def get_eval_model(self):
# Load model, or construct model and load weights.
num_anchors = len(self.anchors)
num_classes = len(self.class_names)
#YOLOv3 model has 9 anchors and 3 feature layers but
#Tiny YOLOv3 model has 6 anchors and 2 feature layers,
#so we can calculate feature layers number to get model type
num_feature_layers = num_anchors // 3
if num_anchors == 5:
# YOLOv2 use 5 anchors
eval_model, _ = get_yolo2_model(self.model_type,
num_anchors,
num_classes,
input_shape=self.model_image_size +
(3, ),
model_pruning=self.model_pruning)
else:
eval_model, _ = get_yolo3_model(self.model_type,
num_feature_layers,
num_anchors,
num_classes,
input_shape=self.model_image_size +
(3, ),
model_pruning=self.model_pruning)
return eval_model
def _generate_model(self):
'''to generate the bounding boxes'''
weights_path = os.path.expanduser(self.weights_path)
assert weights_path.endswith('.h5'), 'Keras model or weights must be a .h5 file.'
# Load model, or construct model and load weights.
num_anchors = len(self.anchors)
num_classes = len(self.class_names)
#YOLOv3 model has 9 anchors and 3 feature layers but
#Tiny YOLOv3 model has 6 anchors and 2 feature layers,
#so we can calculate feature layers number to get model type
num_feature_layers = num_anchors//3
try:
if num_anchors == 5:
# YOLOv2 use 5 anchors
yolo_model, _ = get_yolo2_model(self.model_type, num_anchors, num_classes, input_shape=self.model_image_size + (3,), model_pruning=self.pruning_model)
else:
yolo_model, _ = get_yolo3_model(self.model_type, num_feature_layers, num_anchors, num_classes, input_shape=self.model_image_size + (3,), model_pruning=self.pruning_model)
yolo_model.load_weights(weights_path) # make sure model, anchors and classes match
if self.pruning_model:
yolo_model = sparsity.strip_pruning(yolo_model)
yolo_model.summary()
except Exception as e:
print(repr(e))
assert yolo_model.layers[-1].output_shape[-1] == \
num_anchors/len(yolo_model.output) * (num_classes + 5), \
'Mismatch between model and given anchor and class sizes'
print('{} model, anchors, and classes loaded.'.format(weights_path))
if self.gpu_num>=2:
yolo_model = multi_gpu_model(yolo_model, gpus=self.gpu_num)
return yolo_model
def _generate_model(self):
'''to generate the bounding boxes'''
weights_path = os.path.expanduser(self.weights_path)
assert weights_path.endswith(
'.h5'), 'Keras model or weights must be a .h5 file.'
# Load model, or construct model and load weights.
num_anchors = len(self.anchors)
num_classes = len(self.class_names)
#YOLOv3 model has 9 anchors and 3 feature layers but
#Tiny YOLOv3 model has 6 anchors and 2 feature layers,
#so we can calculate feature layers number to get model type
num_feature_layers = num_anchors // 3
yolo_model, _ = get_yolo3_model(self.model_type,
num_feature_layers,
num_anchors,
num_classes,
input_shape=self.model_image_size +
(3, ),
model_pruning=self.pruning_model)
yolo_model.load_weights(
weights_path) # make sure model, anchors and classes match
#if self.pruning_model:
# yolo_model = sparsity.strip_pruning(yolo_model)
yolo_model.summary()
print('{} model, anchors, and classes loaded.'.format(weights_path))
return yolo_model
def get_eval_model(self):
# Load model, or construct model and load weights.
num_anchors = len(self.anchors)
num_classes = len(self.class_names)
#YOLOv3 model has 9 anchors and 3 feature layers but
#Tiny YOLOv3 model has 6 anchors and 2 feature layers,
#so we can calculate feature layers number to get model type
num_feature_layers = num_anchors // 3
if self.model_type.startswith(
'scaled_yolo4_') or self.model_type.startswith('yolo5_'):
# Scaled-YOLOv4 & YOLOv5 entrance
eval_model, _ = get_yolo5_model(self.model_type,
num_feature_layers,
num_anchors,
num_classes,
input_shape=self.model_image_size +
(3, ),
model_pruning=self.model_pruning)
self.v5_decode = True
elif self.model_type.startswith('yolo3_') or self.model_type.startswith('yolo4_') or \
self.model_type.startswith('tiny_yolo3_') or self.model_type.startswith('tiny_yolo4_'):
# YOLOv3 & v4 entrance
eval_model, _ = get_yolo3_model(self.model_type,
num_feature_layers,
num_anchors,
num_classes,
input_shape=self.model_image_size +
(3, ),
model_pruning=self.model_pruning)
self.v5_decode = False
elif self.model_type.startswith(
'yolo2_') or self.model_type.startswith('tiny_yolo2_'):
# YOLOv2 entrance
eval_model, _ = get_yolo2_model(self.model_type,
num_anchors,
num_classes,
input_shape=self.model_image_size +
(3, ),
model_pruning=self.model_pruning)
self.v5_decode = False
else:
raise ValueError('Unsupported model type')
return eval_model
anchors_path = 'configs/tiny_yolo_anchors.txt'
classes_path = 'yolo_format_training_classes.txt'
model_image_size = (416, 416)
gpu_num = 1
# Load Classes and get anchors for comparision
class_names = get_classes(classes_path)
anchors = get_anchors(anchors_path)
colors = get_colors(class_names)
num_anchors = len(anchors)
num_classes = len(class_names)
num_feature_layers = num_anchors//3
# Initialise the model with same architecture and load the trained model provided in the config
yolo_model, _ = get_yolo3_model(model_type, num_feature_layers, num_anchors, num_classes, input_shape=model_image_size + (3,))
yolo_model.load_weights(model_path) # make sure model, anchors and classes match
print('{} model, anchors, and classes loaded.'.format(model_path))
# Define the Prediction with config parameters
def predict(image_data, image_shape):
out_boxes, out_classes, out_scores = yolo3_postprocess_np(yolo_model.predict(image_data),
image_shape,
anchors,
len(class_names),
model_image_size,
max_boxes=100, confidence = 0.3, iou_threshold=0.4)
return out_boxes, out_classes, out_scores
def _generate_model(self):
'''to generate the bounding boxes'''
weights_path = None
#assert weights_path.endswith('.h5'), 'Keras model or weights must be a .h5 file.'
# Load model, or construct model and load weights.
num_anchors = len(self.anchors)
num_classes = len(self.class_names)
#YOLOv3 model has 9 anchors and 3 feature layers but
#Tiny YOLOv3 model has 6 anchors and 2 feature layers,
#so we can calculate feature layers number to get model type
num_feature_layers = num_anchors // 3
try:
if self.model_type.startswith(
'scaled_yolo4_') or self.model_type.startswith('yolo5_'):
# Scaled-YOLOv4 & YOLOv5 entrance
yolo_model, _ = get_yolo5_model(
self.model_type,
num_feature_layers,
num_anchors,
num_classes,
input_shape=self.model_image_size + (3, ),
model_pruning=self.pruning_model)
elif self.model_type.startswith('yolo3_') or self.model_type.startswith('yolo4_') or \
self.model_type.startswith('tiny_yolo3_') or self.model_type.startswith('tiny_yolo4_'):
# YOLOv3 & v4 entrance
yolo_model, _ = get_yolo3_model(
self.model_type,
num_feature_layers,
num_anchors,
num_classes,
input_shape=self.model_image_size + (3, ),
model_pruning=self.pruning_model)
elif self.model_type.startswith(
'yolo2_') or self.model_type.startswith('tiny_yolo2_'):
# YOLOv2 entrance
yolo_model, _ = get_yolo2_model(
self.model_type,
num_anchors,
num_classes,
input_shape=self.model_image_size + (3, ),
model_pruning=self.pruning_model)
else:
raise ValueError('Unsupported model type')
yolo_model.load_weights(
weights_path) # make sure model, anchors and classes match
if self.pruning_model:
yolo_model = sparsity.strip_pruning(yolo_model)
yolo_model.summary()
except Exception as e:
print(repr(e))
assert yolo_model.layers[-1].output_shape[-1] == \
num_anchors/len(yolo_model.output) * (num_classes + 5), \
'Mismatch between model and given anchor and class sizes'
print('{} model, anchors, and classes loaded.'.format(weights_path))
# tf2.4와의 호환을 위해 아래를 삭제.
'''
if self.gpu_num>=2:
yolo_model = multi_gpu_model(yolo_model, gpus=self.gpu_num)
'''
return yolo_model
