Skip to content

Fix issue 163 and 164 in 1_feature_extraction.ipynb #171

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
+584 −635
Open

Fix issue 163 and 164 in 1_feature_extraction.ipynb #171

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1,219 changes: 584 additions & 635 deletions code/chapter-4/1_feature_extraction.ipynb
View file
Original file line number Diff line number Diff line change
@@ -1,293 +1,4 @@
{
"nbformat": 4,
"nbformat_minor": 0,
"metadata": {
"kernelspec": {
"display_name": "Python 2",
"language": "python",
"name": "python2"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 2
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython2",
"version": "2.7.10"
},
"toc": {
"base_numbering": 1,
"nav_menu": {},
"number_sections": true,
"sideBar": true,
"skip_h1_title": false,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
},
"colab": {
"name": "Copy of 1-feature-extraction.ipynb",
"provenance": [],
"collapsed_sections": []
},
"accelerator": "GPU",
"widgets": {
"application/vnd.jupyter.widget-state+json": {
"2f0c84e6e48b40cb9f1360ba4451513c": {
"model_module": "@jupyter-widgets/controls",
"model_name": "HBoxModel",
"state": {
"_view_name": "HBoxView",
"_dom_classes": [],
"_model_name": "HBoxModel",
"_view_module": "@jupyter-widgets/controls",
"_model_module_version": "1.5.0",
"_view_count": null,
"_view_module_version": "1.5.0",
"box_style": "",
"layout": "IPY_MODEL_470d8a8b00c5475684ab84b55c79b760",
"_model_module": "@jupyter-widgets/controls",
"children": [
"IPY_MODEL_a0ffda7bcf1a46a4bdd6352c43cfda61",
"IPY_MODEL_f80ec0bf78de458995da4f9c964f3722"
]
}
},
"470d8a8b00c5475684ab84b55c79b760": {
"model_module": "@jupyter-widgets/base",
"model_name": "LayoutModel",
"state": {
"_view_name": "LayoutView",
"grid_template_rows": null,
"right": null,
"justify_content": null,
"_view_module": "@jupyter-widgets/base",
"_model_module_version": "1.2.0",
"_view_count": null,
"flex_flow": null,
"min_width": null,
"border": null,
"align_items": null,
"bottom": null,
"_model_module": "@jupyter-widgets/base",
"top": null,
"grid_column": null,
"width": null,
"overflow_x": null,
"grid_auto_flow": null,
"grid_area": null,
"grid_template_columns": null,
"flex": null,
"_model_name": "LayoutModel",
"justify_items": null,
"grid_row": null,
"max_height": null,
"align_content": null,
"visibility": null,
"overflow": null,
"height": null,
"min_height": null,
"padding": null,
"grid_auto_rows": null,
"grid_gap": null,
"overflow_y": null,
"max_width": null,
"order": null,
"_view_module_version": "1.2.0",
"align_self": null,
"grid_template_areas": null,
"object_position": null,
"object_fit": null,
"grid_auto_columns": null,
"margin": null,
"display": null,
"left": null
}
},
"a0ffda7bcf1a46a4bdd6352c43cfda61": {
"model_module": "@jupyter-widgets/controls",
"model_name": "IntProgressModel",
"state": {
"_view_name": "ProgressView",
"style": "IPY_MODEL_9792cd696fa74cc68ef6b35095fb8685",
"_view_module": "@jupyter-widgets/controls",
"_dom_classes": [],
"orientation": "horizontal",
"min": 0,
"bar_style": "success",
"max": 8677,
"_model_name": "IntProgressModel",
"_model_module_version": "1.5.0",
"value": 8677,
"_view_count": null,
"_view_module_version": "1.5.0",
"description_tooltip": null,
"_model_module": "@jupyter-widgets/controls",
"layout": "IPY_MODEL_4d80db8732434d119c3f3b3bb8469ce9",
"description": ""
}
},
"f80ec0bf78de458995da4f9c964f3722": {
"model_module": "@jupyter-widgets/controls",
"model_name": "HTMLModel",
"state": {
"_view_name": "HTMLView",
"style": "IPY_MODEL_fe1544db2c8e4df2a5c7097575da68ab",
"_dom_classes": [],
"description": "",
"_model_name": "HTMLModel",
"placeholder": "​",
"_view_module": "@jupyter-widgets/controls",
"_model_module_version": "1.5.0",
"value": "100% 8677/8677 [06:16<00:00, 23.04it/s]",
"_view_count": null,
"_view_module_version": "1.5.0",
"description_tooltip": null,
"_model_module": "@jupyter-widgets/controls",
"layout": "IPY_MODEL_145e7cf97a5844a48e321c8a4878d13b"
}
},
"9792cd696fa74cc68ef6b35095fb8685": {
"model_module": "@jupyter-widgets/controls",
"model_name": "ProgressStyleModel",
"state": {
"_view_name": "StyleView",
"_model_name": "ProgressStyleModel",
"description_width": "",
"_view_module": "@jupyter-widgets/base",
"_model_module_version": "1.5.0",
"_view_count": null,
"_view_module_version": "1.2.0",
"bar_color": null,
"_model_module": "@jupyter-widgets/controls"
}
},
"4d80db8732434d119c3f3b3bb8469ce9": {
"model_module": "@jupyter-widgets/base",
"model_name": "LayoutModel",
"state": {
"_view_name": "LayoutView",
"grid_template_rows": null,
"right": null,
"justify_content": null,
"_view_module": "@jupyter-widgets/base",
"_model_module_version": "1.2.0",
"_view_count": null,
"flex_flow": null,
"min_width": null,
"border": null,
"align_items": null,
"bottom": null,
"_model_module": "@jupyter-widgets/base",
"top": null,
"grid_column": null,
"width": null,
"overflow_x": null,
"grid_auto_flow": null,
"grid_area": null,
"grid_template_columns": null,
"flex": null,
"_model_name": "LayoutModel",
"justify_items": null,
"grid_row": null,
"max_height": null,
"align_content": null,
"visibility": null,
"overflow": null,
"height": null,
"min_height": null,
"padding": null,
"grid_auto_rows": null,
"grid_gap": null,
"overflow_y": null,
"max_width": null,
"order": null,
"_view_module_version": "1.2.0",
"align_self": null,
"grid_template_areas": null,
"object_position": null,
"object_fit": null,
"grid_auto_columns": null,
"margin": null,
"display": null,
"left": null
}
},
"fe1544db2c8e4df2a5c7097575da68ab": {
"model_module": "@jupyter-widgets/controls",
"model_name": "DescriptionStyleModel",
"state": {
"_view_name": "StyleView",
"_model_name": "DescriptionStyleModel",
"description_width": "",
"_view_module": "@jupyter-widgets/base",
"_model_module_version": "1.5.0",
"_view_count": null,
"_view_module_version": "1.2.0",
"_model_module": "@jupyter-widgets/controls"
}
},
"145e7cf97a5844a48e321c8a4878d13b": {
"model_module": "@jupyter-widgets/base",
"model_name": "LayoutModel",
"state": {
"_view_name": "LayoutView",
"grid_template_rows": null,
"right": null,
"justify_content": null,
"_view_module": "@jupyter-widgets/base",
"_model_module_version": "1.2.0",
"_view_count": null,
"flex_flow": null,
"min_width": null,
"border": null,
"align_items": null,
"bottom": null,
"_model_module": "@jupyter-widgets/base",
"top": null,
"grid_column": null,
"width": null,
"overflow_x": null,
"grid_auto_flow": null,
"grid_area": null,
"grid_template_columns": null,
"flex": null,
"_model_name": "LayoutModel",
"justify_items": null,
"grid_row": null,
"max_height": null,
"align_content": null,
"visibility": null,
"overflow": null,
"height": null,
"min_height": null,
"padding": null,
"grid_auto_rows": null,
"grid_gap": null,
"overflow_y": null,
"max_width": null,
"order": null,
"_view_module_version": "1.2.0",
"align_self": null,
"grid_template_areas": null,
"object_position": null,
"object_fit": null,
"grid_auto_columns": null,
"margin": null,
"display": null,
"left": null
}
}
}
}
},
"cells": [
{
"cell_type": "markdown",
@@ -342,28 +53,30 @@
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true,
"id": "-_byh3zWnz6c"
},
"outputs": [],
"source": [
"!mkdir -p ../../datasets\n",
"!pip install gdown\n",
"!gdown https://drive.google.com/uc?id=137RyRjvTBkBiIfeYBNZBtViDHQ6_Ewsp --output ../../datasets/caltech101.tar.gz\n",
"!gdown \"https://drive.google.com/uc?id=137RyRjvTBkBiIfeYBNZBtViDHQ6_Ewsp\" --output ../../datasets/caltech101.tar.gz\n",
"!tar -xvzf ../../datasets/caltech101.tar.gz --directory ../../datasets\n",
"!mv ../../datasets/101_ObjectCategories ../../datasets/caltech101\n",
"!rm -rf ../../datasets/caltech101/BACKGROUND_Google"
],
"execution_count": null,
"outputs": []
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {
"collapsed": true,
"scrolled": false,
"id": "yN9yirdLnz6d"
"id": "yN9yirdLnz6d",
"scrolled": false
},
"outputs": [],
"source": [
"import numpy as np\n",
"from numpy.linalg import norm\n",
@@ -383,9 +96,7 @@
"from tensorflow.keras.applications.inception_v3 import InceptionV3\n",
"from tensorflow.keras.models import Model\n",
"from tensorflow.keras.layers import Input, Flatten, Dense, Dropout, GlobalAveragePooling2D\n"
],
"execution_count": 2,
"outputs": []
]
},
{
"cell_type": "markdown",
@@ -398,10 +109,12 @@
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {
"collapsed": true,
"id": "n39Tv0YZnz6e"
},
"outputs": [],
"source": [
"def model_picker(name):\n",
" if (name == 'vgg16'):\n",
@@ -439,9 +152,7 @@
" else:\n",
" print(\"Specified model not available\")\n",
" return model"
],
"execution_count": 3,
"outputs": []
]
},
{
"cell_type": "markdown",
@@ -454,29 +165,19 @@
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {
"collapsed": true,
"colab": {
"base_uri": "https://localhost:8080/"
},
"collapsed": true,
"id": "UCz180_jnz6g",
"outputId": "b6d47be1-b290-46b1-9e54-235b5ba0d2d3"
},
"source": [
"model_architecture = 'resnet'\n",
"model = model_picker(model_architecture)"
],
"execution_count": 4,
"outputs": [
{
"output_type": "stream",
"text": [
"Downloading data from https://github.com/keras-team/keras-applications/releases/download/resnet/resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
"94773248/94765736 [==============================] - 2s 0us/step\n",
"94781440/94765736 [==============================] - 2s 0us/step\n"
],
"name": "stdout"
}
},
"outputs": [],
"source": [
"model_architecture = 'resnet'\n",
"model = model_picker(model_architecture)"
]
},
{
@@ -490,10 +191,12 @@
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {
"collapsed": true,
"id": "2Fc9vtRenz6g"
},
"outputs": [],
"source": [
"def extract_features(img_path, model):\n",
" input_shape = (224, 224, 3)\n",
@@ -506,9 +209,7 @@
" flattened_features = features.flatten()\n",
" normalized_features = flattened_features / norm(flattened_features)\n",
" return normalized_features"
],
"execution_count": 5,
"outputs": []
]
},
{
"cell_type": "markdown",
@@ -521,70 +222,61 @@
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {
"id": "qC_isBQ7okco"
},
"outputs": [],
"source": [
"try:\n",
" import google.colab\n",
" IS_COLAB_ENV = True\n",
"except:\n",
" IS_COLAB_ENV = False"
],
"execution_count": 6,
"outputs": []
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "jU3c7aAbqR-P",
"outputId": "7cc823ea-5d84-4eb6-ab0a-acc777ba21be"
},
"outputs": [],
"source": [
"IMG_PATH = '../../sample-images/cat.jpg'\n",
"if IS_COLAB_ENV:\n",
" !curl https://raw.githubusercontent.com/PracticalDL/Practical-Deep-Learning-Book/master/sample-images/cat.jpg --output cat.jpg\n",
" IMG_PATH = 'cat.jpg'"
],
"execution_count": 7,
"outputs": [
{
"output_type": "stream",
"text": [
" % Total % Received % Xferd Average Speed Time Time Time Current\n",
" Dload Upload Total Spent Left Speed\n",
"\r 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0\r100 661k 100 661k 0 0 6964k 0 --:--:-- --:--:-- --:--:-- 6964k\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {
"collapsed": true,
"colab": {
"base_uri": "https://localhost:8080/"
},
"collapsed": true,
"id": "_USxAyninz6h",
"outputId": "c2c588db-95cc-4d55-a686-1ce5a11bc99f"
},
"source": [
"features = extract_features('cat.jpg', model)\n",
"print(\"Total length of features for one image: \", len(features))"
],
"execution_count": 8,
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"2048\n"
],
"name": "stdout"
"1/1 [==============================] - 1s 1s/step\n",
"Total length of features for one image: 2048\n"
]
}
],
"source": [
"features = extract_features(IMG_PATH, model)\n",
"print(\"Total length of features for one image: \", len(features))"
]
},
{
@@ -598,26 +290,18 @@
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true,
"colab": {
"base_uri": "https://localhost:8080/"
},
"collapsed": true,
"id": "Rs8nZZCLnz6i",
"outputId": "cb7d8dba-ec9c-4fc0-ee23-864f2e66bfb9"
},
"outputs": [],
"source": [
"%timeit features = extract_features('cat.jpg', model)"
],
"execution_count": 9,
"outputs": [
{
"output_type": "stream",
"text": [
"10 loops, best of 3: 81.4 ms per loop\n"
],
"name": "stdout"
}
"%timeit features = extract_features(IMG_PATH, model)"
]
},
{
@@ -642,10 +326,12 @@
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {
"collapsed": true,
"id": "biSAC49dnz6j"
},
"outputs": [],
"source": [
"extensions = ['.jpg', '.JPG', '.jpeg', '.JPEG', '.png', '.PNG']\n",
"\n",
@@ -660,9 +346,7 @@
" else:\n",
" print(filepath)\n",
" return file_list"
],
"execution_count": 10,
"outputs": []
]
},
{
"cell_type": "markdown",
@@ -675,34 +359,34 @@
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "XC75Z2Dau77Z",
"outputId": "9140e060-cd2a-42a2-fd6c-a1d782822e28"
},
"source": [
"# path to the your datasets\n",
"root_dir = '../../datasets/caltech101'\n",
"filenames = sorted(get_file_list(root_dir))\n",
"print(len(filenames))"
],
"execution_count": 11,
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"8677\n"
],
"name": "stdout"
]
}
],
"source": [
"# path to the your datasets\n",
"root_dir = '../../datasets/caltech101'\n",
"filenames = sorted(get_file_list(root_dir))\n",
"print(len(filenames))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true,
"colab": {
"base_uri": "https://localhost:8080/",
"height": 84,
@@ -717,40 +401,15 @@
"145e7cf97a5844a48e321c8a4878d13b"
]
},
"collapsed": true,
"id": "8gfnWvUNnz6j",
"outputId": "9563b02d-607e-4b47-a823-d345875b146d"
},
"outputs": [],
"source": [
"standard_feature_list = []\n",
"for i in tqdm_notebook(range(len(filenames))):\n",
" standard_feature_list.append(extract_features(filenames[i], model))"
],
"execution_count": 12,
"outputs": [
{
"output_type": "display_data",
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "2f0c84e6e48b40cb9f1360ba4451513c",
"version_minor": 0,
"version_major": 2
},
"text/plain": [
"HBox(children=(IntProgress(value=0, max=8677), HTML(value=u'')))"
]
},
"metadata": {
"tags": []
}
},
{
"output_type": "stream",
"text": [
"\n",
"8677\n"
],
"name": "stdout"
}
]
},
{
@@ -764,64 +423,77 @@
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true,
"colab": {
"base_uri": "https://localhost:8080/"
},
"collapsed": true,
"id": "gsEQDJuKnz6k",
"outputId": "255b0843-9fd2-4947-94d5-fd400f729eba"
},
"source": [
"batch_size = 128\n",
"datagen = tensorflow.keras.preprocessing.image.ImageDataGenerator(preprocessing_function=preprocess_input)\n",
"\n",
"generator = datagen.flow_from_directory(root_dir,\n",
" target_size=(224, 224),\n",
" class_mode=None,\n",
" shuffle=False)\n",
"\n",
"num_images = len(generator.filenames)\n",
"num_epochs = int(math.ceil(num_images / batch_size))\n",
"\n",
"start_time = time.time()\n",
"feature_list = []\n",
"feature_list = model.predict_generator(generator, num_epochs)\n",
"end_time = time.time()"
],
"execution_count": 14,
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"WARNING: Logging before flag parsing goes to stderr.\n",
"W0725 00:00:57.965632 140262955202432 deprecation.py:323] From <ipython-input-14-4a2e3dbd89a6>:16: predict_generator (from tensorflow.python.keras.engine.training) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Please use Model.predict, which supports generators.\n"
],
"name": "stderr"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found 8677 images belonging to 101 classes.\n",
"(8677, 'num images')\n",
"(67, 'num epochs')\n"
],
"name": "stdout"
]
}
],
"source": [
"batch_size = 128\n",
"datagen = tensorflow.keras.preprocessing.image.ImageDataGenerator(preprocessing_function=preprocess_input)\n",
"\n",
"generator = datagen.flow_from_directory(root_dir,\n",
" target_size=(224, 224),\n",
" class_mode=None,\n",
" batch_size=batch_size,\n",
" shuffle=False)\n",
"\n",
"num_images = len(generator.filenames)\n",
"num_epochs = int(math.ceil(num_images / batch_size))\n",
"\n",
"start_time = time.time()\n",
"feature_list = []\n",
"feature_list = model.predict_generator(generator, num_epochs)\n",
"end_time = time.time()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true,
"colab": {
"base_uri": "https://localhost:8080/"
},
"collapsed": true,
"id": "oMPld5V0nz6k",
"outputId": "f552f9ef-123e-4f46-f11d-3bfa8ca96247"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"('Num images = ', 8677)\n",
"('Shape of feature_list = ', (2144, 2048))\n",
"('Time taken in sec = ', 8.890146017074585)\n"
]
}
],
"source": [
"for i, features in enumerate(feature_list):\n",
" feature_list[i] = features / norm(features)\n",
@@ -831,18 +503,6 @@
"print(\"Num images = \", len(generator.classes))\n",
"print(\"Shape of feature_list = \", feature_list.shape)\n",
"print(\"Time taken in sec = \", end_time - start_time)"
],
"execution_count": 18,
"outputs": [
{
"output_type": "stream",
"text": [
"('Num images = ', 8677)\n",
"('Shape of feature_list = ', (2144, 2048))\n",
"('Time taken in sec = ', 8.890146017074585)\n"
],
"name": "stdout"
}
]
},
{
@@ -921,265 +581,554 @@
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true,
"id": "JXxQ9oi7nz6m"
},
"outputs": [],
"source": [
"filenames = [root_dir + '/' + s for s in generator.filenames]"
],
"execution_count": 19,
"outputs": []
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"code_folding": [],
"collapsed": true,
"id": "6azYdG-Anz6m"
},
"outputs": [],
"source": [
"pickle.dump(generator.classes, open('./data/class_ids-caltech101.pickle',\n",
" 'wb'))\n",
"pickle.dump(filenames, open('./data/filenames-caltech101.pickle', 'wb'))\n",
"pickle.dump(\n",
" feature_list,\n",
" open('./data/features-caltech101-' + model_architecture + '.pickle', 'wb'))"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "ZBpWWLPNnz6m"
},
"source": [
"Let's train a finetuned model as well and save the features for that as well."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true,
"id": "tdq3wz3Qnz6m"
},
"outputs": [],
"source": [
"TRAIN_SAMPLES = 8677\n",
"NUM_CLASSES = 101\n",
"IMG_WIDTH, IMG_HEIGHT = 224, 224"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true,
"id": "ePCFVIHznz6n"
},
"outputs": [],
"source": [
"train_datagen = ImageDataGenerator(preprocessing_function=preprocess_input,\n",
" rotation_range=20,\n",
" width_shift_range=0.2,\n",
" height_shift_range=0.2,\n",
" zoom_range=0.2)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"collapsed": true,
"id": "Yh_NVL5Tnz6n",
"outputId": "129fa193-a5fa-4e02-9db0-f74525b5f9b2"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found 8677 images belonging to 101 classes.\n"
]
}
],
"source": [
"train_generator = train_datagen.flow_from_directory(root_dir,\n",
" target_size=(IMG_WIDTH,\n",
" IMG_HEIGHT),\n",
" shuffle=True,\n",
" seed=12345,\n",
" class_mode='categorical')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true,
"id": "bR6oZOPlnz6n"
},
"outputs": [],
"source": [
"def model_maker():\n",
" base_model = ResNet50(include_top=False,\n",
" input_shape=(IMG_WIDTH, IMG_HEIGHT, 3))\n",
" for layer in base_model.layers[:]:\n",
" layer.trainable = False\n",
" input = Input(shape=(IMG_WIDTH, IMG_HEIGHT, 3))\n",
" custom_model = base_model(input)\n",
" custom_model = GlobalAveragePooling2D()(custom_model)\n",
" custom_model = Dense(64, activation='relu')(custom_model)\n",
" custom_model = Dropout(0.5)(custom_model)\n",
" predictions = Dense(NUM_CLASSES, activation='softmax')(custom_model)\n",
" return Model(inputs=input, outputs=predictions)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"collapsed": true,
"id": "Sghlu4Rfnz6n",
"outputId": "19c97bb7-bc5b-4de0-e48c-60c088998823"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"W0725 00:01:11.765636 140262955202432 deprecation.py:323] From <ipython-input-25-3632d4ca8dfa>:8: fit_generator (from tensorflow.python.keras.engine.training) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Please use Model.fit, which supports generators.\n",
"W0725 00:01:12.087389 140262955202432 data_adapter.py:1091] sample_weight modes were coerced from\n",
" ...\n",
" to \n",
" ['...']\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Train for 68.0 steps\n",
"Epoch 1/10\n",
"68/68 [==============================] - 27s 391ms/step - loss: 3.8677 - acc: 0.2043\n",
"Epoch 2/10\n",
"68/68 [==============================] - 26s 379ms/step - loss: 2.8975 - acc: 0.3625\n",
"Epoch 3/10\n",
"68/68 [==============================] - 26s 382ms/step - loss: 2.5821 - acc: 0.4187\n",
"Epoch 4/10\n",
"68/68 [==============================] - 25s 363ms/step - loss: 2.3700 - acc: 0.4444\n",
"Epoch 5/10\n",
"68/68 [==============================] - 26s 389ms/step - loss: 2.0201 - acc: 0.5175\n",
"Epoch 6/10\n",
"68/68 [==============================] - 26s 386ms/step - loss: 1.8416 - acc: 0.5386\n",
"Epoch 7/10\n",
"68/68 [==============================] - 25s 371ms/step - loss: 1.7514 - acc: 0.5510\n",
"Epoch 8/10\n",
"68/68 [==============================] - 25s 370ms/step - loss: 1.6617 - acc: 0.5694\n",
"Epoch 9/10\n",
"68/68 [==============================] - 26s 383ms/step - loss: 1.5282 - acc: 0.5896\n",
"Epoch 10/10\n",
"68/68 [==============================] - 25s 366ms/step - loss: 1.5246 - acc: 0.5979\n"
]
},
{
"data": {
"text/plain": [
"<tensorflow.python.keras.callbacks.History at 0x7f8ebb7ec790>"
]
},
"execution_count": 25,
"metadata": {
"tags": []
},
"output_type": "execute_result"
}
],
"execution_count": 20,
"outputs": []
},
{
"cell_type": "markdown",
"metadata": {
"id": "ZBpWWLPNnz6m"
},
"source": [
"Let's train a finetuned model as well and save the features for that as well."
"model_finetuned = model_maker()\n",
"model_finetuned.compile(loss='categorical_crossentropy',\n",
" optimizer=tensorflow.keras.optimizers.Adam(0.001),\n",
" metrics=['acc'])\n",
"model_finetuned.fit_generator(\n",
" train_generator,\n",
" steps_per_epoch=math.ceil(float(TRAIN_SAMPLES) / batch_size),\n",
" epochs=10)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true,
"id": "tdq3wz3Qnz6m"
},
"source": [
"TRAIN_SAMPLES = 8677\n",
"NUM_CLASSES = 101\n",
"IMG_WIDTH, IMG_HEIGHT = 224, 224"
],
"execution_count": 21,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"collapsed": true,
"id": "ePCFVIHznz6n"
"id": "GcPDg7Wunz6o"
},
"outputs": [],
"source": [
"train_datagen = ImageDataGenerator(preprocessing_function=preprocess_input,\n",
" rotation_range=20,\n",
" width_shift_range=0.2,\n",
" height_shift_range=0.2,\n",
" zoom_range=0.2)"
],
"execution_count": 22,
"outputs": []
"model_finetuned.save('./data/model-finetuned.h5')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true,
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "Yh_NVL5Tnz6n",
"outputId": "129fa193-a5fa-4e02-9db0-f74525b5f9b2"
"collapsed": true,
"id": "cSE2qbwtnz6o",
"outputId": "dd898777-459f-44f2-b561-c2ed9792c4ef"
},
"source": [
"train_generator = train_datagen.flow_from_directory(root_dir,\n",
" target_size=(IMG_WIDTH,\n",
" IMG_HEIGHT),\n",
" shuffle=True,\n",
" seed=12345,\n",
" class_mode='categorical')"
],
"execution_count": 23,
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found 8677 images belonging to 101 classes.\n"
],
"name": "stdout"
"2144\n",
"('Num images = ', 2144)\n",
"('Shape of feature_list = ', (2144, 101))\n",
"('Time taken in sec = ', 8.284640073776245)\n"
]
}
],
"source": [
"start_time = time.time()\n",
"feature_list_finetuned = []\n",
"feature_list_finetuned = model_finetuned.predict_generator(generator, num_epochs)\n",
"end_time = time.time()\n",
"\n",
"for i, features_finetuned in enumerate(feature_list_finetuned):\n",
" feature_list_finetuned[i] = features_finetuned / norm(features_finetuned)\n",
"\n",
"feature_list = feature_list_finetuned.reshape(len(feature_list_finetuned), -1)\n",
"\n",
"print(\"Num images = \", len(feature_list_finetuned) )\n",
"print(\"Shape of feature_list = \", feature_list.shape)\n",
"print(\"Time taken in sec = \", end_time - start_time)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"collapsed": true,
"id": "bR6oZOPlnz6n"
"id": "L8FvRo-6nz6o"
},
"outputs": [],
"source": [
"def model_maker():\n",
" base_model = ResNet50(include_top=False,\n",
" input_shape=(IMG_WIDTH, IMG_HEIGHT, 3))\n",
" for layer in base_model.layers[:]:\n",
" layer.trainable = False\n",
" input = Input(shape=(IMG_WIDTH, IMG_HEIGHT, 3))\n",
" custom_model = base_model(input)\n",
" custom_model = GlobalAveragePooling2D()(custom_model)\n",
" custom_model = Dense(64, activation='relu')(custom_model)\n",
" custom_model = Dropout(0.5)(custom_model)\n",
" predictions = Dense(NUM_CLASSES, activation='softmax')(custom_model)\n",
" return Model(inputs=input, outputs=predictions)"
],
"execution_count": 24,
"outputs": []
"pickle.dump(\n",
" feature_list,\n",
" open('./data/features-caltech101-resnet-finetuned.pickle', 'wb'))"
]
}
],
"metadata": {
"accelerator": "GPU",
"colab": {
"collapsed_sections": [],
"name": "Copy of 1-feature-extraction.ipynb",
"provenance": []
},
{
"cell_type": "code",
"metadata": {
"collapsed": true,
"colab": {
"base_uri": "https://localhost:8080/"
"kernelspec": {
"display_name": "lewagon",
"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": "3.8.12"
},
"toc": {
"base_numbering": 1,
"nav_menu": {},
"number_sections": true,
"sideBar": true,
"skip_h1_title": false,
"title_cell": "Table of Contents",
"title_sidebar": "Contents",
"toc_cell": false,
"toc_position": {},
"toc_section_display": true,
"toc_window_display": false
},
"widgets": {
"application/vnd.jupyter.widget-state+json": {
"145e7cf97a5844a48e321c8a4878d13b": {
"model_module": "@jupyter-widgets/base",
"model_name": "LayoutModel",
"state": {
"_model_module": "@jupyter-widgets/base",
"_model_module_version": "1.2.0",
"_model_name": "LayoutModel",
"_view_count": null,
"_view_module": "@jupyter-widgets/base",
"_view_module_version": "1.2.0",
"_view_name": "LayoutView",
"align_content": null,
"align_items": null,
"align_self": null,
"border": null,
"bottom": null,
"display": null,
"flex": null,
"flex_flow": null,
"grid_area": null,
"grid_auto_columns": null,
"grid_auto_flow": null,
"grid_auto_rows": null,
"grid_column": null,
"grid_gap": null,
"grid_row": null,
"grid_template_areas": null,
"grid_template_columns": null,
"grid_template_rows": null,
"height": null,
"justify_content": null,
"justify_items": null,
"left": null,
"margin": null,
"max_height": null,
"max_width": null,
"min_height": null,
"min_width": null,
"object_fit": null,
"object_position": null,
"order": null,
"overflow": null,
"overflow_x": null,
"overflow_y": null,
"padding": null,
"right": null,
"top": null,
"visibility": null,
"width": null
}
},
"2f0c84e6e48b40cb9f1360ba4451513c": {
"model_module": "@jupyter-widgets/controls",
"model_name": "HBoxModel",
"state": {
"_dom_classes": [],
"_model_module": "@jupyter-widgets/controls",
"_model_module_version": "1.5.0",
"_model_name": "HBoxModel",
"_view_count": null,
"_view_module": "@jupyter-widgets/controls",
"_view_module_version": "1.5.0",
"_view_name": "HBoxView",
"box_style": "",
"children": [
"IPY_MODEL_a0ffda7bcf1a46a4bdd6352c43cfda61",
"IPY_MODEL_f80ec0bf78de458995da4f9c964f3722"
],
"layout": "IPY_MODEL_470d8a8b00c5475684ab84b55c79b760"
}
},
"470d8a8b00c5475684ab84b55c79b760": {
"model_module": "@jupyter-widgets/base",
"model_name": "LayoutModel",
"state": {
"_model_module": "@jupyter-widgets/base",
"_model_module_version": "1.2.0",
"_model_name": "LayoutModel",
"_view_count": null,
"_view_module": "@jupyter-widgets/base",
"_view_module_version": "1.2.0",
"_view_name": "LayoutView",
"align_content": null,
"align_items": null,
"align_self": null,
"border": null,
"bottom": null,
"display": null,
"flex": null,
"flex_flow": null,
"grid_area": null,
"grid_auto_columns": null,
"grid_auto_flow": null,
"grid_auto_rows": null,
"grid_column": null,
"grid_gap": null,
"grid_row": null,
"grid_template_areas": null,
"grid_template_columns": null,
"grid_template_rows": null,
"height": null,
"justify_content": null,
"justify_items": null,
"left": null,
"margin": null,
"max_height": null,
"max_width": null,
"min_height": null,
"min_width": null,
"object_fit": null,
"object_position": null,
"order": null,
"overflow": null,
"overflow_x": null,
"overflow_y": null,
"padding": null,
"right": null,
"top": null,
"visibility": null,
"width": null
}
},
"id": "Sghlu4Rfnz6n",
"outputId": "19c97bb7-bc5b-4de0-e48c-60c088998823"
},
"source": [
"model_finetuned = model_maker()\n",
"model_finetuned.compile(loss='categorical_crossentropy',\n",
" optimizer=tensorflow.keras.optimizers.Adam(0.001),\n",
" metrics=['acc'])\n",
"model_finetuned.fit_generator(\n",
" train_generator,\n",
" steps_per_epoch=math.ceil(float(TRAIN_SAMPLES) / batch_size),\n",
" epochs=10)"
],
"execution_count": 25,
"outputs": [
{
"output_type": "stream",
"text": [
"W0725 00:01:11.765636 140262955202432 deprecation.py:323] From <ipython-input-25-3632d4ca8dfa>:8: fit_generator (from tensorflow.python.keras.engine.training) is deprecated and will be removed in a future version.\n",
"Instructions for updating:\n",
"Please use Model.fit, which supports generators.\n",
"W0725 00:01:12.087389 140262955202432 data_adapter.py:1091] sample_weight modes were coerced from\n",
" ...\n",
" to \n",
" ['...']\n"
],
"name": "stderr"
"4d80db8732434d119c3f3b3bb8469ce9": {
"model_module": "@jupyter-widgets/base",
"model_name": "LayoutModel",
"state": {
"_model_module": "@jupyter-widgets/base",
"_model_module_version": "1.2.0",
"_model_name": "LayoutModel",
"_view_count": null,
"_view_module": "@jupyter-widgets/base",
"_view_module_version": "1.2.0",
"_view_name": "LayoutView",
"align_content": null,
"align_items": null,
"align_self": null,
"border": null,
"bottom": null,
"display": null,
"flex": null,
"flex_flow": null,
"grid_area": null,
"grid_auto_columns": null,
"grid_auto_flow": null,
"grid_auto_rows": null,
"grid_column": null,
"grid_gap": null,
"grid_row": null,
"grid_template_areas": null,
"grid_template_columns": null,
"grid_template_rows": null,
"height": null,
"justify_content": null,
"justify_items": null,
"left": null,
"margin": null,
"max_height": null,
"max_width": null,
"min_height": null,
"min_width": null,
"object_fit": null,
"object_position": null,
"order": null,
"overflow": null,
"overflow_x": null,
"overflow_y": null,
"padding": null,
"right": null,
"top": null,
"visibility": null,
"width": null
}
},
{
"output_type": "stream",
"text": [
"Train for 68.0 steps\n",
"Epoch 1/10\n",
"68/68 [==============================] - 27s 391ms/step - loss: 3.8677 - acc: 0.2043\n",
"Epoch 2/10\n",
"68/68 [==============================] - 26s 379ms/step - loss: 2.8975 - acc: 0.3625\n",
"Epoch 3/10\n",
"68/68 [==============================] - 26s 382ms/step - loss: 2.5821 - acc: 0.4187\n",
"Epoch 4/10\n",
"68/68 [==============================] - 25s 363ms/step - loss: 2.3700 - acc: 0.4444\n",
"Epoch 5/10\n",
"68/68 [==============================] - 26s 389ms/step - loss: 2.0201 - acc: 0.5175\n",
"Epoch 6/10\n",
"68/68 [==============================] - 26s 386ms/step - loss: 1.8416 - acc: 0.5386\n",
"Epoch 7/10\n",
"68/68 [==============================] - 25s 371ms/step - loss: 1.7514 - acc: 0.5510\n",
"Epoch 8/10\n",
"68/68 [==============================] - 25s 370ms/step - loss: 1.6617 - acc: 0.5694\n",
"Epoch 9/10\n",
"68/68 [==============================] - 26s 383ms/step - loss: 1.5282 - acc: 0.5896\n",
"Epoch 10/10\n",
"68/68 [==============================] - 25s 366ms/step - loss: 1.5246 - acc: 0.5979\n"
],
"name": "stdout"
"9792cd696fa74cc68ef6b35095fb8685": {
"model_module": "@jupyter-widgets/controls",
"model_name": "ProgressStyleModel",
"state": {
"_model_module": "@jupyter-widgets/controls",
"_model_module_version": "1.5.0",
"_model_name": "ProgressStyleModel",
"_view_count": null,
"_view_module": "@jupyter-widgets/base",
"_view_module_version": "1.2.0",
"_view_name": "StyleView",
"bar_color": null,
"description_width": ""
}
},
{
"output_type": "execute_result",
"data": {
"text/plain": [
"<tensorflow.python.keras.callbacks.History at 0x7f8ebb7ec790>"
]
},
"metadata": {
"tags": []
},
"execution_count": 25
}
]
},
{
"cell_type": "code",
"metadata": {
"collapsed": true,
"id": "GcPDg7Wunz6o"
},
"source": [
"model_finetuned.save('./data/model-finetuned.h5')"
],
"execution_count": 26,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"collapsed": true,
"colab": {
"base_uri": "https://localhost:8080/"
"a0ffda7bcf1a46a4bdd6352c43cfda61": {
"model_module": "@jupyter-widgets/controls",
"model_name": "IntProgressModel",
"state": {
"_dom_classes": [],
"_model_module": "@jupyter-widgets/controls",
"_model_module_version": "1.5.0",
"_model_name": "IntProgressModel",
"_view_count": null,
"_view_module": "@jupyter-widgets/controls",
"_view_module_version": "1.5.0",
"_view_name": "ProgressView",
"bar_style": "success",
"description": "",
"description_tooltip": null,
"layout": "IPY_MODEL_4d80db8732434d119c3f3b3bb8469ce9",
"max": 8677,
"min": 0,
"orientation": "horizontal",
"style": "IPY_MODEL_9792cd696fa74cc68ef6b35095fb8685",
"value": 8677
}
},
"id": "cSE2qbwtnz6o",
"outputId": "dd898777-459f-44f2-b561-c2ed9792c4ef"
},
"source": [
"start_time = time.time()\n",
"feature_list_finetuned = []\n",
"feature_list_finetuned = model_finetuned.predict_generator(generator, num_epochs)\n",
"end_time = time.time()\n",
"\n",
"for i, features_finetuned in enumerate(feature_list_finetuned):\n",
" feature_list_finetuned[i] = features_finetuned / norm(features_finetuned)\n",
"\n",
"feature_list = feature_list_finetuned.reshape(len(feature_list_finetuned), -1)\n",
"\n",
"print(\"Num images = \", len(feature_list_finetuned) )\n",
"print(\"Shape of feature_list = \", feature_list.shape)\n",
"print(\"Time taken in sec = \", end_time - start_time)"
],
"execution_count": 27,
"outputs": [
{
"output_type": "stream",
"text": [
"2144\n",
"('Num images = ', 2144)\n",
"('Shape of feature_list = ', (2144, 101))\n",
"('Time taken in sec = ', 8.284640073776245)\n"
],
"name": "stdout"
"f80ec0bf78de458995da4f9c964f3722": {
"model_module": "@jupyter-widgets/controls",
"model_name": "HTMLModel",
"state": {
"_dom_classes": [],
"_model_module": "@jupyter-widgets/controls",
"_model_module_version": "1.5.0",
"_model_name": "HTMLModel",
"_view_count": null,
"_view_module": "@jupyter-widgets/controls",
"_view_module_version": "1.5.0",
"_view_name": "HTMLView",
"description": "",
"description_tooltip": null,
"layout": "IPY_MODEL_145e7cf97a5844a48e321c8a4878d13b",
"placeholder": "​",
"style": "IPY_MODEL_fe1544db2c8e4df2a5c7097575da68ab",
"value": "100% 8677/8677 [06:16&lt;00:00, 23.04it/s]"
}
},
"fe1544db2c8e4df2a5c7097575da68ab": {
"model_module": "@jupyter-widgets/controls",
"model_name": "DescriptionStyleModel",
"state": {
"_model_module": "@jupyter-widgets/controls",
"_model_module_version": "1.5.0",
"_model_name": "DescriptionStyleModel",
"_view_count": null,
"_view_module": "@jupyter-widgets/base",
"_view_module_version": "1.2.0",
"_view_name": "StyleView",
"description_width": ""
}
}
]
},
{
"cell_type": "code",
"metadata": {
"collapsed": true,
"id": "L8FvRo-6nz6o"
},
"source": [
"pickle.dump(\n",
" feature_list,\n",
" open('./data/features-caltech101-resnet-finetuned.pickle', 'wb'))"
],
"execution_count": 28,
"outputs": []
}
}
]
},
"nbformat": 4,
"nbformat_minor": 0
}