def test_train(setup_tf_eager, setup_darknet_weights, setup_train_dirs):
ann_fnames, image_root = setup_train_dirs
darknet_weights = setup_darknet_weights
# 1. create generator
generator = BatchGenerator(ann_fnames,
image_root,
batch_size=2,
labels_naming=["raccoon"],
jitter=False)
valid_generator = BatchGenerator(ann_fnames,
image_root,
batch_size=2,
labels_naming=["raccoon"],
jitter=False)
# 2. create model
model = Yolonet(n_classes=1)
model.load_darknet_params(darknet_weights, True)
# 3. training
loss_history = train_fn(model, generator, valid_generator, num_epoches=3)
assert loss_history[0] > loss_history[-1]
batch_size=config["train"]["batch_size"],
labels_naming=config["model"]["labels"],
anchors=config["model"]["anchors"],
jitter=False,
shuffle=False)
print(train_generator.steps_per_epoch)
# 2. create model
model = Yolonet(n_classes=len(config["model"]["labels"]))
model.load_darknet_params(config["pretrained"]["darknet_format"], skip_detect_layer=True)
# 4. traini
train_fn(model,
train_generator,
valid_generator,
learning_rate=config["train"]["learning_rate"],
save_dname=config["train"]["save_folder"],
num_epoches=config["train"]["num_epoch"])
# 5. prepare sample images
img_fnames = glob.glob(os.path.join(config["train"]["train_image_folder"], "*.*"))
imgs = [cv2.imread(fname)[:,:,::-1] for fname in img_fnames]
# 6. create new model & load trained weights
model = Yolonet(n_classes=len(config["model"]["labels"]))
model.load_weights(os.path.join(config["train"]["save_folder"], "weights.h5"))
detector = YoloDetector(model)
# 7. predict & plot
boxes = detector.detect(imgs[0], config["model"]["anchors"])
image = draw_boxes(imgs[0], boxes, labels=config["model"]["labels"])
help='config file')
if __name__ == '__main__':
args = argparser.parse_args()
config_parser = ConfigParser(args.config)
# Select device
#tf.debugging.set_log_device_placement(True)
gpus = tf.config.experimental.list_physical_devices('GPU')
device = "/CPU:0" if len(gpus) == 0 else "/GPU:0"
device = "/GPU:1" if TF_GPU_SETUP == "iis" else device
with tf.device(device):
# Create data generator
train_generator, valid_generator = config_parser.create_generator()
# Create the YoloV3 model
model = config_parser.create_model()
# Train the (preloaded) model
learning_rate, save_dname, n_epoches = config_parser.get_train_params()
summary_dir = save_dname + "/summary"
train_fn(config_parser,
model,
train_generator,
valid_generator,
summary_dir=summary_dir,
learning_rate=learning_rate,
save_dname=save_dname,
epoch=n_epoches)
argparser.add_argument('-c',
'--config',
default="configs/test.json",
help='config file')
if __name__ == '__main__':
args = argparser.parse_args()
# config = './configs/svhn.json'
config = args.config
config_parser = ConfigParser(config)
# 1. create generator
split_train_valid = config_parser.split_train_val()
train_generator, valid_generator = config_parser.create_generator(
split_train_valid=split_train_valid)
# 2. create model
model = config_parser.create_model()
# 3. training
learning_rate, save_dir, weight_name, n_epoches, checkpoint_path = config_parser.get_train_params(
)
train_fn(model,
train_generator,
valid_generator,
learning_rate=learning_rate,
save_dir=save_dir,
weight_name=weight_name,
num_epoches=n_epoches,
configs=config_parser)
