Releases: obss/sahi
Releases · obss/sahi
0.11.22
What's Changed
- Improve suppot for latest mmdet (v3.3.0) by @fcakyon in #1129
- Improve support for latest yolov5-pip and ultralytics versions by @fcakyon in #1130
- support latest huggingface/transformers models by @fcakyon in #1131
- refctor coco to yolo conversion, update docs by @fcakyon in #1132
- bump version by @fcakyon in #1134
Full Changelog: 0.11.21...0.11.22
Core Documentation Files
Prediction Utilities
- Detailed guide for performing object detection inference
- Standard and sliced inference examples
- Batch prediction usage
- Class exclusion during inference
- Visualization parameters and export formats
- Interactive examples with various model integrations (YOLOv8, MMDetection, etc.)
Slicing Utilities
- Guide for slicing large images and datasets
- Image slicing examples
- COCO dataset slicing examples
- Interactive demo notebook reference
COCO Utilities
- Comprehensive guide for working with COCO format datasets
- Dataset creation and manipulation
- Slicing COCO datasets
- Dataset splitting (train/val)
- Category filtering and updates
- Area-based filtering
- Dataset merging
- Format conversion (COCO ↔ YOLO)
- Dataset sampling utilities
- Statistics calculation
- Result validation
CLI Commands
- Complete reference for SAHI command-line interface
- Prediction commands
- FiftyOne integration
- COCO dataset operations
- Environment information
- Version checking
- Custom script usage
FiftyOne Integration
- Guide for visualizing and analyzing predictions with FiftyOne
- Dataset visualization
- Result exploration
- Interactive analysis
Interactive Examples
All documentation files are complemented by interactive Jupyter notebooks in the demo directory:
slicing.ipynb- Slicing operations demonstrationinference_for_ultralytics.ipynb- YOLOv8/YOLO11/YOLO12 integrationinference_for_yolov5.ipynb- YOLOv5 integrationinference_for_mmdetection.ipynb- MMDetection integrationinference_for_huggingface.ipynb- HuggingFace models integrationinference_for_torchvision.ipynb- TorchVision models integrationinference_for_rtdetr.ipynb- RT-DETR integrationinference_for_sparse_yolov5.ipynb- DeepSparse optimized inference
Getting Started
If you're new to SAHI:
- Start with the prediction utilities to understand basic inference
- Explore the slicing utilities to learn about processing large images
- Check out the CLI commands for command-line usage
- Dive into COCO utilities for dataset operations
- Try the interactive notebooks in the demo directory for hands-on experience
Contributors
fcakyon
Assets 2
0
Join discussion
0.11.21
What's Changed
- Exclude classes from inference using pretrained or custom models by @gguzzy in #1104
- pyproject.toml, pre-commit, ruff, uv and typing issues, fixes #1119 by @Dronakurl in #1120
- add class exclusion example into predict docs by @gguzzy in #1125
- Add OBB demo by @fcakyon in #1126
- fix a type hint typo in predict func by @fcakyon in #1111
- Remove numpy<2 upper pin by @weiji14 in #1112
- fix ci badge on readme by @fcakyon in #1124
- fix version in pyproject.toml by @fcakyon in #1127
New Contributors
- @Dronakurl made their first contribution in #1120
- @gguzzy made their first contribution in #1104
Full Changelog: 0.11.20...0.11.21
0.11.20
0.11.19
What's Changed
- fix ci actions by @fcakyon in #1073
- Update has_mask method for mmdet models (handle an edge case) by @ccomkhj in #1066
- Another self-intersection corner case handling by @sergiev in #982
- Update README.md by @fcakyon in #1077
- drop non-working yolonas support by @fcakyon in #1097
- drop yolonas support part2 by @fcakyon in #1098
- Update has_mask method for mmdet models (handle ConcatDataset) by @ccomkhj in #1092
New Contributors
Full Changelog: 0.11.18...0.11.19
0.11.18
What's Changed
- add yolov8 mask support, improve mask processing speed by 4-5x by @mayrajeo in #1039
- fix has_mask method for mmdet models by @Alias-z in #1054
- Fix
TypeError: 'GeometryCollection' object is not subscriptablewhen slicing COCO by @Alias-z in #1047 - support opencv-python version 4.9 by @iokarkan in #1041
- add upperlimit to numpy dep by @fcakyon in #1057
- add more unit tests by @MMerling in #1048
- upgrade ci actions by @fcakyon in #1049
New Contributors
- @iokarkan made their first contribution in #1041
- @MMerling made their first contribution in #1048
- @Alias-z made their first contribution in #1047
Full Changelog: 0.11.16...0.11.18
v0.11.16
What's Changed
- Update README.md by @fcakyon in #966
- Updated the constant variables for Yolov8 by @AmoghDhaliwal in #917
- Slicing: add unique slice index number to output file name by @jokober in #943
- update version by @fcakyon in #967
- Correcting type hints for
get_slice_bboxes(). by @S-aiueo32 in #930 - Correcting
slice_image()by @S-aiueo32 in #931 - Customize YOLOv8 image_size & device + Allow Saving Slices by @lakshaymehra in #929
- fix package testing workflow paths by @fcakyon in #968
- remove detectron2 from package tests by @fcakyon in #979
- ONNX runtime support by @karl-joan in #922
- Fix RLE when segmentation is None by @bobyard-com in #996
- update functions docstrings and type hinting by @fcakyon in #1016
- fix postprocess type options description by @williamlung in #1013
- revert back package version by @fcakyon in #1017
- Improve printout readability by @jacobmarks in #1009
- improve readme by @fcakyon in #1018
- remove an unused list in postprocess by @developer0hye in #1002
- add more contributors to readme by @fcakyon in #1019
- add more contributors to readme by @fcakyon in #1020
- Adds link to new FiftyOne tutorial by @jacobmarks in #1023
- RTDETR implementation by @edugzlez in #940
- Improve yolov8 config by @GuillaumeBruand in #988
- update readme by @fcakyon in #1034
- relax opencv version by @fcakyon in #1035
- disable slice export by default by @fcakyon in #1036
- remove quality param in slice export due to errors by @fcakyon in #1037
- fix for using bgr image in inference instead of rgb by @bilkosem in #1022
New Contributors
- @AmoghDhaliwal made their first contribution in #917
- @S-aiueo32 made their first contribution in #930
- @lakshaymehra made their first contribution in #929
- @karl-joan made their first contribution in #922
- @bobyard-com made their first contribution in #996
- @williamlung made their first contribution in #1013
- @jacobmarks made their first contribution in #1009
- @developer0hye made their first contribution in #1002
- @edugzlez made their first contribution in #940
- @GuillaumeBruand made their first contribution in #988
- @bilkosem made their first contribution in #1022
Full Changelog: 0.11.15...0.11.16
v0.11.15
What's Changed
- add more yolov8 download utils by @pranavdurai10 in #887
- add learnopencv research article url into readme by @pranavdurai10 in #892
- update supergradients version in tests by @fcakyon in #894
- added yolov8x download utils + update README.md by @pranavdurai10 in #898
- Replace mmdet v2 for v3 by @i-aki-y in #893
- Update README.md by @pranavdurai10 in #925
- fix continious integration workflows by @fcakyon in #963
- Self-intersection handling by @sergiev in #961
- Slicing: Quality focused default image format selection and jpg quality by @jokober in #956
- update version by @fcakyon in #964
New Contributors
- @pranavdurai10 made their first contribution in #887
- @i-aki-y made their first contribution in #893
- @sergiev made their first contribution in #961
- @jokober made their first contribution in #956
Full Changelog: 0.11.14...0.11.15
0.11.14
What's Changed
- support Deci-AI YOLO-NAS models by @ssahinnkadir in #874
- Significant speed improvement for Detectron2 models by @MyosQ in #865
- support ultralytics>=8.0.99 by @eVen-gits in #873
- Documentation typo, and missing value by @Hamzalopode in #859
- update version by @fcakyon in #876
- update black version by @fcakyon in #877
New Contributors
- @Hamzalopode made their first contribution in #859
- @eVen-gits made their first contribution in #873
- @MyosQ made their first contribution in #865
Full Changelog: 0.11.13...0.11.14
v0.11.13
What's Changed
- Add hide labels/confidence by @aphilas in #844
- Add support for sparse YOLOv5 models by @mwitiderrick in #848
- Add support for DeepSparse dynamic image pipelines by @mwitiderrick in #850
- fix package testing by @fcakyon in #853
New Contributors
- @aphilas made their first contribution in #844
- @mwitiderrick made their first contribution in #848
Full Changelog: 0.11.12...0.11.13
v0.11.12
What's Changed
- update readme by @fcakyon in #825
- update yolov5 in tests by @fcakyon in #826
- fix coco2yolov5 cli by @fcakyon in #828
- from_coco_dict_or_path, added threading for processing coco_dict["images"] by @ilkermanap in #827
- update yolo in tests by @fcakyon in #829
- update mmdet in tests by @fcakyon in #830
- Add yolov8 to SAHI by @NguyenTheAn in #833
- Added heuristic for bounding bbox ordering by @MLDovakin in #835
- update citation info by @fcakyon in #837
- Fix RGB channel ordering for yolov8 by @mayrajeo in #836
- update minor version by @fcakyon in #842
New Contributors
- @ilkermanap made their first contribution in #827
- @NguyenTheAn made their first contribution in #833
- @MLDovakin made their first contribution in #835
- @mayrajeo made their first contribution in #836
Full Changelog: 0.11.11...0.11.12
