autoOD

Name	autoOD JSON
Version	1.0.23 JSON
	download
home_page
Summary	Module to automate object detection process.
upload_time	2023-02-02 17:43:59
maintainer
docs_url	None
author	Elina Chertova
requires_python
license
keywords	objectdetection automl autood
VCS
bugtrack_url
requirements	No requirements were recorded.
Travis-CI	No Travis.
coveralls test coverage	No coveralls.

            # autoOD

## Docker

## Augmentation
Available augmentations' methods: [link](docs/augmentation/Документация_по_аугментации.pdf)

## Tutorial
### Установка пакета
Ссылка на проект на pypi: `https://pypi.org/project/autoOD/`.
```angular2html
pip install autoOD
```

## Конфигурация проекта
В корне рабочей директории необходимо создать файл `object_detection.ini`. Пример:
```angular2html
[Environment]
; required fields to change: dataset_dir, path_to_work_dir, image_folder
dataset_dir=/Users/elinachertova/PycharmProjects/autoOD/all_images_data
path_to_work_dir=/Users/elinachertova/PycharmProjects/autoOD/
image_folder=all_images_data
test_folder=test
train_folder=train
all_data_folder=all_data
annotations_folder=annotations
result_folder=result
models_folder=models
additional_folder=additional

[Dataset]
extension=jpg

[Dataset.Preparation]
train_set_percent=0.8
is_shuffle=True

[Training.Params]
; If is_custom_settings == False, the params will be selected automatically.
is_custom_settings=True
; is_custom_settings==False
model_id=16
batch_size=16
num_steps=1000

;
[Training.EarlyStopping]
early_stopping=True
patience=5
min_delta=0.001
save_step=3
;

[Training.Overfitting]
is_prevent_overfitting=True
use_dropout=True
use_score_threshold=True

[Training.Loss]
is_loss_function=False
classification_weight=1.0
localization_weight=1.0

[Training.Augmentation]
; If is_auto_augmentation == True, AutoAu
is_auto_augmentation=False
; Available values: v0, v1, v2, v3. v0 preferable
policy=v0


; If is_augmentation == False, the section will be skipped.
is_augmentation=True


normalize_image=False


horizontal_flip=True
probability_horizontal_flip=0.2


vertical_flip=True
probability_vertical_flip=0.5


rotation90=False
probability_rotation90=0.3


pixel_value_scale=False


entire_rgb_to_grey=True
probability_entire_rgb_to_grey=0.07


adjust_brightness=False
max_delta_adjust_brightness=0.2


adjust_contrast=True
min_delta_adjust_contrast=0.8
max_delta_adjust_contrast=1.25


adjust_hue=False
max_delta_adjust_hue=0.02


adjust_saturation=False
min_delta_adjust_saturation=0.8
max_delta_adjust_saturation=1.25


distort_color=False
color_ordering=1


crop_image=False
min_object_covered_crop_image=1.0
min_aspect_ratio_crop_image=0.75
max_aspect_ratio_crop_image=1.33
min_area_crop_image=0.1
max_area_crop_image=1.0
overlap_thresh_crop_image=0.3
clip_boxes_crop_image=True
random_coef_crop_image=0.0


pad_image=False


crop_pad_image=False


crop_to_aspect_ratio=False
aspect_ratio_crop_to_aspect_ratio=1.0
overlap_thresh_crop_to_aspect_ratio=0.3
clip_boxes_crop_to_aspect_ratio=True


black_patches=True
max_black_patches=10
probability_black_patches=0.2
size_to_image_ratio=0.1


rgb_to_grey=False


drop_label_probabilistically=False
label=1
drop_probability=1.0


jpeg_quality=False
random_coef_jpeg_quality=0.0
min_jpeg_quality=0
max_jpeg_quality=100


patch_gaussian=False
random_coef_patch_gaussian=0.0
min_patch_size=1
max_patch_size=250
min_gaussian_stddev=0.0
max_gaussian_stddev=1.0

```

Необходимо заменить конфигурацию из раздела Environment на пользовательскую. Остальные параметры настраиваются по желанию.


            

Raw data

            {
    "_id": null,
    "home_page": "",
    "name": "autoOD",
    "maintainer": "",
    "docs_url": null,
    "requires_python": "",
    "maintainer_email": "",
    "keywords": "ObjectDetection,autoML,autoOD",
    "author": "Elina Chertova",
    "author_email": "elas.2015@mail.ru",
    "download_url": "https://files.pythonhosted.org/packages/fd/89/37ce588956718b165663932344b89bbfd5e30d041fa0cc326e6a05d1ff55/autoOD-1.0.23.tar.gz",
    "platform": null,
    "description": "# autoOD\n\n## Docker\n\n## Augmentation\nAvailable augmentations' methods: [link](docs/augmentation/\u0414\u043e\u043a\u0443\u043c\u0435\u043d\u0442\u0430\u0446\u0438\u044f_\u043f\u043e_\u0430\u0443\u0433\u043c\u0435\u043d\u0442\u0430\u0446\u0438\u0438.pdf)\n\n## Tutorial\n### \u0423\u0441\u0442\u0430\u043d\u043e\u0432\u043a\u0430 \u043f\u0430\u043a\u0435\u0442\u0430\n\u0421\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u043f\u0440\u043e\u0435\u043a\u0442 \u043d\u0430 pypi: `https://pypi.org/project/autoOD/`.\n```angular2html\npip install autoOD\n```\n\n## \u041a\u043e\u043d\u0444\u0438\u0433\u0443\u0440\u0430\u0446\u0438\u044f \u043f\u0440\u043e\u0435\u043a\u0442\u0430\n\u0412 \u043a\u043e\u0440\u043d\u0435 \u0440\u0430\u0431\u043e\u0447\u0435\u0439 \u0434\u0438\u0440\u0435\u043a\u0442\u043e\u0440\u0438\u0438 \u043d\u0435\u043e\u0431\u0445\u043e\u0434\u0438\u043c\u043e \u0441\u043e\u0437\u0434\u0430\u0442\u044c \u0444\u0430\u0439\u043b `object_detection.ini`. \u041f\u0440\u0438\u043c\u0435\u0440:\n```angular2html\n[Environment]\n; required fields to change: dataset_dir, path_to_work_dir, image_folder\ndataset_dir=/Users/elinachertova/PycharmProjects/autoOD/all_images_data\npath_to_work_dir=/Users/elinachertova/PycharmProjects/autoOD/\nimage_folder=all_images_data\ntest_folder=test\ntrain_folder=train\nall_data_folder=all_data\nannotations_folder=annotations\nresult_folder=result\nmodels_folder=models\nadditional_folder=additional\n\n[Dataset]\nextension=jpg\n\n[Dataset.Preparation]\ntrain_set_percent=0.8\nis_shuffle=True\n\n[Training.Params]\n; If is_custom_settings == False, the params will be selected automatically.\nis_custom_settings=True\n; is_custom_settings==False\nmodel_id=16\nbatch_size=16\nnum_steps=1000\n\n;\n[Training.EarlyStopping]\nearly_stopping=True\npatience=5\nmin_delta=0.001\nsave_step=3\n;\n\n[Training.Overfitting]\nis_prevent_overfitting=True\nuse_dropout=True\nuse_score_threshold=True\n\n[Training.Loss]\nis_loss_function=False\nclassification_weight=1.0\nlocalization_weight=1.0\n\n[Training.Augmentation]\n; If is_auto_augmentation == True, AutoAu\nis_auto_augmentation=False\n; Available values: v0, v1, v2, v3. v0 preferable\npolicy=v0\n\n\n; If is_augmentation == False, the section will be skipped.\nis_augmentation=True\n\n\nnormalize_image=False\n\n\nhorizontal_flip=True\nprobability_horizontal_flip=0.2\n\n\nvertical_flip=True\nprobability_vertical_flip=0.5\n\n\nrotation90=False\nprobability_rotation90=0.3\n\n\npixel_value_scale=False\n\n\nentire_rgb_to_grey=True\nprobability_entire_rgb_to_grey=0.07\n\n\nadjust_brightness=False\nmax_delta_adjust_brightness=0.2\n\n\nadjust_contrast=True\nmin_delta_adjust_contrast=0.8\nmax_delta_adjust_contrast=1.25\n\n\nadjust_hue=False\nmax_delta_adjust_hue=0.02\n\n\nadjust_saturation=False\nmin_delta_adjust_saturation=0.8\nmax_delta_adjust_saturation=1.25\n\n\ndistort_color=False\ncolor_ordering=1\n\n\ncrop_image=False\nmin_object_covered_crop_image=1.0\nmin_aspect_ratio_crop_image=0.75\nmax_aspect_ratio_crop_image=1.33\nmin_area_crop_image=0.1\nmax_area_crop_image=1.0\noverlap_thresh_crop_image=0.3\nclip_boxes_crop_image=True\nrandom_coef_crop_image=0.0\n\n\npad_image=False\n\n\ncrop_pad_image=False\n\n\ncrop_to_aspect_ratio=False\naspect_ratio_crop_to_aspect_ratio=1.0\noverlap_thresh_crop_to_aspect_ratio=0.3\nclip_boxes_crop_to_aspect_ratio=True\n\n\nblack_patches=True\nmax_black_patches=10\nprobability_black_patches=0.2\nsize_to_image_ratio=0.1\n\n\nrgb_to_grey=False\n\n\ndrop_label_probabilistically=False\nlabel=1\ndrop_probability=1.0\n\n\njpeg_quality=False\nrandom_coef_jpeg_quality=0.0\nmin_jpeg_quality=0\nmax_jpeg_quality=100\n\n\npatch_gaussian=False\nrandom_coef_patch_gaussian=0.0\nmin_patch_size=1\nmax_patch_size=250\nmin_gaussian_stddev=0.0\nmax_gaussian_stddev=1.0\n\n```\n\n\u041d\u0435\u043e\u0431\u0445\u043e\u0434\u0438\u043c\u043e \u0437\u0430\u043c\u0435\u043d\u0438\u0442\u044c \u043a\u043e\u043d\u0444\u0438\u0433\u0443\u0440\u0430\u0446\u0438\u044e \u0438\u0437 \u0440\u0430\u0437\u0434\u0435\u043b\u0430 Environment \u043d\u0430 \u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u0442\u0435\u043b\u044c\u0441\u043a\u0443\u044e. \u041e\u0441\u0442\u0430\u043b\u044c\u043d\u044b\u0435 \u043f\u0430\u0440\u0430\u043c\u0435\u0442\u0440\u044b \u043d\u0430\u0441\u0442\u0440\u0430\u0438\u0432\u0430\u044e\u0442\u0441\u044f \u043f\u043e \u0436\u0435\u043b\u0430\u043d\u0438\u044e.\n\n",
    "bugtrack_url": null,
    "license": "",
    "summary": "Module to automate object detection process.",
    "version": "1.0.23",
    "split_keywords": [
        "objectdetection",
        "automl",
        "autood"
    ],
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "blake2b_256": "fd8937ce588956718b165663932344b89bbfd5e30d041fa0cc326e6a05d1ff55",
                "md5": "1c01539e530a2fae03cd5f28c38acb0e",
                "sha256": "5e0c9c405c1b96e8bcb43ffd6afcb50a615b8a328dddd36feb7fb56a11d328ad"
            },
            "downloads": -1,
            "filename": "autoOD-1.0.23.tar.gz",
            "has_sig": false,
            "md5_digest": "1c01539e530a2fae03cd5f28c38acb0e",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 57199,
            "upload_time": "2023-02-02T17:43:59",
            "upload_time_iso_8601": "2023-02-02T17:43:59.497093Z",
            "url": "https://files.pythonhosted.org/packages/fd/89/37ce588956718b165663932344b89bbfd5e30d041fa0cc326e6a05d1ff55/autoOD-1.0.23.tar.gz",
            "yanked": false,
            "yanked_reason": null
        }
    ],
    "upload_time": "2023-02-02 17:43:59",
    "github": false,
    "gitlab": false,
    "bitbucket": false,
    "lcname": "autood"
}