configs/textrecog/nrtr/_base_nrtr_resnet31.py

dictionary = dict(
    type='Dictionary',
    dict_file='{{ fileDirname }}/../../../dicts/english_digits_symbols.txt',
    with_padding=True,
    with_unknown=True,
    same_start_end=True,
    with_start=True,
    with_end=True)

model = dict(
    type='NRTR',
    backbone=dict(
        type='ResNet31OCR',
        layers=[1, 2, 5, 3],
        channels=[32, 64, 128, 256, 512, 512],
        stage4_pool_cfg=dict(kernel_size=(2, 1), stride=(2, 1)),
        last_stage_pool=True),
    encoder=dict(type='NRTREncoder'),
    decoder=dict(
        type='NRTRDecoder',
        module_loss=dict(
            type='CEModuleLoss', ignore_first_char=True, flatten=True),
        postprocessor=dict(type='AttentionPostprocessor'),
        dictionary=dictionary,
        max_seq_len=30,
    ),
    data_preprocessor=dict(
        type='TextRecogDataPreprocessor',
        mean=[123.675, 116.28, 103.53],
        std=[58.395, 57.12, 57.375]))

train_pipeline = [
    dict(type='LoadImageFromFile', ignore_empty=True, min_size=0),
    dict(type='LoadOCRAnnotations', with_text=True),
    dict(
        type='RescaleToHeight',
        height=32,
        min_width=32,
        max_width=160,
        width_divisor=4),
    dict(type='PadToWidth', width=160),
    dict(
        type='PackTextRecogInputs',
        meta_keys=('img_path', 'ori_shape', 'img_shape', 'valid_ratio'))
]

test_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(
        type='RescaleToHeight',
        height=32,
        min_width=32,
        max_width=160,
        width_divisor=16),
    dict(type='PadToWidth', width=160),
    # add loading annotation after ``Resize`` because ground truth
    # does not need to do resize data transform
    dict(type='LoadOCRAnnotations', with_text=True),
    dict(
        type='PackTextRecogInputs',
        meta_keys=('img_path', 'ori_shape', 'img_shape', 'valid_ratio'))
]

tta_pipeline = [
    dict(type='LoadImageFromFile'),
    dict(
        type='TestTimeAug',
        transforms=[
            [
                dict(
                    type='ConditionApply',
                    true_transforms=[
                        dict(
                            type='ImgAugWrapper',
                            args=[dict(cls='Rot90', k=0, keep_size=False)])
                    ],
                    condition="results['img_shape'][1]<results['img_shape'][0]"
                ),
                dict(
                    type='ConditionApply',
                    true_transforms=[
                        dict(
                            type='ImgAugWrapper',
                            args=[dict(cls='Rot90', k=1, keep_size=False)])
                    ],
                    condition="results['img_shape'][1]<results['img_shape'][0]"
                ),
                dict(
                    type='ConditionApply',
                    true_transforms=[
                        dict(
                            type='ImgAugWrapper',
                            args=[dict(cls='Rot90', k=3, keep_size=False)])
                    ],
                    condition="results['img_shape'][1]<results['img_shape'][0]"
                ),
            ],
            [
                dict(
                    type='RescaleToHeight',
                    height=32,
                    min_width=32,
                    max_width=160,
                    width_divisor=16)
            ],
            [dict(type='PadToWidth', width=160)],
            # add loading annotation after ``Resize`` because ground truth
            # does not need to do resize data transform
            [dict(type='LoadOCRAnnotations', with_text=True)],
            [
                dict(
                    type='PackTextRecogInputs',
                    meta_keys=('img_path', 'ori_shape', 'img_shape',
                               'valid_ratio'))
            ]
        ])
]