여러개의 TFRecord 사용 법.
https://github.com/tensorflow/models/issues/3031
This issue is closed since I found answers in the code
The object detection API use parallel reader to import your dataset, here are the comments by the developer
Usage:
data_sources = ['path_to/train*']
key, value = parallel_read(data_sources, tf.CSVReader, num_readers=4)
Args:
data_sources: a list/tuple of files or the location of the data, i.e.
/path/to/train@128, /path/to/train* or /tmp/.../train*
So basically you can define a list of input path as @byungjae89 mentioned above, or simply provide the input directory like
input_path: my_dataset/train/*
The reader will read the entire folder for you.
http://songheqi.me/2017/05/23/Custom-Image-Classify-with-Tensorflow/
TFRecord
14
|
Tensorflow's inception model has a file build_image_data.py that can accomplish the same thing with the assumption that each subdirectory represents a label.
|
object_dectection 폴더의 부모 폴더에서 실행해야 함.
if you are running object_detection.
go to 'object_detection' dir and run your code! 'util' dir is within object_detection |
/work/tensorflow/models/research/data$ python data/generate_tfrecord.py --csv_input=data/data/train_labels.csv --output_path=data/data/train.record
chulminson@chulminson-Z71RH-AD7501E:/work/tensorflow/models/research/data$ ls -laR
.:
total 24
drwxrwxr-x 4 chulminson chulminson 4096 2월 7 15:39 .
drwxrwxr-x 44 chulminson chulminson 4096 2월 7 09:59 ..
drwxrwxr-x 2 chulminson chulminson 4096 2월 7 15:39 data
-rw-rw-r-- 1 chulminson chulminson 3381 2월 7 15:39 generate_tfrecord.py
drwxrwxr-x 2 chulminson chulminson 4096 2월 7 15:39 images
-rw-rw-r-- 1 chulminson chulminson 1171 2월 7 08:36 xml_to_csv.py
./data:
total 216
drwxrwxr-x 2 chulminson chulminson 4096 2월 7 15:39 .
drwxrwxr-x 4 chulminson chulminson 4096 2월 7 15:39 ..
-rw-rw-r-- 1 chulminson chulminson 186 2월 7 15:24 test_labels.csv
-rw-rw-r-- 1 chulminson chulminson 24441 2월 7 15:39 test.record
-rw-rw-r-- 1 chulminson chulminson 1263 2월 7 15:24 train_labels.csv
-rw-rw-r-- 1 chulminson chulminson 178030 2월 7 15:39 train.record
./images:
total 444
drwxrwxr-x 2 chulminson chulminson 4096 2월 7 15:39 .
drwxrwxr-x 4 chulminson chulminson 4096 2월 7 15:39 ..
-rw-rw-r-- 1 chulminson chulminson 1032 2월 6 10:27 a10.png
-rw-rw-r-- 1 chulminson chulminson 509 2월 6 16:48 a10.xml
-rw-rw-r-- 1 chulminson chulminson 2472 2월 6 10:27 a11.png
-rw-rw-r-- 1 chulminson chulminson 508 2월 6 16:48 a11.xml
-rw-rw-r-- 1 chulminson chulminson 3194 2월 6 10:27 a12.png
-rw-rw-r-- 1 chulminson chulminson 510 2월 6 16:48 a12.xml
-rw-rw-r-- 1 chulminson chulminson 625 2월 6 10:28 a13.png
-rw-rw-r-- 1 chulminson chulminson 510 2월 6 16:49 a13.xml
-rw-rw-r-- 1 chulminson chulminson 1571 2월 6 10:28 a14.png
-rw-rw-r-- 1 chulminson chulminson 510 2월 6 16:49 a14.xml
-rw-rw-r-- 1 chulminson chulminson 5767 2월 6 10:27 a15.jpg
-rw-rw-r-- 1 chulminson chulminson 508 2월 6 16:49 a15.xml
-rw-rw-r-- 1 chulminson chulminson 1711 2월 6 10:27 a16.png
-rw-rw-r-- 1 chulminson chulminson 510 2월 6 16:49 a16.xml
-rw-rw-r-- 1 chulminson chulminson 9582 2월 6 10:27 a17.jpg
-rw-rw-r-- 1 chulminson chulminson 508 2월 6 16:50 a17.xml
-rw-rw-r-- 1 chulminson chulminson 11717 2월 6 10:27 a18.jpg
-rw-rw-r-- 1 chulminson chulminson 508 2월 6 16:50 a18.xml
-rw-rw-r-- 1 chulminson chulminson 4631 2월 6 10:27 a19.jpg
-rw-rw-r-- 1 chulminson chulminson 508 2월 6 16:50 a19.xml
-rw-rw-r-- 1 chulminson chulminson 6609 2월 6 10:27 a1.jpg
-rw-rw-r-- 1 chulminson chulminson 506 2월 6 16:47 a1.xml
-rw-rw-r-- 1 chulminson chulminson 10786 2월 6 16:46 a20.jpg
-rw-rw-r-- 1 chulminson chulminson 508 2월 6 16:51 a20.xml
-rw-rw-r-- 1 chulminson chulminson 2467 2월 6 10:27 a2.png
-rw-rw-r-- 1 chulminson chulminson 506 2월 6 16:50 a2.xml
-rw-rw-r-- 1 chulminson chulminson 9816 2월 6 10:28 a3.jpg
-rw-rw-r-- 1 chulminson chulminson 506 2월 6 16:51 a3.xml
-rw-rw-r-- 1 chulminson chulminson 1434 2월 6 10:27 a4.png
-rw-rw-r-- 1 chulminson chulminson 507 2월 6 16:51 a4.xml
-rw-rw-r-- 1 chulminson chulminson 5883 2월 6 10:28 a5.jpg
-rw-rw-r-- 1 chulminson chulminson 506 2월 6 16:51 a5.xml
-rw-rw-r-- 1 chulminson chulminson 2616 2월 6 10:27 a6.png
-rw-rw-r-- 1 chulminson chulminson 506 2월 6 16:52 a6.xml
-rw-rw-r-- 1 chulminson chulminson 5599 2월 6 10:28 a7.jpg
-rw-rw-r-- 1 chulminson chulminson 508 2월 6 16:52 a7.xml
-rw-rw-r-- 1 chulminson chulminson 1190 2월 6 10:27 a8.png
-rw-rw-r-- 1 chulminson chulminson 507 2월 6 16:52 a8.xml
-rw-rw-r-- 1 chulminson chulminson 3775 2월 6 10:28 a9.jpg
-rw-rw-r-- 1 chulminson chulminson 508 2월 6 16:53 a9.xml
-rw-rw-r-- 1 chulminson chulminson 4010 2월 6 10:29 p10.png
-rw-rw-r-- 1 chulminson chulminson 505 2월 6 16:54 p10.xml
-rw-rw-r-- 1 chulminson chulminson 6239 2월 6 10:29 p11.jpg
-rw-rw-r-- 1 chulminson chulminson 505 2월 6 16:54 p11.xml
-rw-rw-r-- 1 chulminson chulminson 2538 2월 6 10:29 p12.png
-rw-rw-r-- 1 chulminson chulminson 506 2월 6 16:54 p12.xml
-rw-rw-r-- 1 chulminson chulminson 4394 2월 6 10:29 p13.jpg
-rw-rw-r-- 1 chulminson chulminson 506 2월 6 16:54 p13.xml
-rw-rw-r-- 1 chulminson chulminson 941 2월 6 10:29 p14.png
-rw-rw-r-- 1 chulminson chulminson 505 2월 6 16:55 p14.xml
-rw-rw-r-- 1 chulminson chulminson 4616 2월 6 10:30 p15.jpg
-rw-rw-r-- 1 chulminson chulminson 507 2월 6 16:55 p15.xml
-rw-rw-r-- 1 chulminson chulminson 1431 2월 6 10:31 p16.png
-rw-rw-r-- 1 chulminson chulminson 505 2월 6 16:55 p16.xml
-rw-rw-r-- 1 chulminson chulminson 10310 2월 6 10:30 p17.jpg
-rw-rw-r-- 1 chulminson chulminson 507 2월 6 16:55 p17.xml
-rw-rw-r-- 1 chulminson chulminson 4910 2월 6 10:31 p18.jpg
-rw-rw-r-- 1 chulminson chulminson 505 2월 6 16:56 p18.xml
-rw-rw-r-- 1 chulminson chulminson 3717 2월 6 10:31 p19.jpg
-rw-rw-r-- 1 chulminson chulminson 507 2월 6 16:56 p19.xml
-rw-rw-r-- 1 chulminson chulminson 6228 2월 6 10:25 p1.jpg
-rw-rw-r-- 1 chulminson chulminson 503 2월 6 16:54 p1.xml
-rw-rw-r-- 1 chulminson chulminson 3717 2월 6 10:31 p20.jpg
-rw-rw-r-- 1 chulminson chulminson 507 2월 6 16:56 p20.xml
-rw-rw-r-- 1 chulminson chulminson 6130 2월 6 10:25 p2.png
-rw-rw-r-- 1 chulminson chulminson 503 2월 6 16:56 p2.xml
-rw-rw-r-- 1 chulminson chulminson 5332 2월 6 10:25 p3.jpg
-rw-rw-r-- 1 chulminson chulminson 505 2월 6 16:56 p3.xml
-rw-rw-r-- 1 chulminson chulminson 6130 2월 6 10:25 p4.png
-rw-rw-r-- 1 chulminson chulminson 503 2월 6 16:57 p4.xml
-rw-rw-r-- 1 chulminson chulminson 4744 2월 6 10:26 p5.jpg
-rw-rw-r-- 1 chulminson chulminson 503 2월 6 16:57 p5.xml
-rw-rw-r-- 1 chulminson chulminson 687 2월 6 10:26 p6.png
-rw-rw-r-- 1 chulminson chulminson 505 2월 6 16:57 p6.xml
-rw-rw-r-- 1 chulminson chulminson 4999 2월 6 10:26 p7.jpg
-rw-rw-r-- 1 chulminson chulminson 505 2월 6 16:57 p7.xml
-rw-rw-r-- 1 chulminson chulminson 4633 2월 6 10:26 p8.png
-rw-rw-r-- 1 chulminson chulminson 503 2월 6 16:58 p8.xml
-rw-rw-r-- 1 chulminson chulminson 8425 2월 6 10:26 p9.jpg
-rw-rw-r-- 1 chulminson chulminson 503 2월 6 16:58 p9.xml
python object_detection/export_inference_graph.py \
--input_type image_tensor \
--pipeline_config_path object_detection/training/pipeline.config \
--trained_checkpoint_prefix object_detection/training/model.ckpt-1456 \
--output_directory play_pause_graph
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Object Detection Demo\n",
"Welcome to the object detection inference walkthrough! This notebook will walk you step by step through the process of using a pre-trained model to detect objects in an image. Make sure to follow the [installation instructions](https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/installation.md) before you start."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Imports"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import numpy as np\n",
"import os\n",
"import six.moves.urllib as urllib\n",
"import sys\n",
"import tarfile\n",
"import tensorflow as tf\n",
"import zipfile\n",
"\n",
"from collections import defaultdict\n",
"from io import StringIO\n",
"from matplotlib import pyplot as plt\n",
"from PIL import Image\n",
"\n",
"if tf.__version__ < '1.4.0':\n",
" raise ImportError('Please upgrade your tensorflow installation to v1.4.* or later!')\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Env setup"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"# This is needed to display the images.\n",
"%matplotlib inline\n",
"\n",
"# This is needed since the notebook is stored in the object_detection folder.\n",
"sys.path.append(\"..\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Object detection imports\n",
"Here are the imports from the object detection module."
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"from utils import label_map_util\n",
"\n",
"from utils import visualization_utils as vis_util"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Model preparation "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Variables\n",
"\n",
"Any model exported using the `export_inference_graph.py` tool can be loaded here simply by changing `PATH_TO_CKPT` to point to a new .pb file. \n",
"\n",
"By default we use an \"SSD with Mobilenet\" model here. See the [detection model zoo](https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md) for a list of other models that can be run out-of-the-box with varying speeds and accuracies."
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"# What model to download.\n",
"MODEL_NAME = 'play_pause_graph'\n",
"\n",
"# Path to frozen detection graph. This is the actual model that is used for the object detection.\n",
"PATH_TO_CKPT = MODEL_NAME + '/frozen_inference_graph.pb'\n",
"\n",
"# List of the strings that is used to add correct label for each box.\n",
"PATH_TO_LABELS = os.path.join('training', 'play_pause_dectection.pbtxt')\n",
"\n",
"NUM_CLASSES = 2"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Download Model"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Load a (frozen) Tensorflow model into memory."
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
"detection_graph = tf.Graph()\n",
"with detection_graph.as_default():\n",
" od_graph_def = tf.GraphDef()\n",
" with tf.gfile.GFile(PATH_TO_CKPT, 'rb') as fid:\n",
" serialized_graph = fid.read()\n",
" od_graph_def.ParseFromString(serialized_graph)\n",
" tf.import_graph_def(od_graph_def, name='')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Loading label map\n",
"Label maps map indices to category names, so that when our convolution network predicts `5`, we know that this corresponds to `airplane`. Here we use internal utility functions, but anything that returns a dictionary mapping integers to appropriate string labels would be fine"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"label_map = label_map_util.load_labelmap(PATH_TO_LABELS)\n",
"categories = label_map_util.convert_label_map_to_categories(label_map, max_num_classes=NUM_CLASSES, use_display_name=True)\n",
"category_index = label_map_util.create_category_index(categories)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Helper code"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"def load_image_into_numpy_array(image):\n",
" (im_width, im_height) = image.size\n",
" return np.array(image.getdata()).reshape(\n",
" (im_height, im_width, 3)).astype(np.uint8)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Detection"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [],
"source": [
"# For the sake of simplicity we will use only 2 images:\n",
"# image1.jpg\n",
"# image2.jpg\n",
"# If you want to test the code with your images, just add path to the images to the TEST_IMAGE_PATHS.\n",
"PATH_TO_TEST_IMAGES_DIR = 'test_images'\n",
"TEST_IMAGE_PATHS = [ os.path.join(PATH_TO_TEST_IMAGES_DIR, 'image{}.jpg'.format(i)) for i in range(7, 7) ]\n",
"\n",
"# Size, in inches, of the output images.\n",
"IMAGE_SIZE = (12, 8)"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [],
"source": [
"with detection_graph.as_default():\n",
" with tf.Session(graph=detection_graph) as sess:\n",
" # Definite input and output Tensors for detection_graph\n",
" image_tensor = detection_graph.get_tensor_by_name('image_tensor:0')\n",
" # Each box represents a part of the image where a particular object was detected.\n",
" detection_boxes = detection_graph.get_tensor_by_name('detection_boxes:0')\n",
" # Each score represent how level of confidence for each of the objects.\n",
" # Score is shown on the result image, together with the class label.\n",
" detection_scores = detection_graph.get_tensor_by_name('detection_scores:0')\n",
" detection_classes = detection_graph.get_tensor_by_name('detection_classes:0')\n",
" num_detections = detection_graph.get_tensor_by_name('num_detections:0')\n",
" for image_path in TEST_IMAGE_PATHS:\n",
" image = Image.open(image_path)\n",
" # the array based representation of the image will be used later in order to prepare the\n",
" # result image with boxes and labels on it.\n",
" image_np = load_image_into_numpy_array(image)\n",
" # Expand dimensions since the model expects images to have shape: [1, None, None, 3]\n",
" image_np_expanded = np.expand_dims(image_np, axis=0)\n",
" # Actual detection.\n",
" (boxes, scores, classes, num) = sess.run(\n",
" [detection_boxes, detection_scores, detection_classes, num_detections],\n",
" feed_dict={image_tensor: image_np_expanded})\n",
" # Visualization of the results of a detection.\n",
" vis_util.visualize_boxes_and_labels_on_image_array(\n",
" image_np,\n",
" np.squeeze(boxes),\n",
" np.squeeze(classes).astype(np.int32),\n",
" np.squeeze(scores),\n",
" category_index,\n",
" use_normalized_coordinates=True,\n",
" line_thickness=8)\n",
" plt.figure(figsize=IMAGE_SIZE)\n",
" plt.imshow(image_np)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"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": "2.7.6"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
Adesh commented on Jul 11, 2017