VideoLLaMA2.1-7B-AV, fine-tuned with contrastive DPO on the audio projector so it
grounds captions in the audio that's actually present instead of hallucinating
sounds from visual cues.
Checkpoint from "Don't Let the Video Speak: Audio-Contrastive Preference
Optimization for Audio-Visual Language Models" (ECCV 2026).
Training: 5k VALOR clips, audio projector only, β=0.1, lr=2e-5, 10 frames.
Standard VideoLLaMA2 checkpoint — load with the
VideoLLaMA2 codebase.
1@inproceedings{acpo2026,
2 title = {Don't Let the Video Speak: Audio-Contrastive Preference Optimization for Audio-Visual Language Models},
3 author = {Baid, Ami and Xue, Zihui and Grauman, Kristen},
4 booktitle = {ECCV},
5 year = {2026}
6}
Released under Apache-2.0. The weights are additionally subject to the VideoLLaMA2
base-model license; the benchmark to AVHBench; training data to VALOR.
