RS-HyRe-R1 combines three rewards in RL training to overcome perceptual inertia in remote sensing VLMs, achieving SOTA results on REC, OVD, and VQA with a 3B-parameter model that outperforms larger ones.
VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Model
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
RS-HyRe-R1: A Hybrid Reward Mechanism to Overcome Perceptual Inertia for Remote Sensing Images Understanding
RS-HyRe-R1 combines three rewards in RL training to overcome perceptual inertia in remote sensing VLMs, achieving SOTA results on REC, OVD, and VQA with a 3B-parameter model that outperforms larger ones.