| # Stoicheia — Released Models and Datasets |
|
|
| All models are character-level masked-diffusion Transformers for Ancient Greek (405M |
| params, `d_model=1024`, depth 32), loadable via |
| `AutoModel.from_pretrained(repo_id, trust_remote_code=True)`. See each model card for |
| usage examples specific to that checkpoint's task. |
|
|
| ## Pretrained backbones (11) |
|
|
| Ten rotated, work-level-decontaminated literary folds plus one documentary-clean model. |
| For any passage in the open training corpus, at least one of these eleven has provably |
| never seen it during pretraining. |
|
|
| | model | decontamination | |
| |---|---| |
| | [`Ericu950/Stoicheia-doc_clean`](https://huggingface.co/Ericu950/Stoicheia-doc_clean) | zero documentary (inscription/papyrus) exposure of any kind | |
| | [`Ericu950/Stoicheia-fold-0`](https://huggingface.co/Ericu950/Stoicheia-fold-0) … [`fold-9`](https://huggingface.co/Ericu950/Stoicheia-fold-9) | rotated 80/10/10 literary split, fold *k*'s test set unseen by fold *k*'s model | |
|
|
| ## Fine-tuned downstream models (4) |
|
|
| All built on `Stoicheia-doc_clean`. |
|
|
| | model | task | |
| |---|---| |
| | [`Ericu950/Stoicheia-restoration-test0`](https://huggingface.co/Ericu950/Stoicheia-restoration-test0) … [`-test9`](https://huggingface.co/Ericu950/Stoicheia-restoration-test9) | documentary restoration, ten checkpoints — checkpoint *k* holds out every PHI/TM identifier ending in digit *k*, so every document in the corpus has a model that provably never saw it | |
| | [`Ericu950/Stoicheia-tagger-parser`](https://huggingface.co/Ericu950/Stoicheia-tagger-parser) | morphosyntactic tagging (XPOS/UPOS/lemma) + dependency parsing | |
| | [`Ericu950/Stoicheia-meter`](https://huggingface.co/Ericu950/Stoicheia-meter) | macronization (vowel length) + metrical scansion, trained jointly | |
| | [`Ericu950/Stoicheia-macronizer`](https://huggingface.co/Ericu950/Stoicheia-macronizer) | macronization only — the arm the paper's macronization ablation is measured on | |
|
|
| ## Datasets |
|
|
| | dataset | what it is | |
| |---|---| |
| | [`Ericu950/AncientGreek`](https://huggingface.co/datasets/Ericu950/AncientGreek) | pretraining corpus, ~361M words in `pristine` and `repaired` tiers | |
| | [`Ericu950/SyntheticAncientGreek-CorpusCorporum`](https://huggingface.co/datasets/Ericu950/SyntheticAncientGreek-CorpusCorporum) | bronze synthetic augmentation tier, machine-translated from Latin | |
| | [`Ericu950/Inscriptions_2`](https://huggingface.co/datasets/Ericu950/Inscriptions_2) | PHI inscriptions used for the 10-fold split and restoration fine-tuning | |
| | [`Ericu950/Stoicheia-meter-silver`](https://huggingface.co/datasets/Ericu950/Stoicheia-meter-silver) | silver macronization/scansion training data (Hypotactic-derived + constraint-solver mined) | |
| | [`Ericu950/norma`](https://huggingface.co/datasets/Ericu950/norma) | mirror of the *Norma Syllabarum Graecarum* benchmark, in the exact split the paper evaluates on | |
|
|
| `AncientGreek` omits the Database of Byzantine Book Epigrams, which is CC BY-NC-SA and so |
| cannot travel inside a CC BY-SA compilation; `scripts/fetch_dbbe.py` refetches it under its |
| own terms. Not re-released at all: the OGA/AGDT treebank splits, which come from Celano's |
| own repository and are cited rather than duplicated. |
|
|
