def setup_cfg(args):
# load config from file and command-line arguments
cfg = get_cfg()
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
# Set score_threshold for builtin models
#############################################################
#추가한 코드
cfg.MODEL.WEIGHTS = "/data_2/jongwon/output/stefan_dcn/model_0043999.pth"
num_classes = 5
cfg.MODEL.ROI_HEADS.NUM_CLASSES = num_classes
cfg.MODEL.SEM_SEG_HEAD.NUM_CLASSES = num_classes
cfg.MODEL.RETINANET.NUM_CLASSES = num_classes
cfg.MODEL.FCOS.NUM_CLASSES = num_classes
#############################################################
cfg.MODEL.RETINANET.SCORE_THRESH_TEST = args.confidence_threshold
cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = args.confidence_threshold
cfg.MODEL.FCOS.INFERENCE_TH_TEST = args.confidence_threshold
cfg.MODEL.PANOPTIC_FPN.COMBINE.INSTANCES_CONFIDENCE_THRESH = args.confidence_threshold
cfg.freeze()
return cfg
def setup(args):
"""
Create configs and perform basic setups.
"""
cfg = get_cfg()
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
cfg.freeze()
default_setup(cfg, args)
return cfg
def setup_cfg(args):
# load config from file and command-line arguments
cfg = get_cfg()
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
# Set score_threshold for builtin models
cfg.MODEL.RETINANET.SCORE_THRESH_TEST = args.confidence_threshold
cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = args.confidence_threshold
cfg.MODEL.FCOS.INFERENCE_TH_TEST = args.confidence_threshold
cfg.MODEL.PANOPTIC_FPN.COMBINE.INSTANCES_CONFIDENCE_THRESH = args.confidence_threshold
cfg.freeze()
return cfg
def setup(args):
"""
Create configs and perform basic setups.
"""
cfg = get_cfg()
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
if args.eval_only:
cfg.MODEL.WEIGHTS = "/root/centermask2/log_50_50/CenterMask-R-50-FPN-ms-3x/model_0009999.pth"
cfg.SOLVER.IMS_PER_BATCH = 6
cfg.freeze()
default_setup(cfg, args)
return cfg
def setup(args):
"""
Create configs and perform basic setups.
"""
cfg = get_cfg()
cfg.merge_from_file(CONFIG_FILE_PATH)
cfg.merge_from_list(args.opts)
cfg.MODEL.WEIGHTS = WEIGHT_PATH
cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = THRESH_TEST
cfg.MODEL.RETINANET.SCORE_THRESH_TEST = THRESH_TEST
cfg.MODEL.PANOPTIC_FPN.COMBINE.INSTANCES_CONFIDENCE_THRESH = THRESH_TEST
cfg.MODEL.FCOS.INFERENCE_TH_TEST = THRESH_TEST
cfg.MODEL.DEVICE = 'cpu'
cfg.freeze()
default_setup(cfg, args)
return cfg
def loadModel(self, nomeroffnet_path = "../",
subdir="./NomeroffNet/configs/centermask2/numberplates/",
config_file='centermask_numberplate_V_39_eSE_FPN_ms_3x.yaml'):
"""
Create configs and perform basic setups.
TODO: create folder config/centermask2/ and put all architecture them
"""
centermask2_path= os.path.join(nomeroffnet_path, "centermask2")
sys.path.append(centermask2_path)
from centermask.config import get_cfg
if get_mode() == "cpu":
config_file = f"cpu_{config_file}"
config_file = os.path.join(nomeroffnet_path, subdir, config_file)
cfg = get_cfg()
cfg.merge_from_file(config_file)
cfg.freeze()
self.predictor = DefaultPredictor(cfg)
def setup_cfg(params):
# load config from file and command-line arguments
cfg = get_cfg()
cfg.merge_from_file(params["is_config_file"])
cfg.merge_from_list([])
# Set score_threshold for builtin models
cfg.MODEL.WEIGHTS = params["is_weight_path"]
num_classes = len(params["class_names"])
cfg.MODEL.ROI_HEADS.NUM_CLASSES = num_classes
cfg.MODEL.SEM_SEG_HEAD.NUM_CLASSES = num_classes
cfg.MODEL.RETINANET.NUM_CLASSES = num_classes
cfg.MODEL.FCOS.NUM_CLASSES = num_classes
cfg.MODEL.RETINANET.SCORE_THRESH_TEST = params["is_thresh"]
cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = params["is_thresh"]
cfg.MODEL.FCOS.INFERENCE_TH_TEST = params["is_thresh"]
cfg.MODEL.PANOPTIC_FPN.COMBINE.INSTANCES_CONFIDENCE_THRESH = params[
"is_thresh"]
cfg.freeze()
return cfg
def setup(args):
"""
Create configs and perform basic setups.
"""
cfg = get_cfg()
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
cfg.SOLVER.BASE_LR = 0.005
cfg.SOLVER.IMS_PER_BATCH = 2
# cfg.SOLVER.CHECKPOINT_PERIOD = 5000
# cfg.SOLVER.MAX_ITER = 200
cfg.MODEL.ROI_HEADS.NUM_CLASSES = 5
cfg.MODEL.FCOS.NUM_CLASSES = 5
cfg.MY_CUSTOM.LOG_FILE = os.path.join(cfg.OUTPUT_DIR, 'my_log.txt')
cfg.freeze()
default_setup(
cfg, args
) # if you don't like any of the default setup, write your own setup code
return cfg
def setup(args):
"""
Create configs and perform basic setups.
"""
cfg = get_cfg()
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
num_classes = 5
cfg.MODEL.ROI_HEADS.NUM_CLASSES = num_classes
cfg.MODEL.SEM_SEG_HEAD.NUM_CLASSES = num_classes
cfg.MODEL.RETINANET.NUM_CLASSES = num_classes
cfg.MODEL.FCOS.NUM_CLASSES = num_classes
#cfg.DATALOADER.NUM_WORKERS = 0
cfg.INPUT.ALBUMENTATIONS = "mapper/albu-config.json"
cfg.freeze()
default_setup(cfg, args)
return cfg
def setup(args):
"""
Create configs and perform basic setups.
"""
DatasetCatalog.register(
"carplate_train", lambda x='train': get_carplate_dicts(x, ROOT, 0.001))
DatasetCatalog.register("carplate_val",
lambda x='val': get_carplate_dicts(x, ROOT, 0.001))
MetadataCatalog.get("carplate_val").set(thing_classes=["carplate"])
# carplate_metadata = MetadataCatalog.get("carplate_train")
MetadataCatalog.get("carplate_val").set(evaluator_type='coco')
cfg = get_cfg()
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
cfg.DATASETS.TRAIN = ("carplate_train", )
cfg.DATASETS.TEST = ("carplate_val", )
cfg.MODEL.DEVICE = 'cuda'
cfg.TEST.EVAL_PERIOD = 1000
cfg.SOLVER.WARMUP_ITERS = 1000
cfg.SOLVER.CHECKPOINT_PERIOD = 3000
cfg.MODEL.ROI_HEADS.BATCH_SIZE_PER_IMAGE = 64
cfg.DATALOADER.NUM_WORKERS = 2
cfg.SOLVER.IMS_PER_BATCH = 1
cfg.MODEL.ROI_HEADS.NUM_CLASSES = 1
cfg.SOLVER.GAMMA = 0.05
cfg.SOLVER.MAX_ITER = 30000
cfg.SOLVER.STEPS = (6000, 10000, 15000, 19000, 25000, 29000)
cfg.SOLVER.BASE_LR = 0.00005
cfg.MODEL.WEIGHTS = os.path.join(ROOT, WEIGHTS, "X-101-32x8d.pkl")
# cfg.freeze()
default_setup(cfg, args)
return cfg
#Register Datasets
DatasetCatalog.register('openimages_train', get_train_dicts)
MetadataCatalog.get('openimages_train').set(thing_classes=classes)
openimages_train_metadata = MetadataCatalog.get('openimages_train')
#Visualizing datasets
# train_dicts = get_train_dicts()
# for d in random.sample(train_dicts, 10):
# print(d)
# img = cv2.imread(d["file_name"])
# visualizer = Visualizer(img[:,:,::-1], metadata=openimages_train_metadata, scale=0.5)
# vis = visualizer.draw_dataset_dict(d)
# cv2.imshow("image", vis.get_image()[:,:,::-1])
# cv2.waitKey()
cfg = get_cfg()
#Training Configs
cfg.merge_from_file(
'configs/centermask/Base-CenterMask-Lite-EfficientNet.yml')
cfg.MODEL.WEIGHTS = 'C:\\Users\\Admin\\Documents\\detectron2\\projects\\CenterMask2\\output\\centermask\\CenterMask-Lite-Efficientnet-2x\\efficientnet-pretrained\\model_final_wo_solver_states.pth'
cfg.DATASETS.TRAIN = ('openimages_train', )
cfg.DATASETS.TEST = ()
cfg.DATALOADER.NUM_WORKERS = 2
cfg.SOLVER.IMS_PER_BATCH = 2
cfg.SOLVER.MAX_ITER = 300000
cfg.SOLVER.BASE_LR = 0.00003
cfg.SOLVER.GAMMA = 0.2
cfg.SOLVER.STEPS = (
150000,
220000,
