def run_test(cfg, model, distributed):
torch.cuda.empty_cache() # TODO check if it helps
iou_types = ("bbox", )
if cfg.MODEL.MASK_ON:
iou_types = iou_types + ("segm", )
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=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 run_test(cfg, model, distributed):
if distributed:
model = model.module
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.RETINANET_ON else cfg.MODEL.RPN_ONLY,
bbox_aug=cfg.TEST.BBOX_AUG.ENABLED,
device=cfg.MODEL.DEVICE,
expected_results=cfg.TEST.EXPECTED_RESULTS,
expected_results_sigma_tol=cfg.TEST.EXPECTED_RESULTS_SIGMA_TOL,
output_folder=output_folder,
)
def train(cfg, local_rank, distributed):
model = build_detection_model(cfg)
optimizer = make_optimizer(cfg, model)
scheduler = make_lr_scheduler(cfg, optimizer)
arguments = {}
arguments["iteration"] = 0
output_dir = cfg.OUTPUT_DIR
checkpointer = DetectronCheckpointer(cfg, model, optimizer, scheduler,
output_dir)
extra_checkpoint_data = checkpointer.load(cfg.MODEL.WEIGHT)
arguments.update(extra_checkpoint_data)
data_loader = make_data_loader(
cfg,
is_train=True,
start_iter=arguments["iteration"],
)
test_period = cfg.SOLVER.TEST_PERIOD
if test_period > 0:
data_loader_val = make_data_loader(cfg,
is_train=False,
is_distributed=distributed,
is_for_period=True)
else:
data_loader_val = None
checkpoint_period = cfg.SOLVER.CHECKPOINT_PERIOD
do_train(
cfg,
model,
data_loader,
data_loader_val,
optimizer,
scheduler,
checkpointer,
checkpoint_period,
test_period,
arguments,
)
return model
def run_inference(config_file):
import jittor as jt
from jittor_utils import auto_diff
from detectron.config import cfg
from detectron.modeling.detector import build_detection_model
from detectron.utils.checkpoint import DetectronCheckpointer
from detectron.data import make_data_loader
from detectron.engine.inference import inference
from detectron.utils.logger import setup_logger
jt.flags.use_cuda = 1
jt.cudnn.set_algorithm_cache_size(0)
cfg.merge_from_file(config_file)
cfg.freeze()
save_dir = ""
logger = setup_logger("maskrcnn_benchmark", save_dir)
model = build_detection_model(cfg)
# hook = auto_diff.Hook('fasterrcnn')
# hook.hook_module(model)
output_dir = cfg.OUTPUT_DIR
checkpointer = DetectronCheckpointer(cfg, model, save_dir=output_dir)
_ = checkpointer.load(cfg.MODEL.WEIGHT)
iou_types = ("bbox", )
if cfg.MODEL.MASK_ON:
iou_types = iou_types + ("segm", )
if cfg.MODEL.KEYPOINT_ON:
iou_types = iou_types + ("keypoints", )
dataset_names = cfg.DATASETS.TEST
data_loaders_val = make_data_loader(cfg, is_train=False)
for dataset_name, data_loader_val in zip(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.EMBED_MASK_ON
or cfg.MODEL.RETINANET_ON else cfg.MODEL.RPN_ONLY,
bbox_aug=cfg.TEST.BBOX_AUG.ENABLED,
expected_results=cfg.TEST.EXPECTED_RESULTS,
expected_results_sigma_tol=cfg.TEST.EXPECTED_RESULTS_SIGMA_TOL,
cfg=cfg)
def train(cfg, distributed=False):
model = build_detection_model(cfg)
device = torch.device(cfg.MODEL.DEVICE)
model.to(device)
optimizer = make_optimizer(cfg, model)
scheduler = make_lr_scheduler(cfg, optimizer)
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,
start_iter=arguments["iteration"],
is_distributed=distributed,
)
checkpoint_period = cfg.SOLVER.CHECKPOINT_PERIOD
do_train(
model,
data_loader,
optimizer,
scheduler,
checkpointer,
device,
checkpoint_period,
arguments,
)
return model
def run_torch_inference(config_file):
import jittor as jt
from jittor_utils import auto_diff
from maskrcnn_benchmark.config import cfg
from maskrcnn_benchmark.modeling.detector import build_detection_model
from maskrcnn_benchmark.utils.checkpoint import DetectronCheckpointer
from maskrcnn_benchmark.data import make_data_loader
from maskrcnn_benchmark.engine.inference import inference
from maskrcnn_benchmark.utils.logger import setup_logger
cfg.merge_from_file(config_file)
cfg.freeze()
save_dir = ""
model = build_detection_model(cfg)
model = model.cuda()
# hook = auto_diff.Hook('fasterrcnn')
# hook.hook_module(model)
output_dir = cfg.OUTPUT_DIR
checkpointer = DetectronCheckpointer(cfg, model, save_dir=output_dir)
_ = checkpointer.load(cfg.MODEL.WEIGHT)
iou_types = ("bbox", )
if cfg.MODEL.MASK_ON:
iou_types = iou_types + ("segm", )
dataset_names = cfg.DATASETS.TEST
data_loaders_val = make_data_loader(cfg, is_train=False)
for dataset_name, data_loader_val in zip(dataset_names, data_loaders_val):
inference(
model,
data_loader_val,
'coco_2014_minival',
iou_types=iou_types,
box_only=cfg.MODEL.RPN_ONLY,
expected_results=cfg.TEST.EXPECTED_RESULTS,
expected_results_sigma_tol=cfg.TEST.EXPECTED_RESULTS_SIGMA_TOL,
)
def do_train(
cfg,
model,
data_loader,
data_loader_val,
optimizer,
scheduler,
checkpointer,
checkpoint_period,
test_period,
arguments,
):
logger = logging.getLogger("detectron.trainer")
logger.info("Start training")
meters = MetricLogger(delimiter=" ")
max_iter = len(data_loader)
start_iter = arguments["iteration"]
model.train()
start_training_time = time.time()
end = time.time()
iou_types = ("bbox",)
if cfg.MODEL.MASK_ON:
iou_types = iou_types + ("segm",)
if cfg.MODEL.KEYPOINT_ON:
iou_types = iou_types + ("keypoints",)
dataset_names = cfg.DATASETS.TEST
for iteration, (images, targets, _) in enumerate(data_loader, start_iter):
if any(len(target) < 1 for target in targets):
logger.error(f"Iteration={iteration + 1} || Image Ids used for training {_} || targets Length={[len(target) for target in targets]}" )
continue
data_time = time.time() - end
iteration = iteration + 1
arguments["iteration"] = iteration
loss_dict = model(images, targets)
losses = sum(loss for loss in loss_dict.values())
# reduce losses over all GPUs for logging purposes
loss_dict_reduced = loss_dict
losses_reduced = sum(loss for loss in loss_dict_reduced.values())
meters.update(loss=losses_reduced, **loss_dict_reduced)
# Note: If mixed precision is not used, this ends up doing nothing
# Otherwise apply loss scaling for mixed-precision recipe
optimizer.step(losses)
scheduler.step()
batch_time = time.time() - end
end = time.time()
meters.update(time=batch_time, data=data_time)
eta_seconds = meters.time.global_avg * (max_iter - iteration)
eta_string = str(datetime.timedelta(seconds=int(eta_seconds)))
if iteration % 20 == 0 or iteration == max_iter:
logger.info(
meters.delimiter.join(
[
"eta: {eta}",
"iter: {iter}",
"{meters}",
"lr: {lr:.6f}",
"max mem: {memory:.0f}",
]
).format(
eta=eta_string,
iter=iteration,
meters=str(meters),
lr=optimizer.param_groups[0]["lr"],
memory=1024 / 1024.0 / 1024.0, # TODO CUDA Memory
)
)
if iteration % checkpoint_period == 0:
checkpointer.save("model_{:07d}".format(iteration), **arguments)
if data_loader_val is not None and test_period > 0 and iteration % test_period == 0:
meters_val = MetricLogger(delimiter=" ")
_ = inference( # The result can be used for additional logging, e. g. for TensorBoard
model,
# The method changes the segmentation mask format in a data loader,
# so every time a new data loader is created:
make_data_loader(cfg, is_train=False, is_distributed=False, is_for_period=True),
dataset_name="[Validation]",
iou_types=iou_types,
box_only=False if 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=None,
)
model.train()
with jt.no_grad():
# Should be one image for each GPU:
for iteration_val, (images_val, targets_val, _) in enumerate(tqdm(data_loader_val)):
loss_dict = model(images_val, targets_val)
losses = sum(loss for loss in loss_dict.values())
loss_dict_reduced = loss_dict
losses_reduced = sum(loss for loss in loss_dict_reduced.values())
meters_val.update(loss=losses_reduced, **loss_dict_reduced)
logger.info(
meters_val.delimiter.join(
[
"[Validation]: ",
"eta: {eta}",
"iter: {iter}",
"{meters}",
"lr: {lr:.6f}",
"max mem: {memory:.0f}",
]
).format(
eta=eta_string,
iter=iteration,
meters=str(meters_val),
lr=optimizer.param_groups[0]["lr"],
memory= 2014 / 1024.0 / 1024.0,# TODO torch.cuda.max_memory_allocated()
)
)
if iteration == max_iter:
checkpointer.save("model_final", **arguments)
total_training_time = time.time() - start_training_time
total_time_str = str(datetime.timedelta(seconds=total_training_time))
logger.info(
"Total training time: {} ({:.4f} s / it)".format(
total_time_str, total_training_time / (max_iter)
)
)
def main():
jt.flags.use_cuda = 1
parent_path = os.path.abspath(__file__).split("/tools/")[0]
parser = argparse.ArgumentParser(description="Object Detection Inference")
parser.add_argument(
"--config-file",
default=
f"{parent_path}/configs/maskrcnn_benchmark/e2e_mask_rcnn_R_50_FPN_1x.yaml",
metavar="FILE",
help="path to config file",
)
parser.add_argument(
"--ckpt",
help=
"The path to the checkpoint for test, default is the latest checkpoint.",
default=None,
)
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 = ""
logger = setup_logger("detectron", save_dir)
logger.info("Using {} GPUs".format(1))
logger.info(cfg)
logger.info("Collecting env info (might take some time)")
logger.info("\n" + collect_env_info())
model = build_detection_model(cfg)
output_dir = cfg.OUTPUT_DIR
checkpointer = DetectronCheckpointer(cfg, model, save_dir=output_dir)
ckpt = cfg.MODEL.WEIGHT if args.ckpt is None else args.ckpt
_ = checkpointer.load(ckpt, use_latest=args.ckpt is None)
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)
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.RETINANET_ON else cfg.MODEL.RPN_ONLY,
bbox_aug=cfg.TEST.BBOX_AUG.ENABLED,
expected_results=cfg.TEST.EXPECTED_RESULTS,
expected_results_sigma_tol=cfg.TEST.EXPECTED_RESULTS_SIGMA_TOL,
output_folder=output_folder,
)
def main(add_eval_flag=False):
parser = argparse.ArgumentParser(
description="PyTorch Object Detection Inference")
parser.add_argument(
"--config-file",
default=cfg_file,
metavar="FILE",
help="path to config file",
)
parser.add_argument("--local_rank",
type=int,
default=int(os.environ['CUDA_VISIBLE_DIVICES']))
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
args.distributed = num_gpus > 1
# if args.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("detectron", save_dir, get_rank())
logger.info("Using {} GPUs".format(num_gpus))
logger.info(cfg)
model = build_detection_model(cfg)
model.to(cfg.MODEL.DEVICE)
output_dir = cfg.OUTPUT_DIR
checkpointer = DetectronCheckpointer(cfg, model, save_dir=output_dir)
_ = checkpointer.load(cfg.MODEL.WEIGHT)
iou_types = ("bbox", )
if cfg.MODEL.MASK_ON:
iou_types = iou_types + ("segm", )
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=args.distributed)
for output_folder, dataset_name, data_loader_val in zip(
output_folders, dataset_names, data_loaders_val):
coco_results, _ = \
inference(
model,
data_loader_val,
dataset_name=dataset_name,
iou_types=iou_types,
box_only=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 add_eval_fields():
ar = coco_results.results['bbox']['AR50']
ap = coco_results.results['bbox']['AP50']
checkpoint_file = checkpointer.get_checkpoint_file()
base_checkpoint_file = os.path.basename(checkpoint_file).split('.')[0]
new_checkpoint_file = os.path.join(
output_dir,
base_checkpoint_file + '_ar{:.03}_ap_{:.03}.pth'.format(ar, ap))
os.rename(checkpoint_file, new_checkpoint_file)
if add_eval_flag:
add_eval_fields()