Article

GroundAnything: Balancing Parallel Decoding and Precise Visual Grounding with Blockwise Denoising

GroundAnything is a 4-billion-parameter visual grounding foundation model that combines parallel decoding with precise object localization through blockwise denoising. Its training integrates pretraining on public datasets, direct conversion from autoregressive to diffusion, supervised fine-tuning, and GRPO-based reinforcement post-training.

Part of the event: GroundAnything Visual Grounding Model Paper Released

This page provides a summary and source link only.