-
Notifications
You must be signed in to change notification settings - Fork 0
/
Setup_COCO_dataset.py
253 lines (189 loc) · 8.8 KB
/
Setup_COCO_dataset.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
import os
import torch
import random
import time
import copy
import numpy as np
import shutil
import wget
from zipfile import ZipFile
import pandas as pd
import cv2
from sklearn.model_selection import train_test_split
import xml.etree.ElementTree as ET
from xml.dom import minidom
from tqdm import tqdm
from PIL import Image, ImageDraw
import matplotlib.pyplot as plt
# ======================================================================
# == Check GPU is connected
# ======================================================================
print("======================")
print("Check GPU is info")
print("======================")
print("How many GPUs are there? Answer:",torch.cuda.device_count())
print("The Current GPU:",torch.cuda.current_device())
print("The Name Of The Current GPU",torch.cuda.get_device_name(torch.cuda.current_device()))
# Is PyTorch using a GPU?
print("Is Pytorch using GPU? Answer:",torch.cuda.is_available())
print("======================")
# switch to False to use CPU
use_cuda = True
use_cuda = use_cuda and torch.cuda.is_available()
device = torch.device("cuda" if use_cuda else "cpu");
# =====================================================
# == Set random seeds
# =====================================================
def set_random_seeds(random_seed=0):
torch.manual_seed(random_seed)
torch.backends.cudnn.deterministic = True
torch.backends.cudnn.benchmark = False
np.random.seed(random_seed)
random.seed(random_seed)
def plot_bounding_box(image, annotation_list):
annotations = np.array(annotation_list)
w, h = image.size
plotted_image = ImageDraw.Draw(image)
transformed_annotations = np.copy(annotations)
transformed_annotations[:,[1,3]] = annotations[:,[1,3]] * w
transformed_annotations[:,[2,4]] = annotations[:,[2,4]] * h
transformed_annotations[:,1] = transformed_annotations[:,1] - (transformed_annotations[:,3] / 2)
transformed_annotations[:,2] = transformed_annotations[:,2] - (transformed_annotations[:,4] / 2)
transformed_annotations[:,3] = transformed_annotations[:,1] + transformed_annotations[:,3]
transformed_annotations[:,4] = transformed_annotations[:,2] + transformed_annotations[:,4]
for ann in transformed_annotations:
obj_cls, x0, y0, x1, y1 = ann
plotted_image.rectangle(((x0,y0), (x1,y1)))
plotted_image.text((x0, y0 - 10), class_id_to_name_mapping[(int(obj_cls))])
plt.imshow(np.array(image))
plt.savefig("temp.png")
#Utility function to move images
def move_files_to_folder(list_of_files, destination_folder):
for f in list_of_files:
try:
shutil.move(f, destination_folder)
except:
print(f)
assert False
# =====================================================
# == Get YOLO model
# =====================================================
random_seed = 0
num_classes = 200
cuda_device = torch.device("cuda:0")
cpu_device = torch.device("cpu:0")
set_random_seeds(random_seed=random_seed)
def get_and_prepare_coco_dataset():
path = "datasets/COCO/"
if not os.path.isdir(path):
os.makedirs(path)
# Download images files
url = "http://images.cocodataset.org/zips/"
fs= ['train2017.zip','val2017.zip']#, 'test2017.zip']
for f in fs:
if not os.path.exists(path+f):
wget.download(url+f,out = path)
print("Downloaded COCO " + f + " zip file.")
else:
print("COCO " + f + " already exist")
# Download COCO dataset annotations in json format
# url = "http://images.cocodataset.org/annotations/"
# fs= ["annotations_trainval2017.zip", "image_info_test2017.zip"]
# for f in fs:
# if not os.path.exists(path+f):
# wget.download(url+f,out = path)
# print("Downloaded COCO " + f + " zip file.")
# else:
# print("COCO " + f + " already exist")
# Download COCO dataset annotations in yolo format
url= "https://github.com/ultralytics/yolov5/releases/download/v1.0/"
f= 'coco2017labels.zip'
if not os.path.exists(path+f):
wget.download(url+f,out = path)
print("Downloaded COCO annotations in yolo format zip file.")
else:
print("COCO labels already exist")
# Extract files
fs= ['train2017.zip', 'val2017.zip', 'coco2017labels.zip']
for f in fs:
if not os.path.exists(path+f[:-4]):
with ZipFile(path+f, "r") as file:
print("Extracting COCO " + f + " zip file, please wait ...")
file.extractall(path)
print("Extracted COCO " + f + " zip file.")
else:
print("COCO " + f + " already extracted")
# Make images and labels directories
if not os.path.exists(path+"images"):
os.mkdir(path+"images")
if not os.path.exists(path+"labels"):
os.mkdir(path+"labels")
temp_paths = ["images/train", "images/val", "images/test", "labels/train", "labels/val", "labels/test"]
for temp_path in temp_paths:
temp_path = path + temp_path
if not os.path.isdir(temp_path):
os.makedirs(temp_path)
# Read images and annotations. Note did work around here to ignore images that do not have an annotation file.
images = [os.path.join(path+'train2017', x[:-3]+"jpg") for x in os.listdir(path+'coco/labels/train2017')]
annotations = [os.path.join(path+'coco/labels/train2017', x) for x in os.listdir(path+'coco/labels/train2017') if x[-3:] == "txt"]
val_images = [os.path.join(path+'val2017', x[:-3]+"jpg") for x in os.listdir(path+'coco/labels/val2017')]
val_annotations = [os.path.join(path+'coco/labels/val2017', x) for x in os.listdir(path+'coco/labels/val2017') if x[-3:] == "txt"]
images.sort()
annotations.sort()
val_images.sort()
val_annotations.sort()
# Split the dataset into train-valid-test splits
train_images, test_images, train_annotations, test_annotations = train_test_split(images, annotations, test_size = 0.2, random_state = 1)
# Move the splits into their folders
move_files_to_folder(train_images, path+'images/train')
move_files_to_folder(val_images, path+'images/val/')
move_files_to_folder(test_images, path+'images/test/')
move_files_to_folder(train_annotations, path+'labels/train/')
move_files_to_folder(val_annotations, path+'labels/val/')
move_files_to_folder(test_annotations, path+'labels/test/')
# Remove files
fs = ["coco", "test2017", "train2017", "val2017"]
for f in fs:
shutil.rmtree(path + f)
# =====================================================
# == Main
# =====================================================
get_and_prepare_coco_dataset()
# # Read images and annotations
# images = [os.path.join('datasets/COCO/images', x) for x in os.listdir('datasets/road_signs/images')]
# annotations = [os.path.join('datasets/COCO/annotations', x) for x in os.listdir('datasets/road_signs/annotations') if x[-3:] == "txt"]
# images.sort()
# annotations.sort()
# # Get any random annotation file
# annotation_file = random.choice(annotations)
# with open(annotation_file, "r") as file:
# annotation_list = file.read().split("\n")[:-1]
# annotation_list = [x.split(" ") for x in annotation_list]
# annotation_list = [[float(y) for y in x ] for x in annotation_list]
# #Get the corresponding image file
# image_file = annotation_file.replace("annotations", "images").replace("txt", "png")
# assert os.path.exists(image_file)
# #Load the image
# image = Image.open(image_file)
# #Plot the Bounding Box
# plot_bounding_box(image, annotation_list)
# # Read images and annotations
# images = [os.path.join('datasets/road_signs/images', x) for x in os.listdir('datasets/road_signs/images')]
# annotations = [os.path.join('datasets/road_signs/annotations', x) for x in os.listdir('datasets/road_signs/annotations') if x[-3:] == "txt"]
# images.sort()
# annotations.sort()
# # Split the dataset into train-valid-test splits
# train_images, val_images, train_annotations, val_annotations = train_test_split(images, annotations, test_size = 0.2, random_state = 1)
# val_images, test_images, val_annotations, test_annotations = train_test_split(val_images, val_annotations, test_size = 0.5, random_state = 1)
# temp_paths = ["images/train", "images/val", "images/test", "labels/train", "labels/val", "labels/test"]
# for temp_path in temp_paths:
# temp_path = "datasets/road_signs/"+ temp_path
# if not os.path.isdir(temp_path):
# os.makedirs(temp_path)
# # Move the splits into their folders
# move_files_to_folder(train_images, 'datasets/road_signs/images/train')
# move_files_to_folder(val_images, 'datasets/road_signs/images/val/')
# move_files_to_folder(test_images, 'datasets/road_signs/images/test/')
# move_files_to_folder(train_annotations, 'datasets/road_signs/labels/train/')
# move_files_to_folder(val_annotations, 'datasets/road_signs/labels/val/')
# move_files_to_folder(test_annotations, 'datasets/road_signs/labels/test/')