def predict_darkflow_model(cfg_obj, args):
options = darkflow_parse_config(cfg_obj)
# erase previous predictions
imgdir_path = options['imgdir']
out_dir = os.path.join(imgdir_path, 'out')
for the_file in os.listdir(out_dir):
file_path = os.path.join(out_dir, the_file)
try:
if os.path.isfile(file_path):
os.unlink(file_path)
except Exception as e:
print(e)
# load args
if args.load is None:
raise AssertionError('Please provide a ckpt number to load from.')
options['load'] = args.load if args.load == 'best' else int(args.load)
if args.threshold is not None:
options['threshold'] = float(args.threshold)
if args.gpu is not None:
options['gpu'] = float(args.gpu)
tfnet = TFNet(options)
tfnet.predict()
# save in json format too
options['json'] = True
tfnet = TFNet(options)
assert tfnet.FLAGS.threshold >= 0
tfnet.predict()
def predict_darkflow_model(cfg_obj,args):
options = darkflow_parse_config(cfg_obj)
if args.load is None:
raise AssertionError('Please provide a ckpt number to load from.')
options['load'] = int(args.load)
if args.threshold is not None:
options['threshold'] = float(args.threshold)
tfnet = TFNet(options)
tfnet.predict()
# save in json format too
options['json'] = True
tfnet = TFNet(options)
tfnet.predict()
def image_prediction():
current_dir = os.getcwd() + "/app/cnnDashboard"
FLAGS = {
'imgdir': f'{current_dir}/sample_img/',
'binary': f'{current_dir}/bin/',
'config': f'{current_dir}/cfg/',
'backup': './ckpt/',
'threshold': -0.1,
'model': f'{current_dir}/cfg/yolo.cfg',
'load': f'{current_dir}/bin/yolo.weights',
'gpu': 0.0,
'batch': 16
}
tfnet = TFNet(FLAGS)
tfnet.predict()
def cliHandler(args):
FLAGS = argHandler()
FLAGS.setDefaults()
FLAGS.parseArgs(args)
# make sure all necessary dirs exist
def _get_dir(dirs):
for d in dirs:
this = os.path.abspath(os.path.join(os.path.curdir, d))
if not os.path.exists(this): os.makedirs(this)
_get_dir([
FLAGS.imgdir, FLAGS.binary, FLAGS.backup,
os.path.join(FLAGS.imgdir, 'out'), FLAGS.summary
])
# fix FLAGS.load to appropriate type
try:
FLAGS.load = int(FLAGS.load)
except:
pass
tfnet = TFNet(FLAGS)
if FLAGS.demo:
tfnet.camera()
exit('Demo stopped, exit.')
if FLAGS.train:
print('Enter training ...')
tfnet.train()
if not FLAGS.savepb:
exit('Training finished, exit.')
if FLAGS.savepb:
print('Rebuild a constant version ...')
tfnet.savepb()
exit('Done')
tfnet.predict()
options = {
'model': 'config/tiny-yolo-dolphin.cfg',
'load': -1,
'threshold': 0.1,
'labels': './labels_dolphin.txt',
'json': False,
}
usage = """
Usage:
./test_yolo.py [options]
Options:
--img=FILE The image file to test.
--imgdir=DIR The image folder to test.
"""
if __name__ == '__main__':
args = docopt(usage, help=True)
if args['--img']:
tfnet = TFNet(options)
imgcv = cv2.imread(os.path.abspath(args['--img']))
result = tfnet.return_predict(imgcv)
pprint(result)
if args['--imgdir']:
options['imgdir'] = os.path.abspath(args['--imgdir'])
tfnet = TFNet(options)
tfnet.predict()
from darkflow.net.build import TFNet
import cv2
options = {
# 訓練済みモデル
"load": "./data/weights/maeharin_yolo-obj_900.weights",
# 訓練するときに使ったyoloのコンフィグファイル
"model": "./data/cfg/maeharin_yolo-obj.cfg",
# ラベル
"labels": "./data/labels/obj.names",
# 閾値
"threshold": 0.3
}
tfnet = TFNet(options)
# 予測対象の画像
imgcv = cv2.imread("./data/sample-images/IMG_0411_frame_1.jpg")
result = tfnet.return_predict(imgcv)
print(result)
tfnet.predict(imgcv)
