Слияние кода завершено, страница обновится автоматически
All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
kitti_raw importer and exporter for dimensions (height, width, length) in meters
(https://github.com/open-edge-platform/datumaro/pull/1596)KaggleImageCsvBase, KaggleImageTxtBase, KaggleImageMaskBase, KaggleVocBase, KaggleYoloBase)
(https://github.com/open-edge-platform/datumaro/pull/1240)__getitem__() for random accessing with O(1) time complexity
(https://github.com/open-edge-platform/datumaro/pull/1247)KaggleCocoBase
(https://github.com/open-edge-platform/datumaro/pull/1273)CityscapesBase and KaggleImageMaskBase by dropping np.unique
(https://github.com/open-edge-platform/datumaro/pull/1261)datum project import command and add filtering example to filter out items containing annotations.
(https://github.com/open-edge-platform/datumaro/pull/1210)urls is not specified
(https://github.com/open-edge-platform/datumaro/pull/1362)ClassificationValidator for multi-label classification datasets with label_groups
(https://github.com/open-edge-platform/datumaro/pull/1116)xml.etree with defusedxml
(https://github.com/open-edge-platform/datumaro/pull/1117)GroupType with IntEnum for, where 0 is EXCLUSIVE
(https://github.com/open-edge-platform/datumaro/pull/1116)bytes property returning None if function is given to data
(https://github.com/open-edge-platform/datumaro/pull/978)person_layout categories and action_classification attributes in imported Pascal-VOC dataset
(https://github.com/open-edge-platform/datumaro/pull/997)Cityscapes to drop ImgsFine directory
(https://github.com/open-edge-platform/datumaro/pull/1023)is_encrypted
(https://github.com/open-edge-platform/datumaro/pull/907)__len__() of Subset when the item is removed
(https://github.com/open-edge-platform/datumaro/pull/854)__repr__ for Dataset
(https://github.com/open-edge-platform/datumaro/pull/750)PointCloud media type,
DatasetItem.media and .media_as(type) members
(https://github.com/open-edge-platform/datumaro/pull/539)media_type
(https://github.com/open-edge-platform/datumaro/pull/539)MultiframeImage
media type (https://github.com/open-edge-platform/datumaro/pull/628)data/ prefix inclusion in YOLO export
(https://github.com/open-edge-platform/datumaro/pull/689)describe-downloads to print information about downloadable datasets
(https://github.com/open-edge-platform/datumaro/pull/678)--depth parameter for detect-dataset CLI command
(https://github.com/open-edge-platform/datumaro/pull/680)download command
(https://github.com/open-edge-platform/datumaro/pull/697)env.detect_dataset() now returns a list of detected formats at all recursion levels
instead of just the lowest one
(https://github.com/open-edge-platform/datumaro/pull/680)backup, names and classes
(instead of just train and valid). The reserved names now raise an error on exporting.
(https://github.com/open-edge-platform/datumaro/pull/688)--save-images is replaced with --save-media in CLI and converter API
(https://github.com/open-edge-platform/datumaro/pull/539)image, point_cloud and related_images of DatasetItem are
replaced with media and media_as(type) members and c-tor parameters
(https://github.com/open-edge-platform/datumaro/pull/539)video_frames format and
to split a video into frames with the datum util split_video command
(https://github.com/open-edge-platform/datumaro/pull/555)--subset parameter in the image_dir format
(https://github.com/open-edge-platform/datumaro/pull/555)MediaManager API to control loaded media resources at runtime
(https://github.com/open-edge-platform/datumaro/pull/555)import datumaro
(https://github.com/open-edge-platform/datumaro/pull/630)export, transform, ...)
(https://github.com/open-edge-platform/datumaro/pull/630)random_sampler, label_random_sampler)
to create smaller datasets from bigger ones
(https://github.com/open-edge-platform/datumaro/pull/636,
https://github.com/open-edge-platform/datumaro/pull/640)remove_items, remove_annotations, remove_attributes)
(https://github.com/open-edge-platform/datumaro/pull/670)datum import. Such sources are imported like
when the rpath parameter is specified, however, only the selected path
is copied into the project
(https://github.com/open-edge-platform/datumaro/pull/555)stats performance, added new filtering parameters,
image stats (unique, repeated) moved to the dataset section,
removed mean and std from the dataset section
(https://github.com/open-edge-platform/datumaro/pull/621)Image creation from just size info
(https://github.com/open-edge-platform/datumaro/pull/634)pyproject.toml-based build
(https://github.com/open-edge-platform/datumaro/pull/671)components.extractor
(https://github.com/open-edge-platform/datumaro/pull/630)add, import and export commands without a project
(https://github.com/open-edge-platform/datumaro/pull/555)make_dataset on empty project tree now produces the error properly
(https://github.com/open-edge-platform/datumaro/pull/555)Resize transform
(https://github.com/open-edge-platform/datumaro/issues/606)Resize transform
(https://github.com/open-edge-platform/datumaro/issues/607)Resize transform
(https://github.com/open-edge-platform/datumaro/issues/618)stats command
(e.g. https://github.com/open-edge-platform/datumaro/issues/607)
(https://github.com/open-edge-platform/datumaro/pull/621)ByteImage
(https://github.com/open-edge-platform/datumaro/pull/595)pycocotools dependency lower bound is raised to 2.0.4.
(https://github.com/open-edge-platform/datumaro/pull/449)smooth_line from datumaro.util.annotation_util - the function
is renamed to approximate_line and has updated interface
(https://github.com/open-edge-platform/datumaro/pull/592)score attribute in KITTI detetion
(https://github.com/open-edge-platform/datumaro/pull/571)ade20k2017, ade20k2020, camvid, coco, cvat, datumaro,
icdar_text_localization, icdar_text_segmentation,
icdar_word_recognition, imagenet_txt, kitti_raw, label_me, lfw,
mot_seq, open_images, vgg_face2, voc, widerface, yolo
(https://github.com/open-edge-platform/datumaro/pull/531,
https://github.com/open-edge-platform/datumaro/pull/536,
https://github.com/open-edge-platform/datumaro/pull/550,
https://github.com/open-edge-platform/datumaro/pull/557,
https://github.com/open-edge-platform/datumaro/pull/558)datum merge command
(https://github.com/open-edge-platform/datumaro/pull/545)Image, ByteImage from datumaro.util.image - these classes
are moved to datumaro.components.media
(https://github.com/open-edge-platform/datumaro/pull/538)datumaro.components.media.Image
and numpy.ndarray
(https://github.com/open-edge-platform/datumaro/pull/568)people.txt became optional in LFW
(https://github.com/open-edge-platform/datumaro/pull/509)image_ids_and_rotation.csv became optional Open Images
(https://github.com/open-edge-platform/datumaro/pull/509)_) in subset names in COCO
(https://github.com/open-edge-platform/datumaro/pull/509)icdar_text_localization format is no longer detected in every directory
(https://github.com/open-edge-platform/datumaro/pull/531)pycocotools version to 2.0.2
(https://github.com/open-edge-platform/datumaro/pull/534)color attribute into the
icdar_text_segmentation format
(https://github.com/open-edge-platform/datumaro/pull/556)pip install datumaro[default], which should
be used by default. The simple datumaro is supposed for library users.
(https://github.com/open-edge-platform/datumaro/pull/238)diff, merge, explain and info CLI commands
(https://github.com/open-edge-platform/datumaro/pull/238)import, remove, commit, checkout, log, status, info CLI commands
(https://github.com/open-edge-platform/datumaro/pull/238)Coco*Extractor classes now have an option to preserve label IDs from the
original annotation file
(https://github.com/open-edge-platform/datumaro/pull/453)patch CLI command to patch datasets
(https://github.com/open-edge-platform/datumaro/pull/401)ProjectLabels transform to change dataset labels for merging etc.
(https://github.com/open-edge-platform/datumaro/pull/401,
https://github.com/open-edge-platform/datumaro/pull/478)imagenet_txt import
(https://github.com/open-edge-platform/datumaro/pull/434,
https://github.com/open-edge-platform/datumaro/pull/489)datumaro.components.annotation
(https://github.com/open-edge-platform/datumaro/pull/439)Project class from datumaro.components is changed completely
(https://github.com/open-edge-platform/datumaro/pull/238)diff and ediff are joined into a single diff CLI command
(https://github.com/open-edge-platform/datumaro/pull/238)datum project migrate
(https://github.com/open-edge-platform/datumaro/pull/238)CliPlugin is not required in plugin classes
(https://github.com/open-edge-platform/datumaro/pull/238)Importers do not create Projects anymore and just return a list of
extractor configurations
(https://github.com/open-edge-platform/datumaro/pull/238)import, project merge CLI commands
(https://github.com/open-edge-platform/datumaro/pull/238)datumaro_project format
(https://github.com/open-edge-platform/datumaro/pull/238)path field of DatasetItem
(https://github.com/open-edge-platform/datumaro/pull/455)open_images_format.py
(https://github.com/open-edge-platform/datumaro/pull/440)lazy_image returning unrelated data sometimes
(https://github.com/open-edge-platform/datumaro/issues/409)pycocotools.mask.iou
(https://github.com/open-edge-platform/datumaro/pull/450)Dataset.is_modified
(https://github.com/open-edge-platform/datumaro/pull/401)RemapLabels
(https://github.com/open-edge-platform/datumaro/pull/401)label for Points read in the LFW format
(https://github.com/open-edge-platform/datumaro/pull/494)Dataset (https://github.com/open-edge-platform/datumaro/pull/386)Dataset remembers export options on saving / exporting for the first time (https://github.com/open-edge-platform/datumaro/pull/386)remap_labels to dataset categories of different length (https://github.com/open-edge-platform/datumaro/issues/314)*_labelIds.png in Cityscapes export (https://github.com/open-edge-platform/datumaro/issues/325, https://github.com/open-edge-platform/datumaro/issues/342)ProjectDataset.transform() with a string argument (https://github.com/open-edge-platform/datumaro/issues/402)ItemTransform class, which describes item-wise dataset Transforms (https://github.com/open-edge-platform/datumaro/pull/297)keep-empty export parameter in VOC format (https://github.com/open-edge-platform/datumaro/pull/297)Dataset changes done by transforms (https://github.com/open-edge-platform/datumaro/pull/297)remap_labels transform (https://github.com/open-edge-platform/datumaro/pull/203)background class no more introduce an instance (https://github.com/open-edge-platform/datumaro/pull/188)Icdar13/15 dataset format (https://github.com/open-edge-platform/datumaro/pull/96)Dataset (https://github.com/open-edge-platform/datumaro/pull/102)Market-1501 dataset format (https://github.com/open-edge-platform/datumaro/pull/108)LFW dataset format (https://github.com/open-edge-platform/datumaro/pull/110)diff command (https://github.com/open-edge-platform/datumaro/pull/117)diff command invocation problem with missing class methods (https://github.com/open-edge-platform/datumaro/pull/117)WiderFace dataset format (https://github.com/open-edge-platform/datumaro/pull/65, https://github.com/open-edge-platform/datumaro/pull/90)VGGFace2 dataset format (https://github.com/open-edge-platform/datumaro/pull/69, https://github.com/open-edge-platform/datumaro/pull/82)datumaro
Dataset class extended with new operations: save, load, export, import_from, detect, run_model (https://github.com/open-edge-platform/datumaro/pull/71)Extractor-only defined formats
(in Project.import_from, dataset.import_from and CLI/project import) (https://github.com/open-edge-platform/datumaro/pull/71)datum project ... commands replaced with datum ... commands (https://github.com/open-edge-platform/datumaro/pull/84)ImageNet extractors (https://github.com/open-edge-platform/datumaro/pull/85)Importer-defined formats as project sources (source add) (https://github.com/open-edge-platform/datumaro/pull/86)ImageDir format and importers (https://github.com/open-edge-platform/datumaro/pull/86)datum project ... CLI context (https://github.com/open-edge-platform/datumaro/pull/84)Extractor (instead of only SourceExtractor)
(https://github.com/open-edge-platform/datumaro/pull/70)pip for pycocotools (https://github.com/open-edge-platform/datumaro/pull/73)YOLO extractor path matching on Windows (https://github.com/open-edge-platform/datumaro/pull/73)labelmap parameter type checking in VOC converter (https://github.com/open-edge-platform/datumaro/pull/76)CamVid dataset format (https://github.com/open-edge-platform/datumaro/pull/57)opencv-python-headless dependency with DATUMARO_HEADLESS=1
environment variable instead of opencv-python (https://github.com/open-edge-platform/datumaro/pull/62)ImageNet and ImageNetTxt dataset formats (https://github.com/open-edge-platform/datumaro/pull/41)label-map parameter value for VOC converter (https://github.com/open-edge-platform/datumaro/pull/34)Transform.subsets() method (https://github.com/open-edge-platform/datumaro/pull/38)ByteImage class to represent encoded images in memory and avoid recoding
on save (https://github.com/open-edge-platform/datumaro/pull/27)default is now a default subset name, instead of None. The values are
interchangeable. (https://github.com/open-edge-platform/datumaro/pull/22)image/depth value from VOC export (https://github.com/open-edge-platform/datumaro/pull/27)reindex option in COCO and CVAT converters (https://github.com/open-edge-platform/datumaro/pull/18)## [Unreleased]
### New features
- TBD
### Enhancements
- TBD
### Deprecated
- TBD
### Removed
- TBD
### Bug fixes
- TBD
### Security
- TBD
Вы можете оставить комментарий после Вход в систему
Неприемлемый контент может быть отображен здесь и не будет показан на странице. Вы можете проверить и изменить его с помощью соответствующей функции редактирования.
Если вы подтверждаете, что содержание не содержит непристойной лексики/перенаправления на рекламу/насилия/вульгарной порнографии/нарушений/пиратства/ложного/незначительного или незаконного контента, связанного с национальными законами и предписаниями, вы можете нажать «Отправить» для подачи апелляции, и мы обработаем ее как можно скорее.
Опубликовать ( 0 )