hello.fiftyone.dataset module

hello.fiftyone.dataset.add_classification_labels(dataset, label_field, labels_path)

hello.fiftyone.dataset.add_coco_labels(dataset, label_field, labels_path, label_type='detections')

hello.fiftyone.dataset.add_dataset(dataset, skip_existing=True, insert_new=True, fields=None, expand_schema=True)

hello.fiftyone.dataset.add_dataset_dir(dataset_dir, data_path=None, labels_path=None, label_field=None, tags=None)

hello.fiftyone.dataset.add_detection_labels(dataset, label_field, labels_path, classes=None, mode='text', remove_prefix=False)

Adds detection labels to the dataset.

Note

if mode=text, a text row corresponds to a sample prediction result. row format: filepath,height,width,x1,y1,x2,y2,s,l,x1,y1,x2,y2,s,l.

if mode=yolo, a txt file corresponds to a sample prediction result. row format: target,xc,yc,w,h,s.

if mode=coco, a standard COCO format json file. from https://cocodataset.org/#format-data.

Parameters:

• dataset – a fiftyone.core.dataset.Dataset

• label_field (str) – the label field in which to store the labels

• labels_path (str) – the labels load from

• classes (list) – the list of class label strings

• mode (str) – supported values are ("text", "yolo", "coco")

hello.fiftyone.dataset.add_images_dir(dataset, images_dir, tags=None, recursive=True)

Adds the given directory of images to the dataset.

Parameters:

• dataset – a fiftyone.core.dataset.Dataset

• images_dir (str) – a directory of images

• tags (None) – an optional tag or iterable of tags to attach to each sample

• recursive (True) – whether to recursively traverse subdirectories

hello.fiftyone.dataset.add_segmentation_labels(dataset, label_field, labels_path, mask_targets='auto', mode='png')

Adds segmentation labels to the dataset.

Parameters:

• dataset – a fiftyone.core.dataset.Dataset

• label_field (str) – the label field in which to store the labels

• labels_path (str) – the labels load from

• mask_targets (dict) – a dict mapping pixel values to semantic label strings

• mode (str) – supported values are ("png", "coco")

hello.fiftyone.dataset.add_yolo_labels(dataset, label_field, labels_path, classes)

hello.fiftyone.dataset.create_dataset(dataset_name, dataset_type, version='001', classes=[], mask_targets={}, force=False)

Create an empty fiftyone.core.dataset.Dataset with the name.

Parameters:

• dataset_name (str) – a name for the dataset

• dataset_type (str) – supported values are ("detection", "segmentation", "unknown")

• classes (list, optional) – defaults to []

• mask_targets (dict, optional) – defaults to {}

Returns:

a fiftyone.core.dataset.Dataset

hello.fiftyone.dataset.delete_datasets(names=None, non_persistent=False, force=False)

hello.fiftyone.dataset.delete_duplicate_images(dataset)

Delete duplicate images.

Parameters:

dataset – a fiftyone.core.dataset.Dataset

hello.fiftyone.dataset.delete_duplicate_labels(dataset, label_field, iou_thresh=0.999, method='simple', iscrowd=None, classwise=True)

Delete duplicate labels in the given field of the dataset, as defined as labels with an IoU greater than a chosen threshold with another label in the field.

Parameters:

• dataset – a fiftyone.core.dataset.Dataset

• label_field – a label field of type fiftyone.core.labels.Detections or fiftyone.core.labels.Polylines

• iou_thresh (0.999) – the IoU threshold to use to determine whether labels are duplicates

• method ("simple") – supported values are ("simple", "greedy")

• iscrowd (None) – an optional name of a boolean attribute

• classwise (True) – different label values as always non-overlapping

hello.fiftyone.dataset.export_classification_dataset(export_dir, dataset, label_field, splits=None, export_media=True)

hello.fiftyone.dataset.export_classification_labels(export_dir, dataset, label_field, splits=None)

hello.fiftyone.dataset.export_dataset(export_dir, dataset, label_field=None, mask_label_field=None, mask_types='stuff', splits=None)

Exports the samples in the collection to disk.

Parameters:

• export_dir – the directory to which to export the samples

• dataset – a fiftyone.core.collections.SampleCollection

• label_field – controls the label field(s) to export

• mask_label_field – controls the label field(s) to export

• mask_types ("stuff") – “stuff”(amorphous regions of pixels), “thing”(connected regions, each representing an instance)

• splits (None) – a list of strings, respectively, specifying the splits to load. If “auto” will computes the distinct tags

hello.fiftyone.dataset.export_detection_dataset(export_dir, dataset, label_field, splits=None)

hello.fiftyone.dataset.export_image_dataset(export_dir, dataset, splits=None)

hello.fiftyone.dataset.export_segmentation_dataset(export_dir, dataset, label_field, mask_types='stuff', splits=None)

hello.fiftyone.dataset.list_datasets()

hello.fiftyone.dataset.load_dataset(name)

Loads the FiftyOne dataset with the given name.

Parameters:

name (str) – the name of the dataset

hello.fiftyone.dataset.load_images_dir(dataset_dir, dataset_name, dataset_type, version='001', classes=[], mask_targets={})

Create a fiftyone.core.dataset.Dataset from the given directory of images.

Parameters:

• dataset_dir (str) – a directory of images

• dataset_name (str) – a name for the dataset

• dataset_type (str) – supported values are ("detection", "segmentation", "unknown")

• classes (list, optional) – defaults to []

• mask_targets (dict, optional) – defaults to {}

Returns:

a fiftyone.core.dataset.Dataset