imagenet_swin_tiny_patch4_window7_224_jpg_torchacc.py

_base_ = 'configs/base.py'

# The bigger the log interval, the better, it will reduce the training speed
log_config = dict(
    interval=200,
    hooks=[
        dict(type='TextLoggerHook'),
        # dict(type='TensorboardLoggerHook')
    ])

# model settings
model = dict(
    type='Classification',
    pretrained=None,
    backbone=dict(
        type='PytorchImageModelWrapper',
        model_name='swin_tiny_patch4_window7_224',
        num_classes=1000,
    ),
    head=dict(
        type='ClsHead',
        loss_config={
            'type': 'SoftTargetCrossEntropy',
        },
        with_fc=False))

data_root = 'data/imagenet_raw/'
data_train_list = data_root + 'meta/train_labeled.txt'
data_train_root = data_root + 'train/'
data_test_list = data_root + 'meta/val_labeled.txt'
data_test_root = data_root + 'val/'
data_all_list = data_root + 'meta/all_labeled.txt'

dataset_type = 'ClsDataset'
img_norm_cfg = dict(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
train_pipeline = [
    dict(type='RandomResizedCrop', size=224),
    dict(type='RandomHorizontalFlip'),
    dict(
        type='MMRandAugment',
        num_policies=2,
        total_level=10,
        magnitude_level=9,
        magnitude_std=0.5,
        hparams=dict(
            pad_val=[round(x * 255) for x in img_norm_cfg['mean'][::-1]],
            interpolation='bicubic')),
    dict(
        type='MMRandomErasing',
        erase_prob=0.25,
        mode='rand',
        min_area_ratio=0.02,
        max_area_ratio=1 / 3,
        fill_color=[x * 255 for x in img_norm_cfg['mean'][::-1]],
        fill_std=[x * 255 for x in img_norm_cfg['std'][::-1]]),
    dict(type='ToTensor'),
    dict(type='Normalize', **img_norm_cfg),
    dict(type='Collect', keys=['img', 'gt_labels'])
]
test_pipeline = [
    dict(type='Resize', size=256),
    dict(type='CenterCrop', size=224),
    dict(type='ToTensor'),
    dict(type='Normalize', **img_norm_cfg),
    dict(type='Collect', keys=['img', 'gt_labels'])
]

data = dict(
    imgs_per_gpu=64,  # total 256
    workers_per_gpu=8,
    drop_last=True,
    train_collate_hooks=[
        dict(
            type='MixupCollateHook',
            mixup_alpha=0.8,
            cutmix_alpha=1.0,
            prob=0.5,
            mode='batch',
            label_smoothing=0.1,
            num_classes=1000)
    ],
    train=dict(
        type=dataset_type,
        data_source=dict(
            list_file=data_train_list,
            root=data_train_root,
            type='ClsSourceImageList'),
        pipeline=train_pipeline),
    val=dict(
        type=dataset_type,
        data_source=dict(
            list_file=data_test_list,
            root=data_test_root,
            type='ClsSourceImageList'),
        pipeline=test_pipeline))
eval_config = dict(initial=False, interval=1, gpu_collect=True)
eval_pipelines = [
    dict(
        mode='test',
        data=data['val'],
        dist_eval=True,
        evaluators=[dict(type='ClsEvaluator', topk=(1, 5))])
]

# additional hooks
custom_hooks = []

# optimizer
paramwise_options = {
    'norm': dict(weight_decay=0.),
    'bias': dict(weight_decay=0.),
    'absolute_pos_embed': dict(weight_decay=0.),
    'relative_position_bias_table': dict(weight_decay=0.)
}
optimizer = dict(
    type='AdamW',
    lr=5e-4 * 1024 / 512,
    weight_decay=0.05,
    eps=1e-8,
    betas=(0.9, 0.999),
    paramwise_options=paramwise_options)
optimizer_config = dict(grad_clip=dict(max_norm=5.0))

# learning policy
lr_config = dict(
    policy='CosineAnnealing',
    by_epoch=False,
    min_lr_ratio=1e-2,
    warmup='linear',
    warmup_ratio=1e-3,
    warmup_iters=20,
    warmup_by_epoch=True)
checkpoint_config = dict(interval=30)

# runtime settings
total_epochs = 300