gca_r34_4xb10-dimaug-200k_comp1k.py

_base_ = ['./gca_r34_4xb10-200k_comp1k.py']

experiment_name = 'gca_r34_4xb10-dimaug-200k_comp1k'
work_dir = f'./work_dirs/{experiment_name}'
save_dir = './work_dirs/'

# model settings
model = dict(backbone=dict(encoder=dict(in_channels=4)))

# dataset settings
train_pipeline = [
    dict(type='LoadImageFromFile', key='alpha', color_type='grayscale'),
    dict(type='LoadImageFromFile', key='merged'),
    dict(
        type='CropAroundUnknown',
        keys=['alpha', 'merged'],
        crop_sizes=[320, 480, 640]),
    dict(type='Flip', keys=['alpha', 'merged']),
    dict(
        type='Resize',
        keys=['alpha', 'merged'],
        scale=(320, 320),
        keep_ratio=False),
    dict(type='GenerateTrimap', kernel_size=(1, 30)),
    dict(type='FormatTrimap', to_onehot=False),
    dict(type='PackInputs'),
]

test_pipeline = [
    dict(
        type='LoadImageFromFile',
        key='alpha',
        color_type='grayscale',
        save_original_img=True),
    dict(
        type='LoadImageFromFile',
        key='trimap',
        color_type='grayscale',
        save_original_img=True),
    dict(type='LoadImageFromFile', key='merged'),
    dict(type='FormatTrimap', to_onehot=False),
    dict(type='PackInputs'),
]

train_dataloader = dict(dataset=dict(pipeline=train_pipeline))

val_dataloader = dict(dataset=dict(pipeline=test_pipeline))

test_dataloader = val_dataloader