浏览代码

ghost->get_step

/asymm-envs
Andrew Cohen 5 年前
当前提交
80fd858a
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/ghost/trainer.py

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


# will push the current_snapshot into the correct queue. Otherwise,
# it will continue skipping and swap_snapshot will continue to handle
# pushing fixed snapshots
next_learning_team = self.controller.get_learning_team(self.ghost_step)
next_learning_team = self.controller.get_learning_team(self.get_step)
for brain_name in self._internal_policy_queues:
internal_policy_queue = self._internal_policy_queues[brain_name]
try:

正在加载...
取消
保存