facebookresearch/slbo

Understanding Logging

Closed this issue · 0 comments

How do I make sense of the log.json? How can I retrieve the avg reward in term of samples (steps) like Figure 1 in the paper?