1.2.5

[1.2.5] - 2024-05-04

Added

• Added exception in mltu.dataProvider.DataProvider to raise ValueError when dataset is not iterable
• Added custom training code for YoloV8 object detector: Tutorials\11_Yolov8\train_yolov8.py
• Added custom trained inference code for YoloV8 object detector:Tutorials\11_Yolov8\test_yolov8.py

Changed

• Fixed RandomElasticTransform in mltu.augmentors to handle elastic transformation not to exceed image boundaries
• Modified YoloPreprocessor in mltu.torch.yolo.preprocessors to output dictionary with np.arrays istead of lists

1.2.4

[1.2.4] - 2024-03-21

Added

• Added RandomElasticTransform to mltu.augmentors to work with Image objects
• Added xyxy_abs to mltu.annotations.detections.Detection object to return absolute bounding boxes

Changes

• Changed ImageShowCV2 transformer in mltu.transformers to display images when running with multiple threads

1.2.3

[1.2.3] - 2024-03-17

Added

• Added Tutorial how to run YOLOv8 pretrained Object Detection model Tutorials.11_Yolov8.README.md

1.2.2

[1.2.2] - 2024-03-15

Changed

• Bug fixed with loss_info local variable in mltu.torch.model.Model object

Added

• Added RandomColorMode and RandomZoom into mltu.augmentors

1.2.1

[1.2.1] - 2024-03-12

Changed

• Fixed many minor bugs

Added

• Added mltu.transformers.ImageNormalizer to normalize and transpose images
• Added mltu.torch.yolo.annotation.VOCAnnotationReader to read VOC annotation files
• Added mltu.torch.yolo.preprocessors.YoloPreprocessor to preprocess images and annotations for YoloV8 detection model

1.2.0

[1.2.0] - 2024-03-05

Changed

• Creating code to work with Ultralytics YoloV8 Detection model (training and inference)
• Updated previous tutorials to work with the latest mltu changes
• Updated mltu.augmentors.RandomRotate to work with Detections objects
• Changed to use importlib to import librosa in mltu.preprocessors to avoid import errors
• Changed mltu.torch.model.Model object to provide more flexibility in training and validation
• Improved mltu.torch.callbacks to provide more flexibility in training and validation

Added

• Added mltu.torch.detection module, that contains Detections and Detection objects, to handle detection annotations
• Added RandomDropBlock and RandomDropBlock augmentors into mltu.augmentors to work with Detections objects
• Added ModelEMA into mltu.torch.model to work with EMA (Exponential Moving Average) model
• Added FpsWrapper into mltu.inferenceModel to automatically calculate FPS (Frames Per Second) when using inference model
• Added mltu.torch.yolo.detector.BaseDetector as a base class for preprocessing and postprocessing detection models
• Added mltu.torch.yolo.detector.onnx_detector.Detector as a class to handle YoloV8 onnx model detection inference
• Added mltu.torch.yolo.detector.torch_detector.Detector as a class to handle YoloV8 torch model detection inference
• Added mltu.torch.yolo.loss.v8DetectionLoss as a class to handle YoloV8 detection loss in training
• Added mltu.torch.yolo.metrics.YoloMetrics as a class to handle YoloV8 detection metrics in training and validation
• Added mltu.torch.yolo.optimizer module, that contains AccumulativeOptimizer object and build_optimizer function, to handle YoloV8 detection optimizer in training
• Added YoloV8 Detection tutorial in Tutorials.11_yolov8 that shows how to do basic inference with torch and exported onnx models

1.1.8

[1.1.8] - 2024-01-02

Changed

• Fixed setup.py to include mltu.torch and mltu.tensorflow packages and other packages that are required for mltu to work properly

1.1.7

[1.1.7] - 2022-10-30

Changed

• Fixed bug in mltu.torch.dataProvider.SequenceHandler, to handle len sequence < queue_size case

1.1.6

Fixing dependency with librosa library

1.1.5

[1.1.5] - 2022-10-17

Changed

• Fixed dependencies with librosa library