inoryy / reaver

Reaver: Modular Deep Reinforcement Learning Framework. Focused on StarCraft II. Supports Gym, Atari, and MuJoCo.

MIT License

554 stars 89 forks source link

Restore previous experiment result issue #32

Open kimbring2 opened 5 years ago

kimbring2 commented 5 years ago

Thank you for always kindly answering.

Suddenly the computer stopped when learning and tried to use the restore function to start with the learning results done so far. However, I discovered in the log file the phenomenon that reading was no longer progressing and stopped at a place.

In this case, which part is the problem?

19-22-02

From Dohyeong

inoryy commented 5 years ago

Hello, from what you're showing it seems to be working fine. If it's completely stuck for longer than a couple of minutes then try re-running, SC2 client itself can sometimes get stuck.

kimbring2 commented 5 years ago

I find the reason of stuck.

It seems that there was a problem reading the log file. When commented line of an actor_critic.py file as follows, it confirm that it operate normally.

` def on_start(self):

self.logger.on_start()

inoryy commented 5 years ago

That is very strange. Can you track down the exact error?

kimbring2 commented 5 years ago

The unexpected error message does not occur, but when I test it using the print function, it stops at line of self.run_time = int (logs [0] .split ("") [- 1]) in logger.py file.