def setup(args):
"""
Create configs and perform basic setups.
"""
cfg = get_cfg()
add_tensormask_config(cfg)
# set config file
cfg.merge_from_file(
"configs/BDD00K-InstanceSegmentation/tensormask_r101_3x_single_scale_bs8_bdd100k.yaml"
)
cfg.DATASETS.TRAIN = ("bdd100k_train", )
cfg.DATASETS.TEST = ("bdd100k_test", )
# cfg.MODEL.WEIGHTS = model_zoo.get_checkpoint_url("COCO-Detection/faster_rcnn_X_101_32x8d_FPN_3x.yaml") # Let training initialize from model zoo
cfg.OUTPUT_DIR = './tensormask_r101_3x_single_scale_bs16_bdd100k'
os.makedirs(cfg.OUTPUT_DIR, exist_ok=True)
# new added solver arguments
cfg.SOLVER.CHECKPOINT_PERIOD = 500
cfg.TEST.EVAL_PERIOD = 500
# end of new arguments
# cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
cfg.freeze()
default_setup(cfg, args)
return cfg
def load_tensormask_model(model_path, cfg_path, device):
"""
Load the pretrained TensorMask model states and prepare the model for image segmentation.
Paramters
---------
model_path: str
Path to the pretrained model states binary file.
cfg_path: str
Path to the model's Configuration file. Located in the .configs folder.
device: torch.device
Device to load the model on.
Returns
-------
model: TensorMask
Model with the loaded pretrained states.
"""
# set up model config
cfg = get_cfg()
tensormask.add_tensormask_config(cfg)
cfg.merge_from_file(cfg_path)
model = build_model(cfg)
# load the model weights
DetectionCheckpointer(model).load(model_path)
model.eval()
return model
def setup(args):
"""
Create configs and perform basic setups.
"""
cfg = get_cfg()
add_tensormask_config(cfg)
cfg.merge_from_file(f'tensormask/configs/{args.config_file}.yaml')
cfg.merge_from_list(args.opts)
if args.log_dir:
cfg.OUTPUT_DIR_BASE = args.log_dir
if args.data_dir:
cfg.DATASETS.TRAIN = (args.data_dir,)
cfg.MODEL.DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
print('device:', cfg.MODEL.DEVICE)
register_datasets(cfg.DATASETS.TRAIN)
register_datasets(cfg.DATASETS.TEST)
# setup up logging directory
cfg.OUTPUT_DIR = os.path.join(cfg.OUTPUT_DIR_BASE, args.config_file)
os.makedirs(cfg.OUTPUT_DIR, exist_ok=True)
cfg.freeze()
default_setup(cfg, args)
return cfg
def setup(args):
"""
Create configs and perform basic setups.
"""
cfg = get_cfg()
add_tensormask_config(cfg)
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
cfg.freeze()
default_setup(cfg, args)
return cfg
def setup(args):
"""
Create configs and perform basic setups.
"""
cfg = get_cfg()
add_tensormask_config(cfg)
cfg.merge_from_file(
"/root/detectron2/projects/TensorMask/configs/tensormask_R_50_FPN_1x.yaml"
)
# cfg.merge_from_list(args.opts)
cfg.MODEL.WEIGHTS = "/root/detectron2/projects/TensorMask/log_80_20/model_0024999.pth"
return cfg
def setup(args):
"""
Create configs and perform basic setups.
"""
cfg = get_cfg()
add_tensormask_config(cfg)
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
if args.eval_only:
cfg.MODEL.WEIGHTS = "/root/detectron2/projects/TensorMask/log_50_50/model_0034999.pth"
cfg.SOLVER.IMS_PER_BATCH = 6
cfg.freeze()
default_setup(cfg, args)
return cfg