AudioRubrics: Audio Reasoning with Evolving Rubric Rewards
Upload an audio clip and ask a question about it. The model reasons step-by-step about what it hears.
Based on Reinforcement Learning with Evolving Rubrics as Rewards for Audio Reasoning | a Qwen2.5-Omni-7B model fine-tuned with GRPO using self-evolving, audio-grounded rubric rewards.
Examples
Model: umd-zhou-lab/AudioRubrics | Paper: arXiv:2608.02831 | Code: GitHub