insect-detect-ml-v1.2

In the v1.2 release, new Google Colab notebooks for custom training of YOLOv6, YOLOv7 and YOLOv8 detection models were added.

The check_images.py script was added, which can be used to automatically remove corrupted images that can sometimes be generated by the automated monitoring script of the DIY camera trap and will cause an error while running the classification script. Running the script will also give you the mean/min/max image width/height of your collected data as output.

New features

• Add YOLOv6, YOLOv7 and YOLOv8 model training notebooks: ba914e0
• Add check_images.py script: 2508fa4

Important Updates

Update license to AGPLv3: 78d55fc

Changelog

Changes between previous release and this release: v1.1...v1.2

insect-detect-ml-v1.1

For the v1.1 release, a new YOLOv5s-cls model in ONNX format for faster CPU inference was added. This model was trained on a new dataset version, including episyr_balt (Episyrphus balteatus) as added class.

New features

Add new YOLOv5n-cls ONNX model: 84e15ec

Important Updates

Update csv_analysis.py: 660a6dc

Changelog

Changes between previous release and this release: v1.0...v1.1

insect-detect-ml-v1.0

First release of the insect-detect-ml repository, containing Jupyter notebooks for YOLOv5 detection and classification model training and Python scripts for classification and analysis of insect images and metadata, collected with the Insect Detect DIY camera trap.