Benchmarks
45 benchmarks in 8 domains
Single-cell
Protocol
3 models
| # | Model ↕ | Mean rank ▲ | Coverage ↕ | Cell type annotation (macro F1) ↕ | Perturbation response (Pearson on Δ expression) ↕ | Single-cell batch integration (scIB) ↕ |
|---|---|---|---|---|---|---|
| Low coverage: below 2 of 3, or nothing to compare with. Listed, not ranked. | ||||||
| — | single-cell/gears | 0.00 | 1/3 | · | 0.543 (1) | · |
| — | single-cell/cpa | 1.00 | 1/3 | · | 0.396 (2) | · |
| — | single-cell/scvi-tools | — | 2/3 | 0.714 (1) | · | 0.610 (1) |
| Reference floors: trivial baselines on the same items, not hub models, unranked. | ||||||
| — | perturbation-additive floor | · | 0.399 | · | ||
| — | perturbation-no-change floor | · | 0.016 | · | ||
| — | perturbation-train-mean floor | · | 0.576 | · | ||
No model matches these filters.
Each cell is the model's value on that benchmark's board, with its rank there in brackets, as the benchmark's own leaderboard ranks it (its primary supervised metric, at its default version); hover for the 95% interval, n and source, click to open that board. A * marks a declared train/test overlap. Rows are ordered by mean normalised rank, (rank − 1) / (board size − 1) averaged over the benchmarks the model has a result on (a board of one entry does not count): ranks, never values, so no average mixes units. Filtering hides rows without re-ranking anyone.