Cannot reproduce training results in evaluation even on same dataset

Lars_Simon_Zehnder · November 20, 2022, 12:53pm

@SVH This has been answered by @mannyv in the reply to a similar problem of yours, I guess.

The explore attribute for evaluation has to be set to True to achieve comparable results to training.

Topic		Replies	Views
Training mean reward vs. evaluation mean rewward RLlib	4	1268	November 17, 2022
Test reward much lower than training reward RLlib	3	425	July 17, 2022
Reproducibility of ray.tune with seeds RLlib	6	2780	July 26, 2022
Why does my evaluation during training with tune return 0 values? RLlib	2	280	September 6, 2022
Anomalous behaviour with some plateaus during training RLlib	0	9	November 28, 2024