zjc
/
yolov5
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Objects365 update

This commit is contained in:
Glenn Jocher 2021-05-01 00:16:58 +02:00
parent 801b469878
commit 54652fe3ff
2 changed files with 24 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
data/scripts/get_objects365.py
View File

@ -7,22 +7,34 @@
# /images # /images
# /labels # /labels
from pycocotools.coco import COCO from pycocotools.coco import COCO
coco = COCO("zhiyuan_objv2_train.json") from utils.general import download, Path
cats = coco.loadCats(coco.getCatIds())
nms = [cat["name"] for cat in cats] # Make Directories
print("COCO categories: \n{}\n".format(" ".join(nms))) dir = Path('../datasets/objects365') # dataset directory
for categoryId, cat in enumerate(nms): for p in 'images', 'labels':
(dir / p).mkdir(parents=True, exist_ok=True)
for q in 'train', 'val':
(dir / p / q).mkdir(parents=True, exist_ok=True)
# Download
url = "https://dorc.ks3-cn-beijing.ksyun.com/data-set/2020Objects365%E6%95%B0%E6%8D%AE%E9%9B%86/train/"
download(url + 'zhiyuan_objv2_train.tar.gz', dir=dir, threads=8) # annotations json
download([url + f for f in [f'patch{i}.tar.gz' for i in range(51)]], dir=dir / 'images' / 'train', threads=8)
# Labels
coco = COCO(dir / 'zhiyuan_objv2_train.json')
names = [x["name"] for x in coco.loadCats(coco.getCatIds())]
for categoryId, cat in enumerate(names):
catIds = coco.getCatIds(catNms=[cat]) catIds = coco.getCatIds(catNms=[cat])
imgIds = coco.getImgIds(catIds=catIds) imgIds = coco.getImgIds(catIds=catIds)
print(cat)
# Create a subfolder in this directory called "labels". This is where the annotations will be saved in YOLO format
for im in coco.loadImgs(imgIds): for im in coco.loadImgs(imgIds):
width, height = im["width"], im["height"] width, height = im["width"], im["height"]
path = im["file_name"].split("/")[-1] # image filename path = Path(im["file_name"]) # image filename
try: try:
with open("labels/train/" + path.replace(".jpg", ".txt"), "a+") as file: with open(dir / 'labels' / 'train' / path.with_suffix('.txt').name, 'a') as file:
annIds = coco.getAnnIds(imgIds=im["id"], catIds=catIds, iscrowd=None) annIds = coco.getAnnIds(imgIds=im["id"], catIds=catIds, iscrowd=None)
for a in coco.loadAnns(annIds): for a in coco.loadAnns(annIds):
x, y, w, h = a['bbox'] # bounding box in xywh (xy top-left corner) x, y, w, h = a['bbox'] # bounding box in xywh (xy top-left corner)

6
utils/general.py
View File

@ -183,7 +183,7 @@ def check_dataset(dict):
raise Exception('Dataset not found.') raise Exception('Dataset not found.')
def download(url, dir='.', multi_thread=False): def download(url, dir='.', threads=1):
# Multi-threaded file download and unzip function # Multi-threaded file download and unzip function
def download_one(url, dir): def download_one(url, dir):
# Download 1 file # Download 1 file
@ -200,8 +200,8 @@ def download(url, dir='.', multi_thread=False):
dir = Path(dir) dir = Path(dir)
dir.mkdir(parents=True, exist_ok=True) # make directory dir.mkdir(parents=True, exist_ok=True) # make directory
if multi_thread: if threads > 1:
ThreadPool(8).imap(lambda x: download_one(*x), zip(url, repeat(dir))) # 8 threads ThreadPool(threads).imap(lambda x: download_one(*x), zip(url, repeat(dir))) # multi-threaded
else: else:
for u in tuple(url) if isinstance(url, str) else url: for u in tuple(url) if isinstance(url, str) else url:
download_one(u, dir) download_one(u, dir)