configs/mask2former/mask2former_r50_8xb2-90k_cityscapes-512x1024.py

_base_ = ['../_base_/default_runtime.py', '../_base_/datasets/cityscapes.py']

crop_size = (512, 1024)
data_preprocessor = dict(
    type='SegDataPreProcessor',
    mean=[123.675, 116.28, 103.53],
    std=[58.395, 57.12, 57.375],
    bgr_to_rgb=True,
    pad_val=0,
    seg_pad_val=255,
    size=crop_size,
    test_cfg=dict(size_divisor=32))
num_classes = 19
model = dict(
    type='EncoderDecoder',
    data_preprocessor=data_preprocessor,
    backbone=dict(
        type='ResNet',
        depth=50,
        deep_stem=False,
        num_stages=4,
        out_indices=(0, 1, 2, 3),
        frozen_stages=-1,
        norm_cfg=dict(type='SyncBN', requires_grad=False),
        style='pytorch',
        init_cfg=dict(type='Pretrained', checkpoint='torchvision://resnet50')),
    decode_head=dict(
        type='Mask2FormerHead',
        in_channels=[256, 512, 1024, 2048],
        strides=[4, 8, 16, 32],
        feat_channels=256,
        out_channels=256,
        num_classes=num_classes,
        num_queries=100,
        num_transformer_feat_level=3,
        align_corners=False,
        pixel_decoder=dict(
            type='mmdet.MSDeformAttnPixelDecoder',
            num_outs=3,
            norm_cfg=dict(type='GN', num_groups=32),
            act_cfg=dict(type='ReLU'),
            encoder=dict(  # DeformableDetrTransformerEncoder
                num_layers=6,
                layer_cfg=dict(  # DeformableDetrTransformerEncoderLayer
                    self_attn_cfg=dict(  # MultiScaleDeformableAttention
                        embed_dims=256,
                        num_heads=8,
                        num_levels=3,
                        num_points=4,
                        im2col_step=64,
                        dropout=0.0,
                        batch_first=True,
                        norm_cfg=None,
                        init_cfg=None),
                    ffn_cfg=dict(
                        embed_dims=256,
                        feedforward_channels=1024,
                        num_fcs=2,
                        ffn_drop=0.0,
                        act_cfg=dict(type='ReLU', inplace=True))),
                init_cfg=None),
            positional_encoding=dict(  # SinePositionalEncoding
                num_feats=128, normalize=True),
            init_cfg=None),
        enforce_decoder_input_project=False,
        positional_encoding=dict(  # SinePositionalEncoding
            num_feats=128, normalize=True),
        transformer_decoder=dict(  # Mask2FormerTransformerDecoder
            return_intermediate=True,
            num_layers=9,
            layer_cfg=dict(  # Mask2FormerTransformerDecoderLayer
                self_attn_cfg=dict(  # MultiheadAttention
                    embed_dims=256,
                    num_heads=8,
                    attn_drop=0.0,
                    proj_drop=0.0,
                    dropout_layer=None,
                    batch_first=True),
                cross_attn_cfg=dict(  # MultiheadAttention
                    embed_dims=256,
                    num_heads=8,
                    attn_drop=0.0,
                    proj_drop=0.0,
                    dropout_layer=None,
                    batch_first=True),
                ffn_cfg=dict(
                    embed_dims=256,
                    feedforward_channels=2048,
                    num_fcs=2,
                    act_cfg=dict(type='ReLU', inplace=True),
                    ffn_drop=0.0,
                    dropout_layer=None,
                    add_identity=True)),
            init_cfg=None),
        loss_cls=dict(
            type='mmdet.CrossEntropyLoss',
            use_sigmoid=False,
            loss_weight=2.0,
            reduction='mean',
            class_weight=[1.0] * num_classes + [0.1]),
        loss_mask=dict(
            type='mmdet.CrossEntropyLoss',
            use_sigmoid=True,
            reduction='mean',
            loss_weight=5.0),
        loss_dice=dict(
            type='mmdet.DiceLoss',
            use_sigmoid=True,
            activate=True,
            reduction='mean',
            naive_dice=True,
            eps=1.0,
            loss_weight=5.0),
        train_cfg=dict(
            num_points=12544,
            oversample_ratio=3.0,
            importance_sample_ratio=0.75,
            assigner=dict(
                type='mmdet.HungarianAssigner',
                match_costs=[
                    dict(type='mmdet.ClassificationCost', weight=2.0),
                    dict(
                        type='mmdet.CrossEntropyLossCost',
                        weight=5.0,
                        use_sigmoid=True),
                    dict(
                        type='mmdet.DiceCost',
                        weight=5.0,
                        pred_act=True,
                        eps=1.0)
                ]),
            sampler=dict(type='mmdet.MaskPseudoSampler'))),
    train_cfg=dict(),
    test_cfg=dict(mode='whole'))

# dataset config
train_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(type='LoadAnnotations'),
    dict(
        type='RandomChoiceResize',
        scales=[int(1024 * x * 0.1) for x in range(5, 21)],
        resize_type='ResizeShortestEdge',
        max_size=4096),
    dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
    dict(type='RandomFlip', prob=0.5),
    dict(type='PhotoMetricDistortion'),
    dict(type='PackSegInputs')
]
train_dataloader = dict(dataset=dict(pipeline=train_pipeline))

# optimizer
embed_multi = dict(lr_mult=1.0, decay_mult=0.0)
optimizer = dict(
    type='AdamW', lr=0.0001, weight_decay=0.05, eps=1e-8, betas=(0.9, 0.999))
optim_wrapper = dict(
    type='OptimWrapper',
    optimizer=optimizer,
    clip_grad=dict(max_norm=0.01, norm_type=2),
    paramwise_cfg=dict(
        custom_keys={
            'backbone': dict(lr_mult=0.1, decay_mult=1.0),
            'query_embed': embed_multi,
            'query_feat': embed_multi,
            'level_embed': embed_multi,
        },
        norm_decay_mult=0.0))
# learning policy
param_scheduler = [
    dict(
        type='PolyLR',
        eta_min=0,
        power=0.9,
        begin=0,
        end=90000,
        by_epoch=False)
]

# training schedule for 90k
train_cfg = dict(type='IterBasedTrainLoop', max_iters=90000, val_interval=5000)
val_cfg = dict(type='ValLoop')
test_cfg = dict(type='TestLoop')
default_hooks = dict(
    timer=dict(type='IterTimerHook'),
    logger=dict(type='LoggerHook', interval=50, log_metric_by_epoch=False),
    param_scheduler=dict(type='ParamSchedulerHook'),
    checkpoint=dict(
        type='CheckpointHook', by_epoch=False, interval=5000,
        save_best='mIoU'),
    sampler_seed=dict(type='DistSamplerSeedHook'),
    visualization=dict(type='SegVisualizationHook'))

# Default setting for scaling LR automatically
#   - `enable` means enable scaling LR automatically
#       or not by default.
#   - `base_batch_size` = (8 GPUs) x (2 samples per GPU).
auto_scale_lr = dict(enable=False, base_batch_size=16)