def run_chunk(self, mode):
self.dataset = get_dataset(self.cfg, self.augmentor, mode)
if mode == 'train':
num_chunk = self.total_iter_nums // self.cfg.DATASET.DATA_CHUNK_ITER
self.total_iter_nums = self.cfg.DATASET.DATA_CHUNK_ITER
for chunk in range(num_chunk):
self.dataset.updatechunk()
self.dataloader = build_dataloader(
self.cfg,
self.augmentor,
mode,
dataset=self.dataset.dataset)
self.dataloader = iter(self.dataloader)
print('start train', chunk)
self.train()
print('finished train', chunk)
self.start_iter += self.cfg.DATASET.DATA_CHUNK_ITER
del self.dataloader
else:
num_chunk = len(self.dataset.chunk_num_ind)
for chunk in range(num_chunk):
self.dataset.updatechunk(do_load=False)
self.inference_output_name = self.cfg.INFERENCE.OUTPUT_NAME + self.dataset.get_coord_name(
) + '.h5'
if not os.path.exists(
os.path.join(self.output_dir,
self.inference_output_name)):
self.dataset.loadchunk()
self.dataloader = build_dataloader(
self.cfg,
self.augmentor,
mode,
dataset=self.dataset.dataset)
self.dataloader = iter(self.dataloader)
self.test()
def __init__(self, cfg, device, mode, checkpoint=None):
self.cfg = cfg
self.device = device
self.output_dir = cfg.DATASET.OUTPUT_PATH
self.mode = mode
self.model = build_model(self.cfg, self.device)
self.optimizer = build_optimizer(self.cfg, self.model)
self.lr_scheduler = build_lr_scheduler(self.cfg, self.optimizer)
self.start_iter = self.cfg.MODEL.PRE_MODEL_ITER
if checkpoint is not None:
self.update_checkpoint(checkpoint)
if self.mode == 'train':
self.augmentor = build_train_augmentor(self.cfg)
else:
self.augmentor = None
self.dataloader = build_dataloader(self.cfg, self.augmentor, self.mode)
self.monitor = build_monitor(self.cfg)
self.criterion = build_criterion(self.cfg, self.device)
# add config details to tensorboard
self.monitor.load_config(self.cfg)
self.dataloader = iter(
self.dataloader
) #iter() function returns an iterator for the given object. The iter() function creates an object which can be iterated one element at a time.
def __init__(self, cfg, device, mode, checkpoint=None):
self.cfg = cfg
self.device = device
self.output_dir = cfg.DATASET.OUTPUT_PATH
self.mode = mode
self.model = build_model(self.cfg, self.device)
self.optimizer = build_optimizer(self.cfg, self.model)
self.lr_scheduler = build_lr_scheduler(self.cfg, self.optimizer)
self.start_iter = self.cfg.MODEL.PRE_MODEL_ITER
if checkpoint is not None:
self.update_checkpoint(checkpoint)
if self.mode == 'train':
self.augmentor = build_train_augmentor(self.cfg)
self.monitor = build_monitor(self.cfg)
self.criterion = build_criterion(self.cfg, self.device)
# add config details to tensorboard
self.monitor.load_config(self.cfg)
else:
self.augmentor = None
if cfg.DATASET.DO_CHUNK_TITLE == 0:
self.dataloader = build_dataloader(self.cfg, self.augmentor,
self.mode)
self.dataloader = iter(self.dataloader)
else:
self.dataset = None
self.dataloader = None
self.total_iter_nums = self.cfg.SOLVER.ITERATION_TOTAL - self.start_iter
self.inference_output_name = self.cfg.INFERENCE.OUTPUT_NAME
self.total_time = 0
def __init__(self,
cfg,
device,
mode,
output_dir='outputs/',
checkpoint=None):
self.cfg = cfg
self.device = device
self.output_dir = output_dir
self.mode = mode
self.model = build_model(self.cfg, self.device, checkpoint)
self.optimizer = build_optimizer(self.cfg, self.model)
self.lr_scheduler = build_lr_scheduler(self.cfg, self.optimizer)
if self.mode == 'train':
self.augmentor = build_train_augmentor(self.cfg)
else:
self.augmentor = None
self.dataloader = build_dataloader(self.cfg, self.augmentor, self.mode)
self.monitor = build_monitor(self.cfg)
self.criterion = build_criterion(self.cfg, self.device)
self.dataloader = iter(self.dataloader)