HOW TO IMPLEMENT ALBUMENTATION ?

[deconasser]

train_pipeline = [ dict(type='LoadImageFromFile'), dict(type='LoadAnnotations', with_bbox=True), dict( type='AutoAugment', policies=[ [ dict( type='Resize', img_scale=[(480, 2048), (512, 2048), (544, 2048), (576, 2048), (608, 2048), (640, 2048), (672, 2048), (704, 2048), (736, 2048), (768, 2048), (800, 2048), (832, 2048), (864, 2048), (896, 2048), (928, 2048), (960, 2048), (992, 2048), (1024, 2048), (1056, 2048), (1088, 2048), (1120, 2048), (1152, 2048), (1184, 2048), (1216, 2048), (1248, 2048), (1280, 2048), (1312, 2048), (1344, 2048), (1376, 2048), (1408, 2048), (1440, 2048), (1472, 2048), (1504, 2048), (1536, 2048)], multiscale_mode='value', keep_ratio=True) ], [ dict( type='Resize', # The radio of all image in train dataset < 7 # follow the original impl img_scale=[(400, 4200), (500, 4200), (600, 4200)], multiscale_mode='value', keep_ratio=True), dict( type='RandomCrop', crop_type='absolute_range', crop_size=(384, 600), allow_negative_crop=True), dict( type='Resize', img_scale=[(480, 2048), (512, 2048), (544, 2048), (576, 2048), (608, 2048), (640, 2048), (672, 2048), (704, 2048), (736, 2048), (768, 2048), (800, 2048), (832, 2048), (864, 2048), (896, 2048), (928, 2048), (960, 2048), (992, 2048), (1024, 2048), (1056, 2048), (1088, 2048), (1120, 2048), (1152, 2048), (1184, 2048), (1216, 2048), (1248, 2048), (1280, 2048), (1312, 2048), (1344, 2048), (1376, 2048), (1408, 2048), (1440, 2048), (1472, 2048), (1504, 2048), (1536, 2048)], multiscale_mode='value', override=True, keep_ratio=True) ] ]), # dict(type='RemoveImgInfo'), # Insert this line dict( type='Albu', transforms=[ dict(type='GaussianBlur', blur_limit=(7, 11), sigma_limit=0, p=0.5), dict(type='ColorJitter', p=0.1), dict(type='MotionBlur', blur_limit=(11, 21), allow_shifted=True, p=0.1), dict(type='RandomBrightnessContrast', brightness_limit=0.2, contrast_limit=0.2, p=0.5), dict(type='Perspective', scale=(0.05, 0.1), p=0.1), dict(type='RandomFog', fog_coef_lower=0.2, fog_coef_upper=0.6, alpha_coef=0.1, p=0.05), ], bbox_params=dict( type='BboxParams', format='coco', # Định dạng chính xác cho COCO label_fields=['gt_labels'], min_visibility=0.1, filter_lost_elements=True), keymap={ 'img': 'image', 'gt_masks': 'masks', 'gt_bboxes': 'bboxes' }, update_pad_shape=False, skip_img_without_anno=True ), dict(type='Normalize', **img_norm_cfg), dict(type='Pad', size_divisor=32), dict(type='DefaultFormatBundle'), dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']) ] The bug is: ValueError: Key img_info is not in available keys.

[TempleX98]

Sorry for the late reply, but I have never tried the albumentation in this codebase.