OBB-PBD

Checkpoints for OBB-PBD: Extending Parallel Box Decoding to Oriented Bounding Boxes for Referring Expression Grounding in Aerial Imagery (Chang Woo Song, Yeong Jun Koh; Remote Sensing, MDPI, accepted, 2026).

Release status

The checkpoints of all nine trained runs (OBB-PBD and the two matched controls, three seeds each) will be uploaded here when the article is published. Each consists of LoRA adapters on the language model and the last six MoonViT blocks, plus the fully trained mlp1 projector, loaded on top of nvidia/LocateAnything-3B.

Code: github.com/ce-dric/OBB-PBD

License

Distributed under the NVIDIA License (LICENSE): non-commercial research and evaluation use only.

Improved using Qwen. The language model of LocateAnything-3B is based on Qwen2.5-3B-Instruct, licensed under the Qwen RESEARCH LICENSE AGREEMENT (QWEN_RESEARCH_LICENSE.txt, notice in NOTICE).

Citation

The citation will be added when the article is published.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for ce-dric/OBB-PBD

Base model

Qwen/Qwen2.5-3B

Finetuned

Qwen/Qwen2.5-3B-Instruct

Finetuned

nvidia/LocateAnything-3B

Adapter

(1)

this model