Hollon desktop buck/doe engine (experimental)

1. Unzip the folder.
2. Open a terminal in that folder and run: python3 -m http.server 8789 --bind 127.0.0.1
   On Windows with Python installed: py -m http.server 8789 --bind 127.0.0.1
3. Open http://localhost:8789 in a recent WebGPU-capable Chrome or Edge desktop browser.
4. Click Load local model. The model downloads from Hugging Face on first use.
5. Select local trail-camera images. Images are processed locally and never uploaded.
6. Download the JSON results if desired. This tool does not sync to your Hollon account.

No account credentials, Reveal credentials, cloud API tokens, or payment details are needed.
Model predictions are experimental. Unknown means manual review, not no deer.
Keep the terminal open while processing; Ctrl+C stops the local server.
Includes the existing Hollon local inference runtime using @huggingface/transformers.
See THIRD-PARTY-NOTICES.txt for bundled software license notices.
