dataset_type = "CustomDataset"
img_dir = "/media/VA/databases/{{cookiecutter.dataset}}/images/"
ann_dir = "results/data/transform/coco_to_mmsegmentation-{{cookiecutter.dataset}}/masks/"
split_dir = "results/data/transform/coco_to_mmsegmentation-{{cookiecutter.dataset}}/"
CLASSES = ["background"]
img_scale = (512, 512)
keep_ratio = False
img_norm_cfg = dict(
mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True
)
train_pipeline = [
dict(type="LoadImageFromFile"),
dict(type="LoadAnnotations"),
dict(type="Resize", img_scale=img_scale, keep_ratio=keep_ratio),
dict(type="Normalize", **img_norm_cfg),
dict(type="DefaultFormatBundle"),
dict(type="Collect", keys=["img", "gt_semantic_seg"]),
]
test_pipeline = [
dict(type="LoadImageFromFile"),
dict(
type="MultiScaleFlipAug",
img_scale=img_scale,
flip=False,
transforms=[
dict(type="Resize", keep_ratio=keep_ratio),
dict(type="Normalize", **img_norm_cfg),
dict(type="ImageToTensor", keys=["img"]),
dict(type="Collect", keys=["img"]),
],
),
]
data = dict(
samples_per_gpu=4,
workers_per_gpu=2,
train=dict(
type=dataset_type,
img_dir=img_dir,
ann_dir=ann_dir + "/train",
split=split_dir + "/{{cookiecutter.dataset}}_train.txt",
pipeline=train_pipeline,
classes=CLASSES,
),
val=dict(
type=dataset_type,
img_dir=img_dir,
ann_dir=ann_dir + "/val",
split=split_dir + "/{{cookiecutter.dataset}}_val.txt",
pipeline=test_pipeline,
classes=CLASSES,
),
test=dict(
type=dataset_type,
img_dir=img_dir,
ann_dir=ann_dir + "/val",
split=split_dir + "/{{cookiecutter.dataset}}_val.txt",
pipeline=test_pipeline,
classes=CLASSES,
),
)