configs/segformer/segformer_mit-b5_8xb2-160k_ade20k-640x640.py

_base_ = ['./segformer_mit-b0_8xb2-160k_ade20k-512x512.py']

checkpoint = 'https://download.openmmlab.com/mmsegmentation/v0.5/pretrain/segformer/mit_b5_20220624-658746d9.pth'  # noqa

# dataset settings
crop_size = (640, 640)
data_preprocessor = dict(size=crop_size)
train_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='LoadAnnotations', reduce_zero_label=True),
    dict(
        type='RandomResize',
        scale=(2048, 640),
        ratio_range=(0.5, 2.0),
        keep_ratio=True),
    dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
    dict(type='RandomFlip', prob=0.5),
    dict(type='PhotoMetricDistortion'),
    dict(type='PackSegInputs')
]
test_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='Resize', scale=(2048, 640), keep_ratio=True),
    # add loading annotation after ``Resize`` because ground truth
    # does not need to do resize data transform
    dict(type='LoadAnnotations', reduce_zero_label=True),
    dict(type='PackSegInputs')
]
train_dataloader = dict(dataset=dict(pipeline=train_pipeline))
val_dataloader = dict(batch_size=1, dataset=dict(pipeline=test_pipeline))
test_dataloader = val_dataloader

# model settings
model = dict(
    data_preprocessor=data_preprocessor,
    backbone=dict(
        init_cfg=dict(type='Pretrained', checkpoint=checkpoint),
        embed_dims=64,
        num_heads=[1, 2, 5, 8],
        num_layers=[3, 6, 40, 3]),
    decode_head=dict(in_channels=[64, 128, 320, 512]))