BatDetect2 Version 2.0 Beta

[mbsantiago]

BatDetect2 Version 2.0 Beta

It is finally here, updated code for improved training, evaluation and finetuning workflows!

This beta release is a major refresh of BatDetect2, with a new end-to-end workflow for running inference, training custom models, fine-tuning existing checkpoints, and evaluating models on labelled datasets.

Some of the main updates are:

• a new BatDetect2API for Python workflows
• updated CLI workflows centred around process, train, finetune, and evaluate
• improved checkpoint handling, including bundled aliases and Hugging Face support
• clearer dataset, targets, outputs, and evaluation configuration workflows
• richer output handling for predictions, evaluation artifacts, and training provenance
• a large docs refresh covering tutorials, how-to guides, explanations, and references
• migration support for older workflows and legacy formats

If you are upgrading from the previous version, please check the migration guide first:

• docs/source/legacy/migration-guide.md

It explains the main workflow changes, renamed commands, updated defaults, and how to move from the older API and CLI patterns.

The docs site is still being built and will be live soon.
In the meantime, if you want a quick look, please check the docs/ folder in this repository.

This is a beta release, and we would really appreciate feedback.
If you run into bugs, rough edges, confusing behaviour, or missing docs, please open an issue and let us know.

Thanks to everyone who contributed ideas, code, testing, and feedback while this update was coming together, and thanks for your patience while we worked through such a big overhaul.

[Copilot]

Copilot wasn't able to review this pull request because it exceeds the maximum number of files (300). Try reducing the number of changed files and requesting a review from Copilot again.