69a8f96
1
2
3
4
5
6
7
8
9
10
11
# model settings _base_ = ['../backbones/convnext_large.py',] model = dict( type='DensePredModel', decode_head=dict( type='HourglassDecoder', in_channels=[192, 384, 768, 1536], decoder_channel=[128, 128, 256, 512], prefix='decode_heads.'), )
