---
tags:
- spacy
- token-classification
- ancient-greek
language:
- grc
license: mit
model-index:
- name: grc_dep_web_trf
  results:
  - task:
      name: POS Tagging
      type: token-classification
    metrics:
    - name: POS Accuracy
      type: accuracy
      value: 0.9728
    - name: TAG (XPOS) Accuracy
      type: accuracy
      value: 0.9740
  - task:
      name: Lemmatization
      type: token-classification
    metrics:
    - name: Lemma Accuracy
      type: accuracy
      value: 0.9399
  - task:
      name: Dependency Parsing
      type: token-classification
    metrics:
    - name: Labeled Attachment Score
      type: f_score
      value: 0.8027
---

# grc_dep_web_trf

**Ancient Greek** pipeline for [spaCy](https://spacy.io), part of the [LatinCy](https://huggingface.co/latincy) project.

**Experimental beta release.** This is part of the first generation of Ancient Greek models porting the [LatinCy](https://huggingface.co/latincy) Latin pipeline infrastructure to Ancient Greek. Expect rough edges; scores and component behavior will improve as training data is harmonized and curated through the LatinCy flywheel (train, evaluate, curate, retrain).

Transformer model powered by PhilBerta (Ancient Greek RoBERTa). Trained on Universal Dependencies Ancient Greek treebanks (PTNK, PROIEL, Perseus) with a 1.2M-entry lookup lemmatizer overlay built from CLTK Morpheus, UD treebanks, and Wiktionary.

| Feature | Description |
| --- | --- |
| **Name** | `grc_dep_web_trf` |
| **Version** | `3.8.1` |
| **spaCy** | `>=3.8.11,<3.9.0` |
| **Default Pipeline** | `senter`, `transformer`, `tagger`, `morphologizer`, `trainable_lemmatizer`, `lookup_lemmatizer`, `parser` |
| **Components** | `senter`, `transformer`, `tagger`, `morphologizer`, `trainable_lemmatizer`, `lookup_lemmatizer`, `parser` |
| **Vectors** | 0 keys, 0 unique vectors (0 dimensions) |
| **License** | `MIT` |
| **Author** | [Patrick J. Burns](https://huggingface.co/latincy) |

## Install

```bash
pip install https://huggingface.co/latincy/grc_dep_web_trf/resolve/main/grc_dep_web_trf-3.8.1-py3-none-any.whl
```

## Usage

```python
import spacy

nlp = spacy.load("grc_dep_web_trf")
doc = nlp("\u03bc\u1fc6\u03bd\u03b9\u03bd \u1f04\u03b5\u03b9\u03b4\u03b5 \u03b8\u03b5\u1f70 \u03a0\u03b7\u03bb\u03b7\u03ca\u03ac\u03b4\u03b5\u03c9 \u1f08\u03c7\u03b9\u03bb\u1fc6\u03bf\u03c2")

for token in doc:
    print(token.text, token.pos_, token.lemma_, token.dep_)
```

## Evaluation

Scores on held-out UD test data (combined PTNK + PROIEL + Perseus).

| Metric | Score |
| --- | --- |
| **POS (UPOS) Accuracy** | 97.28 |
| **TAG (XPOS) Accuracy** | 97.40 |
| **Morph (UFeats) Accuracy** | 93.61 |
| **Lemma Accuracy** | 93.99 |
| **Unlabeled Attachment Score (UAS)** | 85.12 |
| **Labeled Attachment Score (LAS)** | 80.27 |
| **Sentences F-Score** | 88.18 |

## Training Data

| Source | Description |
| --- | --- |
| [UD_Ancient_Greek-PTNK](https://github.com/UniversalDependencies/UD_Ancient_Greek-PTNK) | Septuagint (Codex Alexandrinus) |
| [UD_Ancient_Greek-PROIEL](https://github.com/UniversalDependencies/UD_Ancient_Greek-PROIEL) | PROIEL Ancient Greek treebank |
| [UD_Ancient_Greek-Perseus](https://github.com/UniversalDependencies/UD_Ancient_Greek-Perseus) | Perseus Ancient Greek treebank |

## Components

- **transformer** -- PhilBerta transformer backbone (Ancient Greek RoBERTa)
- **tagger** -- Fine-grained POS tagger (XPOS, harmonized 16-tag tagset)
- **morphologizer** -- Morphological feature assignment (UPOS + UFeats)
- **trainable_lemmatizer** -- Edit-tree lemmatizer
- **lookup_lemmatizer** -- 1.2M-entry dictionary lemmatizer overlay (CLTK Morpheus + UD + Wiktionary); normalizes grave accents to acute at query time
- **parser** -- Dependency parser (transition-based)
- **senter** -- Sentence segmenter

## Label Scheme

<details>

<summary>View label scheme (1796 labels for 3 components)</summary>

**`tagger`**: `adjective`, `adverb`, `conjunction`, `conjunction_adverb`, `conjunction_pronoun`, `determiner`, `interjection`, `noun`, `number`, `particle`, `preposition`, `pronoun`, `proper_noun`, `punc`, `unknown`, `verb`

**`morphologizer`**: 1749 morphological feature combinations

**`parser`**: `ROOT`, `acl`, `advcl`, `advmod`, `amod`, `appos`, `aux`, `case`, `cc`, `ccomp`, `conj`, `cop`, `csubj`, `dep`, `det`, `discourse`, `dislocated`, `fixed`, `flat`, `iobj`, `mark`, `nmod`, `nsubj`, `nummod`, `obj`, `obl`, `orphan`, `parataxis`, `punct`, `vocative`, `xcomp`

</details>