Writing intermediate checkpoints during RL?

[DAlvGar]

Hi all,

I am trying to reproduce parts of this work (https://doi.org/10.1016/j.bmc.2021.116308). In the super detailed paper (thanks!) the authors mention they save an agent checkpoint every 50 steps during the RL runs. I believe this work was run using Reinvent 2. Is there any way to do the same now with Reinvent4? I could not find any obvious way to do this.

Any help is appreciated!
Thanks!

Daniel

[halx]

Hi,

we only save checkpoints at termination. You would need to modify the source code to achieve regular writes.

Cheers,
Hannes.