def train(cfg):
model = build_segmentation_model(cfg)
device = torch.device(cfg.MODEL.DEVICE)
model.to(device)
optimizer = make_optimizer(cfg, model)
# scheduler = make_lr_scheduler(cfg, optimizer)
scheduler = None
arguments = {}
arguments["epoch"] = 0
output_dir = cfg.OUTPUT_DIR
max_epoch = cfg.SOLVER.MAX_EPOCH
checkpointer = SegmentationCheckpointer(
cfg, model, optimizer, scheduler, output_dir, save_to_disk=True
)
extra_checkpoint_data = checkpointer.load(cfg.MODEL.WEIGHT)
arguments.update(extra_checkpoint_data)
train_data_loader = make_data_loader(
cfg,
split='train'
)
val_data_loader = make_data_loader(
cfg,
split='val'
)
checkpoint_period = cfg.SOLVER.CHECKPOINT_PERIOD
do_train(
cfg,
model,
train_data_loader,
val_data_loader,
optimizer,
scheduler,
checkpointer,
device,
checkpoint_period,
arguments,
max_epoch,
)
return model
def train(cfg, local_rank, distributed):
model = build_detection_model(cfg)
device = torch.device(cfg.MODEL.DEVICE)
model.to(device)
if cfg.MODEL.USE_SYNCBN:
assert is_pytorch_1_1_0_or_later(), \
"SyncBatchNorm is only available in pytorch >= 1.1.0"
model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model)
optimizer = make_optimizer(cfg, model)
scheduler = make_lr_scheduler(cfg, optimizer)
if distributed:
model = torch.nn.parallel.DistributedDataParallel(
model,
device_ids=[local_rank],
output_device=local_rank,
# this should be removed if we update BatchNorm stats
broadcast_buffers=False,
# find_unused_parameters=True,
)
arguments = {}
arguments["iteration"] = 0
output_dir = cfg.OUTPUT_DIR
save_to_disk = get_rank() == 0
checkpointer = DetectronCheckpointer(cfg, model, optimizer, scheduler,
output_dir, save_to_disk)
extra_checkpoint_data = checkpointer.load(cfg.MODEL.WEIGHT)
arguments.update(extra_checkpoint_data)
data_loader = make_data_loader(
cfg,
is_train=True,
is_distributed=distributed,
start_iter=arguments["iteration"],
)
checkpoint_period = cfg.SOLVER.CHECKPOINT_PERIOD
vis_period = cfg.VISUALIZE.PERIOD
if 0 < vis_period < cfg.SOLVER.MAX_ITER:
visualizer = SummaryWriterX(
cfg.VISUALIZE.DIR + '/' + cfg.VISUALIZE.ENV, cfg.VISUALIZE.ENV,
vis_period, 20, get_category(cfg.DATASETS.TRAIN[0]))
else:
visualizer = None
meters = MetricLogger(delimiter=" ",
save_dir=os.path.join(output_dir, 'meters.json'))
meters.load(is_main_process=get_rank() == 0)
do_train(model, data_loader, optimizer, scheduler, checkpointer, device,
checkpoint_period, arguments, meters, visualizer)
return model
def main():
parser = argparse.ArgumentParser(description="PyTorch Segmentation Inference")
parser.add_argument(
"--config-file",
default="./configs/Encoder_UNet.yaml",
metavar="FILE",
help="path to config file",
)
parser.add_argument(
"opts",
help="Modify config options using the command-line",
default=None,
nargs=argparse.REMAINDER,
)
args = parser.parse_args()
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
cfg.freeze()
save_dir = cfg.OUTPUT_DIR
logger = setup_logger("core", save_dir)
logger.info(cfg)
logger.info("Collecting env info (might take some time)")
logger.info("\n" + collect_env_info())
model = build_segmentation_model(cfg)
model.to(cfg.MODEL.DEVICE)
output_dir = cfg.OUTPUT_DIR
checkpointer = SegmentationCheckpointer(cfg, model, save_dir=output_dir)
_ = checkpointer.load(cfg.MODEL.WEIGHT)
dataset_names = cfg.DATASETS.TEST
output_folders = [None] * len(cfg.DATASETS.TEST)
if cfg.OUTPUT_DIR:
for idx, dataset_name in enumerate(dataset_names):
output_folder = os.path.join(cfg.OUTPUT_DIR, "inference", cfg.MODEL.ENCODER + '_' + cfg.MODEL.ARCHITECTURE,
dataset_name)
mkdir(output_folder)
output_folders[idx] = output_folder
else:
raise RuntimeError("Output directory is missing!")
test_data_loaders = make_data_loader(cfg, split='test')
for output_folder, dataset_name, test_data_loader in zip(output_folders, dataset_names, test_data_loaders):
inference(
model,
test_data_loader,
dataset_name=dataset_name,
device=cfg.MODEL.DEVICE,
output_folder=output_folder,
)
def run_test(cfg, model):
dataset_names = cfg.DATASETS.TEST
if cfg.OUTPUT_DIR:
for idx, dataset_name in enumerate(dataset_names):
output_folder = os.path.join(cfg.OUTPUT_DIR, "inference", dataset_name)
mkdir(output_folder)
output_folders[idx] = output_folder
else:
raise RuntimeError("Output directory is missing!")
test_data_loaders = make_data_loader(cfg, split='test')
for output_folder, dataset_name, test_data_loader in zip(output_folders, dataset_names, test_data_loaders):
inference(
model,
test_data_loader,
dataset_name=dataset_name,
device=cfg.MODEL.DEVICE,
output_folder=output_folder,
)
def run_test(cfg, model, distributed):
if distributed:
model = model.module
torch.cuda.empty_cache() # TODO check if it helps
iou_types = ("bbox", )
if cfg.MODEL.MASK_ON:
iou_types = iou_types + ("segm", )
if cfg.MODEL.KEYPOINT_ON:
iou_types = iou_types + ("keypoints", )
output_folders = [None] * len(cfg.DATASETS.TEST)
dataset_names = cfg.DATASETS.TEST
if cfg.OUTPUT_DIR:
for idx, dataset_name in enumerate(dataset_names):
output_folder = os.path.join(cfg.OUTPUT_DIR, "inference",
dataset_name)
mkdir(output_folder)
output_folders[idx] = output_folder
data_loaders_val = make_data_loader(cfg,
is_train=False,
is_distributed=distributed)
for output_folder, dataset_name, data_loader_val in zip(
output_folders, dataset_names, data_loaders_val):
inference(
model,
data_loader_val,
dataset_name=dataset_name,
iou_types=iou_types,
box_only=False if cfg.MODEL.FCOS_ON or cfg.MODEL.PACKDET_ON
or cfg.MODEL.RETINAPACK_ON or cfg.MODEL.RETINANET_ON else
cfg.MODEL.RPN_ONLY,
device=cfg.MODEL.DEVICE,
expected_results=cfg.TEST.EXPECTED_RESULTS,
expected_results_sigma_tol=cfg.TEST.EXPECTED_RESULTS_SIGMA_TOL,
output_folder=output_folder,
)
synchronize()
def main():
parser = argparse.ArgumentParser(description="PyTorch Object Detection Inference")
parser.add_argument(
"--config-file",
default="/private/home/fmassa/github/detectron.pytorch_v2/configs/e2e_faster_rcnn_R_50_C4_1x_caffe2.yaml",
metavar="FILE",
help="path to config file",
)
parser.add_argument("--local_rank", type=int, default=0)
parser.add_argument(
"opts",
help="Modify config options using the command-line",
default=None,
nargs=argparse.REMAINDER,
)
args = parser.parse_args()
num_gpus = int(os.environ["WORLD_SIZE"]) if "WORLD_SIZE" in os.environ else 1
distributed = num_gpus > 1
if distributed:
torch.cuda.set_device(args.local_rank)
torch.distributed.init_process_group(
backend="nccl", init_method="env://"
)
synchronize()
cfg.merge_from_file(args.config_file)
cfg.merge_from_list(args.opts)
cfg.freeze()
save_dir = ""
logger = setup_logger("core", save_dir, get_rank())
logger.info("Using {} GPUs".format(num_gpus))
logger.info(cfg)
logger.info("Collecting env info (might take some time)")
logger.info("\n" + collect_env_info())
model = build_detection_model(cfg)
model.to(cfg.MODEL.DEVICE)
num_parameters = sum([param.nelement() for param in model.parameters()])
logger.info('# parameters totally: '+str(num_parameters))
output_dir = cfg.OUTPUT_DIR
checkpointer = DetectronCheckpointer(cfg, model, save_dir=output_dir)
_ = checkpointer.load(cfg.MODEL.WEIGHT, is_train=False)
suffix = cfg.MODEL.WEIGHT.split('/')[-1][:-4]
iou_types = ("bbox",)
if cfg.MODEL.MASK_ON:
iou_types = iou_types + ("segm",)
if cfg.MODEL.KEYPOINT_ON:
iou_types = iou_types + ("keypoints",)
output_folders = [None] * len(cfg.DATASETS.TEST)
dataset_names = cfg.DATASETS.TEST
if cfg.OUTPUT_DIR:
for idx, dataset_name in enumerate(dataset_names):
output_folder = os.path.join(cfg.OUTPUT_DIR, "inference_"+suffix, dataset_name)
mkdir(output_folder)
output_folders[idx] = output_folder
data_loaders_val = make_data_loader(cfg, is_train=False, is_distributed=distributed)
for output_folder, dataset_name, data_loader_val in zip(output_folders, dataset_names, data_loaders_val):
inference(
model,
data_loader_val,
dataset_name=dataset_name,
iou_types=iou_types,
box_only=False if cfg.MODEL.FCOS_ON or
cfg.MODEL.PACKDET_ON or
cfg.MODEL.RETINAPACK_ON or
cfg.MODEL.RETINANET_ON else
cfg.MODEL.RPN_ONLY,
device=cfg.MODEL.DEVICE,
expected_results=cfg.TEST.EXPECTED_RESULTS,
expected_results_sigma_tol=cfg.TEST.EXPECTED_RESULTS_SIGMA_TOL,
output_folder=output_folder,
)
synchronize()