Ml-agents: Is it possible to resume training?
So, if I do a trainnig session, lets say a 5.0e5 steps session. Can I do 5.0e5 more steps starting from the result of the previous training?
GeriBP
All 3 comments
You can add --load to your command. For me it would be like this:
python learn.py name --run-id=13 --train --load
The run id has to be the same as the run you want to resume, and you have to increase the max steps to your new maximum (so 10e5 in your case)
Hengoo
on 22 Mar 2018
👍8
Thats exactly how its done. Thanks!
GeriBP
on 22 Mar 2018
This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.
![lock[bot] picture](https://avatars1.githubusercontent.com/in/6672?v=4&s=40)
lock[bot]
on 3 Jan 2020
Was this page helpful?
0 / 5 - 0 ratings
Related issues
RavenLeeANU
路
4Comments
scotthovestadt
路
4Comments
MarcPilgaard
路
3Comments
jlanis
路
4Comments
Rodnyy
路
3Comments
Most helpful comment
You can add --load to your command. For me it would be like this:
python learn.py name --run-id=13 --train --load
The run id has to be the same as the run you want to resume, and you have to increase the max steps to your new maximum (so 10e5 in your case)