This work presents WeAgent-MMGenEdit, a full-stack recipe including a multimodal harness, a scalable data construction pipeline, a comprehensive benchmark, and post-training methods for the agent policy and image backend that enables a 30B-total/3B-active policy to outperform similarly sized policy models and approach...
Hui Zhang, Zongkai Liu, Li-Qiang Niu et al.· 1 citation
WeAgent-Harness, a multimodal agentic harness that supports native text-vision interaction and runtime recovery, and WeAgent-MMSearch, an integrated system spanning data construction, agentic post-training, and multimodal rollout that outperform similarly sized open-source models and rival models with roughly ten times...
Zongkai Liu, Hui Zhang, Li-Qiang Niu et al.· 1 citation
Proximal Policy Optimization (PPO) for large language models typically trains its critic by mean-squared-error (MSE) regression on scalar value targets. Although scalar MSE is statistically valid for estimating the conditional expected return, sparse binary rewards in reinforcement learning with verifiable rewards (RLV...
Zhi-Jian Zhou, Long Li, Xuan Zhang et al.· 2 citations· ⚡2
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
