StreamOPD: A Post-Training Recipe with Spatio-Temporal Cue Gating for Streaming Video Understanding
StreamOPD improves streaming video understanding via on-policy distillation with verifiable rewards and a spatio-temporal cue-gating mechanism, achieving near-teacher performance w…
Hugging Face · Daily Papers
·Keming Wu, Baoyi Wang
·
·▲ 4 upvotes
Este artigo está em destaque na seleção diária de papers do Hugging Face, curada pela comunidade de pesquisa em IA.
Autores: Keming Wu, Baoyi Wang, Kaichen Zhang, Xiang An, Zuhao Yang, Sudong Wang
- 4 upvotes da comunidade
- Temas: StreamOPD, on-policy distillation, verifiable rewards, thinking mode, instruct mode, Spatio-Temporal CueGate
Resumo
Resumo original (em inglês), extraído do paper:
StreamOPD improves streaming video understanding via on-policy distillation with verifiable rewards and a spatio-temporal cue-gating mechanism, achieving near-teacher performance without inference-time memory.