TOPK = (1, 2)
NUM_CLASSES = 6
CLASSES = ['0-11', '12-24', '25-29', '30-40', '41-64', '65+']
dataset_type = 'CustomDataset'
data_preprocessor = dict(
    num_classes=6,
    mean=[122.7709383, 116.7460125, 104.09373615000001],
    std=[68.5005327, 66.6321579, 70.32316304999999],
    to_rgb=True)
train_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(
        type='RandomResizedCrop',
        scale=448,
        backend='pillow',
        interpolation='bicubic'),
    dict(type='RandomFlip', prob=0.5, direction='horizontal'),
    dict(type='PackInputs')
]
test_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(
        _scope_='mmpretrain',
        type='ResizeEdge',
        scale=448,
        edge='short',
        backend='pillow',
        interpolation='bicubic'),
    dict(type='CenterCrop', crop_size=448),
    dict(type='PackInputs')
]
train_dataloader = dict(
    pin_memory=True,
    persistent_workers=True,
    collate_fn=dict(type='default_collate'),
    batch_size=10,
    num_workers=5,
    dataset=dict(
        type='CustomDataset',
        data_root='../data-raw/external/UTKFace/',
        ann_file='train/train_age.txt',
        data_prefix='train',
        pipeline=[
            dict(type='LoadImageFromFile'),
            dict(
                type='RandomResizedCrop',
                scale=448,
                backend='pillow',
                interpolation='bicubic'),
            dict(type='RandomFlip', prob=0.5, direction='horizontal'),
            dict(type='PackInputs')
        ],
        with_label=True,
        classes=['0-11', '12-24', '25-29', '30-40', '41-64', '65+']),
    sampler=dict(type='DefaultSampler', shuffle=True))
val_dataloader = dict(
    pin_memory=True,
    persistent_workers=True,
    collate_fn=dict(type='default_collate'),
    batch_size=10,
    num_workers=5,
    dataset=dict(
        type='CustomDataset',
        data_root='../data-raw/external/UTKFace/',
        ann_file='val/val_age.txt',
        data_prefix='val',
        pipeline=[
            dict(type='LoadImageFromFile'),
            dict(
                _scope_='mmpretrain',
                type='ResizeEdge',
                scale=448,
                edge='short',
                backend='pillow',
                interpolation='bicubic'),
            dict(type='CenterCrop', crop_size=448),
            dict(type='PackInputs')
        ],
        with_label=True,
        classes=['0-11', '12-24', '25-29', '30-40', '41-64', '65+']),
    sampler=dict(type='DefaultSampler', shuffle=False))
val_evaluator = dict(type='Accuracy', topk=(1, 2))
test_dataloader = dict(
    pin_memory=True,
    persistent_workers=True,
    collate_fn=dict(type='default_collate'),
    batch_size=10,
    num_workers=5,
    dataset=dict(
        type='CustomDataset',
        data_root='../data-raw/external/UTKFace/',
        ann_file='test/test_age.txt',
        data_prefix='test',
        pipeline=[
            dict(type='LoadImageFromFile'),
            dict(
                _scope_='mmpretrain',
                type='ResizeEdge',
                scale=448,
                edge='short',
                backend='pillow',
                interpolation='bicubic'),
            dict(type='CenterCrop', crop_size=448),
            dict(type='PackInputs')
        ],
        with_label=True,
        classes=['0-11', '12-24', '25-29', '30-40', '41-64', '65+']),
    sampler=dict(type='DefaultSampler', shuffle=False))
test_evaluator = dict(type='Accuracy', topk=(1, 2))
optim_wrapper = dict(
    optimizer=dict(type='AdamW', lr=1.5e-07, weight_decay=0.3),
    paramwise_cfg=dict(
        custom_keys=dict({
            '.cls_token': dict(decay_mult=0.0),
            '.pos_embed': dict(decay_mult=0.0)
        })))
warmup_epochs = 15
param_scheduler = [
    dict(
        type='LinearLR',
        start_factor=0.001,
        by_epoch=True,
        end=15,
        convert_to_iter_based=True),
    dict(type='CosineAnnealingLR', eta_min=1e-05, by_epoch=True, begin=15),
    dict(
        type='MultiStepLR',
        by_epoch=True,
        begin=0,
        end=41,
        milestones=[15, 40],
        gamma=0.1)
]
train_cfg = dict(by_epoch=True, max_epochs=100, val_interval=1)
val_cfg = dict()
test_cfg = dict()
auto_scale_lr = dict(base_batch_size=2048)
default_scope = 'mmpretrain'
default_hooks = dict(
    timer=dict(type='IterTimerHook'),
    logger=dict(type='LoggerHook', interval=100),
    param_scheduler=dict(type='ParamSchedulerHook'),
    checkpoint=dict(type='CheckpointHook', interval=1),
    sampler_seed=dict(type='DistSamplerSeedHook'),
    visualization=dict(type='VisualizationHook', enable=False))
custom_hooks = [
    dict(
        type='EarlyStoppingHook',
        monitor='accuracy/top1',
        patience=100,
        stopping_threshold=85.0
        
    )]
env_cfg = dict(
    cudnn_benchmark=False,
    mp_cfg=dict(mp_start_method='fork', opencv_num_threads=0),
    dist_cfg=dict(backend='nccl'))
vis_backends = [dict(type='LocalVisBackend')]
visualizer = dict(
    type='UniversalVisualizer', 
    vis_backends=[
        dict(type='TensorboardVisBackend'), 
        dict(
            type='WandbVisBackend',
            init_kwargs=dict(project='age_estimation'),
            define_metric_cfg=[
                dict(name='accuracy/top1', step_metric='epoch'),
                dict(name='accuracy/top2', step_metric='epoch'),
            ]
        )])
log_level = 'INFO'
load_from = None
resume = False
randomness = dict(seed=None, deterministic=False)
model = dict(
    type='ImageClassifier',
    backbone=dict(
        type='ViTEVA02',
        arch='b',
        img_size=448,
        patch_size=14,
        sub_ln=True,
        final_norm=False,
        out_type='avg_featmap',
        init_cfg=dict(
            type='Pretrained',
            checkpoint=
            '../pre-trained-models/eva02-base-p14_in21k-pre_in21k-medft_3rdparty_in1k-448px_20230505-5cd4d87f.pth',
            prefix='backbone')),
    neck=None,
    head=dict(
        type='LinearClsHead',
        num_classes=6,
        in_channels=768,
        loss=dict(
            type='LabelSmoothLoss', label_smooth_val=0.1, mode='original')),
    init_cfg=[
        dict(type='TruncNormal', layer='Linear', std=0.02),
        dict(type='Constant', layer='LayerNorm', val=1.0, bias=0.0)
    ],
    train_cfg=dict(augments=[
        dict(type='Mixup', alpha=0.8),
        dict(type='CutMix', alpha=1.0)
    ]))
launcher = 'pytorch'
work_dir = './work_dirs/eva02-base-p14_in1k'