feat: Why does evaluation run significantly slower than training when using the same data and the same rollout size? #120
Description
Describe the feature
Why does evaluation run significantly slower than training when using the same data and the same rollout size?
Why do you need this feature?
No response
Additional context
No response