| JVM | Platform | Status |
|---|---|---|
| OpenJDK (Temurin) Current | Linux |  |
| OpenJDK (Temurin) LTS | Linux |  |
| OpenJDK (Temurin) Current | Windows |  |
| OpenJDK (Temurin) LTS | Windows |  |
The laurel package attempts to provide tools to assist
with image captioning within the context of machine
learning.
In particular, the application is geared towards the management of smaller datasets (in the range of thousands of images) for use in techniques such as LORA training.
- A user interface for managing images and captions for those images.
- A caption categorization system for assisting with keeping captions consistent across large datasets.
- The ability to import captions and images into a dataset from a directory hierarchy.
- The ability to export captions and images into a directory for use in training scripts.
- A persistent undo/redo system that can store every change ever made to a dataset, including the ability to effectively revert to an earlier version at any time.
- A carefully-engineered Java API for manipulating datasets; the command-line tools and user interface are thin shells over this API.
- Datasets are backed by SQLite for reliable, transactional updates, and a file format that is designed to endure for decades to come.
- Command line tools for automating operations such as importing, exporting, and interrogating metadata.
- Comprehensive documentation.
- OSGi-ready.
- JPMS-ready.
- ISC license.
See the documentation.