量子位官网T2·62 pts
- Guanglun Intelligence released RoboFinals RL Envs, providing interactive and verifiable simulation environments for robot policy training
- RoboFinals RL Envs handles policy training while RoboFinals Benchmark provides independent evaluation, forming a train-and-diagnose loop
- Post-training supply is expanding from algorithms and compute to training data and interactive environments, with physical-world environments far more complex than digital tasks