AICY hosts the data and models behind CYTransformer — a transformer that generates Fine, Regular, Star Triangulations (FRSTs) of reflexive polytopes, the combinatorial data of smooth Calabi-Yau threefolds. The model learns FRSTs from the training datasets below; downloadable pretrained weights will follow.
Training datasets
Favorable reflexive polytopes paired with reference FRSTs, pre-split into train / validation / test.
Every record is plain JSON: a polytope is [POLYID, DRESVERTS] and a triangulation is
[POLYID, TRIANG]. They ship inside the
CYTransformer repository under
datasets/, ready for cyt-train.
| Configuration | h1,1 | Train | Validation | Test |
|---|---|---|---|---|
| 9 + 1 | 5 | 22,238 | 5,965 | 5,668 |
| 10 + 1 | 6 | 164,172 | 21,084 | 21,101 |
Counts are polytope–FRST pairs. More configurations will be added over time.
Get the datasetsPretrained weights
Coming soon (July 2026) — trained CYTransformer weights you can download and run directly with the software.
Model-generated FRSTs
Coming soon (July 2026) — new FRSTs generated by CYTransformer and automatically validated, a growing catalog of Calabi-Yau data.