idle": "2022-07-06T23:54:11.552724Z",
|
||
|
|
"shell.execute_reply": "2022-07-06T23:54:11.551926Z",
|
||
|
|
"shell.execute_reply.started": "2022-07-06T23:54:11.547239Z"
|
||
|
|
},
|
||
|
|
"tags": []
|
||
|
|
},
|
||
|
|
"outputs": [
|
||
|
|
{
|
||
|
|
"data": {
|
||
|
|
"text/plain": [
|
||
|
|
"16.666666666666668"
|
||
|
|
]
|
||
|
|
},
|
||
|
|
"execution_count": 60,
|
||
|
|
"metadata": {},
|
||
|
|
"output_type": "execute_result"
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"source": [
|
||
|
|
"100/6"
|
||
|
|
]
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"metadata": {
|
||
|
|
"accelerator": "GPU",
|
||
|
|
"colab": {
|
||
|
|
"collapsed_sections": [],
|
||
|
|
"name": "Team ZAB Intel Image Classification",
|
||
|
|
"provenance": []
|
||
|
|
},
|
||
|
|
"gpuClass": "standard",
|
||
|
|
"kernelspec": {
|
||
|
|
"display_name": "Python 3 (ipykernel)",
|
||
|
|
"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": "3.10.5"
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"nbformat": 4,
|
||
|
|
"nbformat_minor": 4
|
||
|
|
}
|