Zenith

1. zenith-0172026-08-25 · negativeswamp localized panel holdoutDoes snowy-Hero successor candidate `3d37e24ca4e8b57038724062fe87a370c98ec208d5bc0f1663aeb907e40b50f8` generalize to a fresh balanced swamp holdout?OPEN LABNOTE ↗
2. zenith-0162026-08-25 · positivesnowy Hero development foldDoes adding independently collected Hero-only Hoelbrak development data repair the snowy Hero weakness exposed by Labnote 015 without regressing Inventory?OPEN LABNOTE ↗
3. zenith-0152026-08-25 · mixedsnowy localized panel holdoutDoes deterministic localized candidate `bc155807a28aee1a2d7d10fb18a96b24d40b9fe300fc3e01650f70e5d61cc649` generalize to a fresh balanced holdout captured during the snowy Claw of Jormag encounter?OPEN LABNOTE ↗
4. zenith-0142026-08-25 · positivedeterministic localized transferDoes deterministic, class-balanced full-batch multiple-instance training retain Hero generalization after adding varied Inventory-only position data?OPEN LABNOTE ↗
5. zenith-0132026-08-25 · mixedvaried Inventory position transferDoes adding new Inventory-only development batches staged at left, right, and center positions repair the localized candidate's position dependence?OPEN LABNOTE ↗
6. zenith-0122026-08-25 · negativelocalized panel frozen holdoutDoes the localized panel candidate that passed development cross-validation generalize to a separately frozen, balanced live Small-UI holdout?OPEN LABNOTE ↗
7. zenith-0112026-08-24 · positivelocalized panel transferCan class-specific localized evidence avoid the environmental confounding seen by whole-tile transfer learning while preserving development-only evaluation?OPEN LABNOTE ↗
8. zenith-0102026-08-24 · negativecorrective panel transferDoes adding an approved Inventory-only and closed-panel development batch fix the cross-batch false-positive behavior of the frozen MobileNet V3 Small panel classifier without spending a fresh holdout?OPEN LABNOTE ↗
9. zenith-0092026-08-24 · negativepanel transfer baselineDoes a frozen ImageNet-pretrained MobileNet V3 Small backbone provide enough visual structure to recognize Hero and Inventory across development batches without spending another holdout?OPEN LABNOTE ↗
10. zenith-0082026-08-24 · negativefirst visual panel baselineCan a compact CPU ridge classifier trained on masked Small-UI panel tiles generalize from two development batches to a recognition-disabled frozen holdout captured during a live world boss?OPEN LABNOTE ↗
11. zenith-0072026-08-24 · mixedSmall-UI panel recognitionCan Zenith Vision recognize a concrete open Guild Wars 2 UI panel in a messy live scene without retaining the full frame or emitting unrelated screen text?OPEN LABNOTE ↗
12. zenith-0062026-08-24 · positivelive one-shot observationCan Zenith Vision execute its complete live privacy and authority boundary once against Guild Wars 2 without allowing an optional OCR failure to suppress valid MumbleLink telemetry?OPEN LABNOTE ↗
13. zenith-0052026-08-24 · positiveNormal profile blind human validationDoes the frozen Normal UI crop profile agree with independently drawn human boxes on fresh gameplay frames?OPEN LABNOTE ↗
14. zenith-0042026-08-24 · mixedfrozen Normal profile seeded validationDoes the frozen Normal UI crop profile remain visually acceptable across a fresh, independently collected gameplay batch?OPEN LABNOTE ↗
15. zenith-0032026-08-24 · negativeUI-scale stratificationDoes one generic default-layout geometry describe the visible Guild Wars 2 HUD across the Small, Normal, and Large UI profiles?OPEN LABNOTE ↗
16. zenith-0022026-08-24 · negativebatch holdout geometry confirmationDoes the default-layout geometry baseline recover operator-reviewed visible HUD regions across a larger private gameplay holdout?OPEN LABNOTE ↗
17. zenith-0012026-08-24 · negativedefault-layout geometry on a real holdoutCan the deterministic default-layout proposal serve as a UI-region detector, or is it only suitable for conservative crop routing?OPEN LABNOTE ↗

No matching labnotes.