浏览代码

Cherry-pick separate critic only for PPO (#4661) (#4666)

/release_10_branch
GitHub 4 年前
当前提交
f1206bed
共有 1 个文件被更改,包括 1 次插入1 次删除
  1. 2
      ml-agents/mlagents/trainers/ppo/trainer.py

2
ml-agents/mlagents/trainers/ppo/trainer.py


behavior_spec,
self.trainer_settings,
condition_sigma_on_obs=False, # Faster training for PPO
separate_critic=behavior_spec.action_spec.is_continuous(),
separate_critic=True, # Match network architecture with TF
)
return policy

正在加载...
取消
保存