def build_hooks(self) -> List[HookBase]:
"""
This method overwrites the default one from DefaultTrainer.
It adds the `LossEvalHook` that allows evaluating results on the validation set.
The default method builds a list of default hooks, including timing, evaluation,
checkpointing, lr scheduling, precise BN, writing events.
Returns:
list[HookBase]: The list of hooks to call during training.
"""
hooks: List[HookBase] = super().build_hooks()
# TODO remove as it can't work
# input_example = next(iter(self.data_loader))
# hooks.append(ModelWriter(model=self.model,
# input_example=input_example,
# log_dir=self.cfg.OUTPUT_DIR))
# We add our custom validation hook
if self.cfg.DATASETS.VALIDATION != "":
data_set_mapper: PanelSegDatasetMapper = PanelSegDatasetMapper(
cfg=self.cfg, is_train=True)
data_loader: DataLoader = build_detection_test_loader(
cfg=self.cfg,
dataset_name=self.cfg.DATASETS.VALIDATION,
mapper=data_set_mapper)
loss_eval_hook: LossEvalHook = LossEvalHook(
eval_period=self.cfg.VALIDATION.VALIDATION_PERIOD,
model=self.model,
data_loader=data_loader)
hooks.insert(-1, loss_eval_hook)
return hooks
def build_hooks(self) -> List[HookBase]:
"""
This method overwrites the default one from DefaultTrainer.
It adds (if necessary) the `LossEvalHook` that allows evaluating the loss on the
validation set.
Returns:
List[HookBase]: The augmented list of hooks.
"""
# Build a list of default hooks, including timing, evaluation,
# checkpointing, lr scheduling, precise BN, writing events.
hooks = super().build_hooks()
# We add our custom validation hook
if self.cfg.DATASETS.VALIDATION != "":
data_set_mapper: DatasetMapper = DatasetMapper.from_config(
cfg=self.cfg, is_train=True)
data_loader: DataLoader = build_detection_test_loader(
cfg=self.cfg,
dataset_name=self.cfg.DATASETS.VALIDATION,
mapper=data_set_mapper)
loss_eval_hook: LossEvalHook = LossEvalHook(
eval_period=self.cfg.VALIDATION.VALIDATION_PERIOD,
model=self.model,
data_loader=data_loader)
hooks.insert(index=-1, obj=loss_eval_hook)
return hooks
def build_test_loader(cls, cfg, dataset_name):
"""
It now calls :func:`detectron2.data.build_detection_test_loader`.
Overwrite it if you'd like a different data loader.
"""
mapr = cls.build_mapper(cfg, is_train=False)
return build_detection_test_loader(cfg, dataset_name, mapper=None)
def build_hooks(self):
hooks = super().build_hooks()
if self.cfg.ISPRS.LABEL.BOXMODE == "ROTATED":
eval_mapper = ISPRSCOCOStyleMapperRotated
else:
eval_mapper = ISPRSCOCOStyleMapperAxisAligned
hooks.insert(
-1,
LossEvalHook(
self.cfg.TEST.EVAL_PERIOD, self._trainer.model,
build_detection_test_loader(self.cfg,
self.cfg.DATASETS.TEST[0],
eval_mapper(self.cfg, True))))
return hooks
def build_test_loader(cls, cfg: CfgNode, dataset_name: str) -> DataLoader:
"""
Instanciate the test data loader.
Args:
cfg (CfgNode): The global config.
dataset_name (str): The name of the test dataset.
Returns:
a DataLoader yielding formatted test examples.
"""
mapper: PanelSegDatasetMapper = PanelSegDatasetMapper(cfg,
is_train=False)
return build_detection_test_loader(cfg,
dataset_name=dataset_name,
mapper=mapper)
def build_test_loader(cls, cfg, dataset_name):
return build_detection_test_loader(cfg,
dataset_name,
mapper=DatasetMapper(cfg, False))
def custom_test_loader(cfg, dataset_name):
return build_detection_test_loader(cfg,
dataset_name,
mapper=CustomDatasetMapper(cfg, False))
def get_lvis_test_dataloader(cfg, h, w):
default_mapper = DatasetMapper(cfg, is_train=False)
mapper = partial(wrapper, default_m=default_mapper, h=h, w=w)
dl = build_detection_test_loader(cfg, 'lvis_v0.5_val', mapper=mapper)
return dl
trainer.resume_or_load(resume=False)
trainer.train()
# Look at training curves in tensorboard:
# %load_ext tensorboard
# %tensorboard --logdir output/run_rdd/
else:
# Inference & evaluation using the trained model
# Now, let's run inference with the trained model on the validation dataset.
# First, let's create a predictor using the model we just trained:
cfg.MODEL.WEIGHTS = os.path.join(cfg.OUTPUT_DIR, "model_final.pth")
cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = 0.7 # set a custom testing threshold for this model
cfg.DATASETS.TEST = ("rdd2020_val", )
predictor = DefaultPredictor(cfg)
trainer.resume_or_load(resume=True)
# Then, we randomly select several samples to visualize the prediction results.
from detectron2.utils.visualizer import ColorMode
from detectron2.evaluation import COCOEvaluator, DatasetEvaluators, inference_on_dataset
from detectron2.data import build_detection_test_loader
evaluator = COCOEvaluator("rdd2020_val", cfg, False, "coco_eval")
val_loader = build_detection_test_loader(cfg, "rdd2020_val")
eval_results = inference_on_dataset(trainer.model, val_loader,
DatasetEvaluators([evaluator]))
# another equivalent way is to use trainer.test
print(eval_results)
# Empty the GPU Memory
torch.cuda.empty_cache()