38 lines
1.1 KiB
Python
38 lines
1.1 KiB
Python
# model settings
|
|
embed_dims = [32, 64, 160, 256]
|
|
norm_cfg = dict(type='SyncBN', requires_grad=True)
|
|
model = dict(
|
|
type='EncoderDecoder',
|
|
pretrained=None,
|
|
backbone=dict(
|
|
type='SimplifiedMixTransformer',
|
|
in_channels=3,
|
|
embed_dims=embed_dims,
|
|
num_stages=4,
|
|
num_layers=[2, 2, 2, 2],
|
|
num_heads=[1, 2, 5, 8],
|
|
patch_sizes=[7, 3, 3, 3],
|
|
strides=[4, 2, 2, 2],
|
|
sr_ratios=[8, 4, 2, 1],
|
|
out_indices=(0, 1, 2, 3),
|
|
mlp_ratios=[8, 8, 4, 4],
|
|
qkv_bias=True,
|
|
drop_rate=0.0,
|
|
attn_drop_rate=0.0,
|
|
drop_path_rate=0.1,
|
|
norm_cfg=norm_cfg),
|
|
decode_head=dict(
|
|
type='DESTHead',
|
|
in_channels=[32, 64, 160, 256],
|
|
in_index=[0, 1, 2, 3],
|
|
channels=32,
|
|
dropout_ratio=0.1,
|
|
num_classes=19,
|
|
norm_cfg=norm_cfg,
|
|
align_corners=False,
|
|
loss_decode=dict(
|
|
type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0)),
|
|
# model training and testing settings
|
|
train_cfg=dict(),
|
|
test_cfg=dict(mode='slide', crop_size=(1024, 1024), stride=(768, 768)))
|