Dedup single-fit GBDTs in live-recompute path (slow path)

%↗ over XGBoost: baseline is xgboost_ensemble (the canonical "XGBoost")

Add static/ assets and minor JS additions

KPI cards as native HTML row above the forest plot (responsive flexbox)

Table leading-group: use Bars rule (all-pairs Wilcoxon-Holm + greedy walk) so the two views agree

Re-enable Win-rate cache fast-path + loosen slider-equality check (was missing every metric switch on HF)

Cache forest plots in figures_cache (fixes 30s Bars view load on HF)

Rebuild figures_cache.json (was stale from 3-cluster collapse)

README: TabBench (capitalisation)

Surface 'classification' scope: title, subtitle, intro, methodology, perfmap, tree

Rebrand a bit the preview

Refresh map prose + 04.B trends (FM2 only) + drop em-dashes for colons

Force light mode at page load (head script) — dark toggle commented out earlier was not enough

Refit PCA on display models only; collapse to 3 clusters; honest decision tree

Add statsmodels to requirements (needed for forest plot Holm correction)

Cluster renames (3-line labels) + perfmap polish (axes, hulls, imbalance, win-rate squares); remove Z-Accuracy

Significance forest plot + TabPFN v3 fix + perfmap regen + table polish

Add TabPFN v3 (fixed) + extras: 189-dataset base, TabArena filter, win-rate cache nesting

Tree boxes in pixel-mode + perfmap calibration overlays removed + dark-mode bar-plot resize fix

TabBench v2: GBDT ensembles, perfmap refresh, dark mode, win-rate view

Performance Map view + 186-dataset base + multi-pass curation

Expand leaderboard mix to 171 curated datasets

Restrict leaderboard to canonical 165-dataset base + mean-GBDT imputation

Refresh leaderboard: new metrics, fixed dataset count, Seldon callout

Fix HF Spaces scrolling: counter Gradio's SET_SCROLLING:false postMessage

Fix HF Spaces page-height: let body grow to content, drop height caps

Table view polish: gold/silver/bronze shimmer + wider sweep + all 8 metrics

Display names + drop nicl and sap models

Verticals filter, table view, methodology, GBDT imputation for missing cells

Pixel-art bars + per-class palette on home leaderboard

Switch academic leaderboard to per-fold results + neuralk restyle

updated xgboost metrics

added pct_improvement metrics and updated space

fixed typo and plots

fixed typo and content

Update app.py

Update app.py

updated styling

Merge branch 'style/iteration'

updated values and added public enterprise data

updated academic scores

working on new presentation of results

fixed typo and plot style

Merge branch 'main' of https://huggingface.co/spaces/Neuralk-AI/tabbench

updated tabbench space with up-to-date data

remove data mapping button

Change data

Update finance

Update

Update plot