Training failures commonly occur due to one of three reasons:
- The AnyLogic model itself contains logical errors. Sometimes, the RL policy can learn behavior that causes it to encounter rare logical errors in your model. Please follow these instructions to try to reproduce them back in AnyLogic.
- You are using custom libraries that are not yet supported in Pathmind.
- Pathmind ran out of memory during training. Complex AnyLogic models (i.e. those with thousands of observation) require more memory to execute. Try simplifying your observations to see if it resolves training failures.
When training fails, please follow these steps to confirm that your AnyLogic runs without error. If training continues to fail, please contact Pathmind support with the following information:
- Link to the experiment
- Copy of the AnyLogic alp (if possible)