Releases · obss/sahi

12 Aug 20:26

fcakyon

0.10.4

072ea62

v0.10.4

What's Changed

hotfix for occasional segmentation fault by @fcakyon in #546
refactor requirement checking by @fcakyon in #549
Code formatting and checks are moved to a single module. by @devrimcavusoglu in #548
fix a typo in type hinting by @ymerkli in #553
fix an incorrect url in comments by @aynursusuz in #554
update version by @fcakyon in #555
fix linting script by @devrimcavusoglu in #558
fix an incorrect url in comments by @aynursusuz in #559

New Contributors

@ymerkli made their first contribution in #553
@aynursusuz made their first contribution in #554

Full Changelog: 0.10.3...0.10.4

Contributors

ymerkli, fcakyon, and 2 other contributors

Assets 2

02 Aug 07:43

fcakyon

0.10.3

1a9ae25

v0.10.3

What's Changed

fix coco2fiftyone script by @fcakyon in #540
fix segmentation fault in same cases by @fcakyon in #541

Full Changelog: 0.10.2...0.10.3

Contributors

fcakyon

Assets 2

28 Jul 07:24

fcakyon

0.10.2

23b2754

v0.10.2

What's Changed

add automatic slice size calculation by @mcvarer in #512
Fix bug that Unable to print for prediction time when verbose=2 by @youngjae-avikus in #521
pybboxes allow oob, strict=False fix. by @devrimcavusoglu in #528
update predict verbose by @fcakyon in #514
update pybboxes version by @fcakyon in #531

New Contributors

@youngjae-avikus made their first contribution in #521

Full Changelog: 0.10.1...0.10.2

Contributors

mcvarer, fcakyon, and 2 other contributors

Assets 2

0 Join discussion

25 Jun 20:07

fcakyon

0.10.1

5912293

v0.10.1

What's Changed

add python 3.10 support by @fcakyon in #503
add file_name to export_visuals by @mcvarer in #507
refactor automodel to lazyload models by @fcakyon in #509
update automodel loading method to from_pretrained by @fcakyon in #510

New Contributors

@mcvarer made their first contribution in #507

Full Changelog: 0.10.0...0.10.1

Contributors

mcvarer and fcakyon

Assets 2

21 Jun 14:15

fcakyon

0.10.0

4934db9

v0.10.0

New Features

- Layer.ai integration

from sahi import AutoDetectionModel

detection_model = AutoDetectionModel.from_layer("layer/yolov5/models/yolov5s")

result = get_sliced_prediction(
    "image.jpeg",
    detection_model,
    slice_height = 512,
    slice_width = 512,
    overlap_height_ratio = 0.2,
    overlap_width_ratio = 0.2
)

- HuggingfFace Transformers object detectors

from sahi.model import HuggingfaceDetectionModel

detection_model = HuggingfaceDetectionModel(
    model_path="facebook/detr-resnet-50",
    image_size=640,
    confidence_threshold=0.5
)

result = get_sliced_prediction(
    "image.jpeg",
    detection_model,
    slice_height = 512,
    slice_width = 512,
    overlap_height_ratio = 0.2,
    overlap_width_ratio = 0.2
)

- TorchVision object detectors

import torchvision
from sahi.model import TorchVisionDetectionModel

model = torchvision.models.detection.fasterrcnn_resnet50_fpn(pretrained=True)

detection_model = TorchVisionDetectionModel(
    model=model,
    image_size=640,
    confidence_threshold=0.5
)

result = get_sliced_prediction(
    "image.jpeg",
    detection_model,
    slice_height = 512,
    slice_width = 512,
    overlap_height_ratio = 0.2,
    overlap_width_ratio = 0.2
)

- Support for exporting predictions in COCO format

from sahi.utils.coco import Coco, CocoImage, CocoAnnotation, CocoPrediction
from sahi.utils.file import save_json
from pycocotools.cocoeval import COCOeval
from pycocotools.coco import COCO

coco_obj = Coco()

# add n images to coco_obj
for _ in range(n):
    image = CocoImage(**kwargs)
    
    # add n annotations to the image
    for _ in ange(n):
        image.add_annotation(CocoAnnotation(**kwargs))
    
    # add n predictions to the image
    for _ in range(n)
        image.add_prediction(CocoPrediction(**kwargs))
    
    # add image to coco object
    coco_obj.add_image(image)

# export ground truth annotations
coco_gt = coco_obj.json
save_json(coco_gt , "ground_truth.json")

# export predictions 
coco_predictions = coco_obj.prediction_array
save_json(coco_predictions, "predictions.json")

coco_ground_truth = COCO(annotation_file="coco_dataset.json")
coco_predictions = coco_ground_truth.loadRes("coco_predictions.json")
coco_evaluator = COCOeval(coco_ground_truth, coco_predictions, "bbox")
coco_evaluator.evaluate()
coco_evaluator.accumulate()
coco_evaluator.summarize()

What's Changed

refactor torch utils by @fcakyon in #468
add automodel structure for unified model loading by @fcakyon in #469
add support to instantiate a DetectionModel from layer by @mecevit in #462
refactor automodel by @fcakyon in #470
update layer versions in workflows by @fcakyon in #471
update version to 0.10.0 by @fcakyon in #474
Added support for exporting predictions in COCO format by @PushpakBhoge in #465
update contributors in readme by @fcakyon in #477
update device priority for Detectron2DetectionModel by @fcakyon in #479
fix pickle export for video by @fcakyon in #481
update continuous integration by @fcakyon in #483
refactor import and torch utils by @fcakyon in #484
make detectionmodel classes more explicit in automodel by @fcakyon in #485
utilize check_requirements in several modules by @fcakyon in #487
update package versions in workflows by @fcakyon in #488
add support for huggingface transformers object detectors by @devrimcavusoglu in #475
add torchvision detector support by @fcakyon in #486
remove legacy image_size parameter by @kadirnar in #494
AutoDetectionModel can be imported from sahi by @fcakyon in #498
add python3.6 support by @fcakyon in #489
refactor exception handling by @kadirnar in #499
improve requirements and import handling by @fcakyon in #502