Lazy fetch top values cross validation. 20k epochs.
Cross validation for lazy fetching with topped selection
| | L0 | L1 |
|---|---|---|
| learning rate | 0.2 | 0.2 |
| | E0 | E1 | E2 |
|---|---|---|---|
| epsilon | 0.01 | 0.01 | 0.01 |
| | D0 | D1 |
|---|---|---|
| discount | 0.3 | 0.3 |
| | M0 |
|---|---|
| mapping | non-linear-3 |
| | R0 |
|---|---|
| reward handler | can-see |
| | F0 | F1 | F2 | F3 | F4 |
|---|---|---|---|---|---|
| fetch mode | lazy-s-t2 | lazy-s-t5 | lazy-s-t10 | lazy-m-t2 | lazy-m-t5 |
| | F5 | F6 | F7 | F8 | |
| fetch mode | lazy-m-t10 | lazy-l-t2 | lazy-l-t5 | lazy-l-t10 |