【Agentic RL / 强化学习框架 / VERL】Uni-Agent Gateway 深度解析 - 罗西的思考
0x00 Uni-AgentGateway VERL RFC #5790Agent Abstractions and Trajectory Gateway for VERLRFC #5790 Gateway Agent/ RL Token OpenAI API loss mask token Agent token Uni-Agent Gateway RFC Uni-Agent … [+48756 chars]