Download language-rules.json from proofbird/gector-bert-it: direct link, hf CLI and curl.
- Browser
- Download file 829 Bytes
-
https://huggingface.co/proofbird/gector-bert-it/resolve/main/language-rules.json
- Command line
-
hf download hf://proofbird/gector-bert-it/language-rules.json
-
curl -L -o language-rules.json https://huggingface.co/proofbird/gector-bert-it/resolve/main/language-rules.json
829 Bytes
| { | |
| "version": 1, | |
| "language": "it", | |
| "_comment": "Per-bundle language facts (data-driven; see C# LanguageRules / ADR-004). stripTrailingTerminator: strip a single trailing . ! ? before the WordPiece model and re-attach it after (the Italian bundle is brittle to a trailing period). elision.nonElidingArticles: full (un-elided) articles the tokenizer treats as glued-elision prefixes (gli'avvocato -> gli avvocato); overrides ItalianElision.DefaultNonElidingArticles. Absent file => None (stripTerminator false, default article list).", | |
| "stripTrailingTerminator": true, | |
| "elision": { | |
| "nonElidingArticles": ["gli", "le", "i", "il", "la", "lo", "una", "uno"] | |
| }, | |
| "clauseConjunctions": ["ma", "e", "ed", "o", "oppure", "però", "mentre", "quindi", "poi", "perché", "anzi", "dunque", "infatti", "tuttavia", "perciò"] | |
| } | |