← Back to Architecture
parallel residuals
ArchitectureUsed in
42 PRs
Best BPB
1.0616
Avg BPB
1.1408
Submissions
PR #1204by msisovicRECORD
1.1063PR #1274by MatoTeziTanka
1.0876PR #1326by aryanbhosale
1.0896PR #1333by aryanbhosale
1.0766PR #1334by aryanbhosaleRECORD
1.0897PR #1338by bigbag
1.0955PR #1339by bigbag
1.0955PR #1381by X-Abhishek-X
1.1604PR #1396by erichroepke
1.1067PR #1412by Robby955RECORD
1.0835PR #1420by abaybektursun
1.0801PR #1425by dentity007
1.4479PR #1435by AbhayAnandUCSD
1.0980PR #1437by dexhunter
1.0780PR #1450by andrewbaggio1
1.0848PR #1477by aryanbhosaleRECORD
1.0822PR #1485by ndokutovich
1.0679PR #1489by joshkmartinez
1.0736PR #1492by bigbag
1.0810PR #1493by bigbagRECORD
1.0810PR #1499by dippatel1994
1.6323PR #1515by dexhunter
1.0872PR #1521by aryanbhosale
1.0802PR #1532by nogakeren
1.0803PR #1534by someone114514
1.0846PR #1541by bigbag
1.0778PR #1570by yufang67
1.0970PR #1578by mikeapedia
1.0668PR #1614by seekerPrice
1.5096PR #1620by shiawyonglim
1.6644PR #1635by PapaFranku4647
1.1063PR #1647by powerpratik
1.0616PR #1661by anderamondarainh-stack
1.1444PR #1667by MarioPaerle
1.0714PR #1720by kiyoaki
1.0818PR #1725by teslaeco
1.0813PR #1731by Victory963
1.0785PR #1733by G3sparky
1.3262PR #1737by sakthivarshans
1.0723PR #1750by teslaeco
1.0809PR #1755by OE-GOD
1.0746PR #1760by BrandtChristian
1.1863Hyperparameters Across PRs
| pr_number | parameters |
|---|---|
| 1204 | {"start_layer":7} |
| 1274 | {"start_layer":7} |
| 1326 | {"start_layer":7} |
| 1333 | {"start_layer":7} |
| 1334 | {"start_layer":7} |
| 1338 | {"start_layer":7} |
| 1339 | {"start_layer":7} |
| 1381 | {"start_layer":7} |
| 1396 | {"start_layer":7} |
| 1412 | {"start_layer":7} |
| 1420 | {"start_layer":7,"end_layer":10} |
| 1425 | {"start_layer":6} |
| 1435 | {"layers":"7+"} |
| 1437 | {"start_layer":7,"end_layer":10} |
| 1450 | {"layers":[7,8,9,10]} |
| 1477 | {"start_layer":7} |
| 1485 | {"start_layer":7} |
| 1489 | {"start_layer":7} |
| 1492 | {"layers":"7+"} |
| 1493 | {"layers":"7+"} |
| 1499 | {"start_layer":7} |
| 1515 | {"start_layer":7} |
| 1521 | — |
| 1532 | {"layers":"7+"} |
| 1534 | — |
| 1541 | {"start_layer":7,"new_scalar_params":66} |
| 1570 | {"start_layer":7} |
| 1578 | — |
| 1614 | {"start_layer":7} |
| 1620 | — |
| 1635 | {"start_layer":7} |
| 1647 | — |
| 1661 | {"start_layer":7} |
| 1667 | {"start_layer":7} |
| 1720 | {"start_layer":7} |
| 1725 | — |
| 1731 | {"start_layer":7} |
| 1733 | — |
| 1737 | {"start_layer":7} |
| 1750 | — |
| 1755 | {"start_layer":7} |
| 1760 | {"start_layer":7} |