SyncPro-RL

SyncPro-RL is the official reinforcement-learning checkpoint for SyncPro: Synchronous Reasoning and Responding for Proactive Streaming Video Understanding.

SyncPro synchronously decodes an internal reasoning stream and an external response stream under a causal streaming attention map.

Model

SyncPro forks the decoder into synchronous internal-reasoning and external-response branches. At each streaming step, it learns whether to remain silent, reason internally, or speak, while preserving causal processing.

This checkpoint starts from SyncPro-SFT and is optimized with the paper's balanced hard-F1 reinforcement-learning recipe:

• 960 ordered training rows, with 240 schedule assignments in each of four proactive behavior groups;
• chronological hard-window F1 rewards;
• query batch size 16 and group size 4;
• 60 sampled frames per trajectory;
• a 60-step training run with fixed gamma = 0.5 branch merging;
• this released model is the selected checkpoint at step 50.

The release contains merged BF16 inference weights. RL optimizer state and rl_trainer_state.pt are intentionally excluded. Inference code and benchmark adapters are documented in the SyncPro repository.

Data and intended use

The model is intended for research on causal streaming-video understanding, proactive response timing, and separated internal/external generation. The public SyncPro-189k release contains the SFT annotations only; the RL annotations used for this checkpoint are not distributed. Source videos are not redistributed and must be obtained from their original publishers under the applicable terms.

The checkpoint may inherit limitations and biases from its base model and training datasets. It should not be treated as a safety-critical monitoring system.

License

The model weights are released under the Apache License 2.0. The annotation dataset has separate licensing and source-specific terms described on its dataset card.

Citation

@misc{tian2026syncpro,
  title  = {SyncPro: Synchronous Reasoning and Responding for
            Proactive Streaming Video Understanding},
  author = {Tian, Xueyun and Xu, Bingbing and Liu, Jiaming and
            Yu, Jinpeng and Shi, Yang and Zhao, Yu and
            Huang, Ruihua and Wang, Yuanzhuo and Shen, Huawei},
  year   = {2026},
  url    = {https://eureka-maggie.github.io/SyncPro/}
}