Skip to content

data_loader_config

AspectRatioBucketConfig

target_resolution 

target_resolution: int

The target resolution for all aspect ratios. When generating aspect ratio buckets, the resolution of each bucket is selected to have roughly target_resolution * target_resolution pixels (i.e. a square image with dimensions equal to target_resolution).

start_dim 

start_dim: int

Aspect ratio bucket resolutions are generated as follows:

  • Iterate over 'first' dimension values from start_dim to end_dim in steps of size divisible_by.
  • Calculate the 'second' dimension to be as close as possible to the total number of pixels in target_resolution, while still being divisible by divisible_by.
Choosing aspect ratio buckets

The aspect ratio bucket resolutions are logged at the start of training with the number of images in each bucket. Review these logs to make sure that images are being split into buckets as expected.

Highly fragmented splits (i.e. many buckets with few examples in each) can 1) limit the extent to which examples can be shuffled, and 2) slow down training if there are many partial batches.

end_dim 

end_dim: int

See explanation under start_dim.

divisible_by 

divisible_by: int

See explanation under start_dim.

ImageCaptionSDDataLoaderConfig

type 

type: Literal['IMAGE_CAPTION_SD_DATA_LOADER'] = 'IMAGE_CAPTION_SD_DATA_LOADER'

dataset 

dataset: ImageCaptionDatasetConfig

aspect_ratio_buckets 

aspect_ratio_buckets: AspectRatioBucketConfig | None = None

resolution 

resolution: int | tuple[int, int] = 512

The resolution for input images. Either a scalar integer representing the square resolution height and width, or a (height, width) tuple. All of the images in the dataset will be resized to this resolution unless the aspect_ratio_buckets config is set.

center_crop 

center_crop: bool = True

If True, input images will be center-cropped to the target resolution. If False, input images will be randomly cropped to the target resolution.

random_flip 

random_flip: bool = False

Whether random flip augmentations should be applied to input images.

caption_prefix 

caption_prefix: str | None = None

A prefix that will be prepended to all captions. If None, no prefix will be added.

dataloader_num_workers 

dataloader_num_workers: int = 0

Number of subprocesses to use for data loading. 0 means that the data will be loaded in the main process.

DreamboothSDDataLoaderConfig

type 

type: Literal['DREAMBOOTH_SD_DATA_LOADER'] = 'DREAMBOOTH_SD_DATA_LOADER'

instance_caption 

instance_caption: str

class_caption 

class_caption: Optional[str] = None

instance_dataset 

instance_dataset: ImageDirDatasetConfig

class_dataset 

class_dataset: Optional[ImageDirDatasetConfig] = None

class_data_loss_weight 

class_data_loss_weight: float = 1.0

The loss weight applied to class dataset examples. Instance dataset examples have an implicit loss weight of 1.0.

aspect_ratio_buckets 

aspect_ratio_buckets: AspectRatioBucketConfig | None = None

The aspect ratio bucketing configuration. If None, aspect ratio bucketing is disabled, and all images will be resized to the same resolution.

resolution 

resolution: int | tuple[int, int] = 512

The resolution for input images. Either a scalar integer representing the square resolution height and width, or a (height, width) tuple. All of the images in the dataset will be resized to this resolution unless the aspect_ratio_buckets config is set.

center_crop 

center_crop: bool = True

If True, input images will be center-cropped to the target resolution. If False, input images will be randomly cropped to the target resolution.

random_flip 

random_flip: bool = False

Whether random flip augmentations should be applied to input images.

dataloader_num_workers 

dataloader_num_workers: int = 0

Number of subprocesses to use for data loading. 0 means that the data will be loaded in the main process.

TextualInversionSDDataLoaderConfig

type 

type: Literal['TEXTUAL_INVERSION_SD_DATA_LOADER'] = 'TEXTUAL_INVERSION_SD_DATA_LOADER'

dataset 

dataset: ImageDirDatasetConfig | ImageCaptionDatasetConfig

caption_preset 

caption_preset: Literal['style', 'object'] | None = None

caption_templates 

caption_templates: list[str] | None = None

A list of caption templates with a single template argument 'slot' in each. E.g.:

  • "a photo of a {}"
  • "a rendering of a {}"
  • "a cropped photo of the {}"

keep_original_captions 

keep_original_captions: bool = False

If True, then the captions generated as a result of the caption_preset or caption_templates will be used as prefixes for the original captions. If False, then the generated captions will replace the original captions.

aspect_ratio_buckets 

aspect_ratio_buckets: AspectRatioBucketConfig | None = None

The aspect ratio bucketing configuration. If None, aspect ratio bucketing is disabled, and all images will be resized to the same resolution.

resolution 

resolution: int | tuple[int, int] = 512

The resolution for input images. Either a scalar integer representing the square resolution height and width, or a (height, width) tuple. All of the images in the dataset will be resized to this resolution unless the aspect_ratio_buckets config is set.

center_crop 

center_crop: bool = True

If True, input images will be center-cropped to the target resolution. If False, input images will be randomly cropped to the target resolution.

random_flip 

random_flip: bool = False

Whether random flip augmentations should be applied to input images.

shuffle_caption_delimiter 

shuffle_caption_delimiter: str | None = None

If None, then no caption shuffling is applied. If set, then captions are split on this delimiter and shuffled.

dataloader_num_workers 

dataloader_num_workers: int = 0

Number of subprocesses to use for data loading. 0 means that the data will be loaded in the main process.