Token Classification
GLiNER
PyTorch
multilingual
Edit model card

Finetuned for way too short on the following PII categories:

PII_CATEGORIES = [
    "prefix", "person name",
    "job title", "company name", "job area",
    "email address",
    "account number", "address", "city", "postal code",
    "password", "job type", "state",
    "county",
    "phone number",
    "personal identification number",
    "gender", "biological sex",
    "religious belief", "nationality",
    "sexual orientation", "medical condition", "age",
]

On these languages (the ones contained in Whisper):

LANGUAGES = {
    "en": "english",
    "zh": "chinese",
    "de": "german",
    "es": "spanish",
    "ru": "russian",
    "ko": "korean",
    "fr": "french",
    "ja": "japanese",
    "pt": "portuguese",
    "tr": "turkish",
    "pl": "polish",
    "ca": "catalan",
    "nl": "dutch",
    "ar": "arabic",
    "sv": "swedish",
    "it": "italian",
    "id": "indonesian",
    "hi": "hindi",
    "fi": "finnish",
    "vi": "vietnamese",
    "he": "hebrew",
    "uk": "ukrainian",
    "el": "greek",
    "ms": "malay",
    "cs": "czech",
    "ro": "romanian",
    "da": "danish",
    "hu": "hungarian",
    "ta": "tamil",
    "no": "norwegian",
    "th": "thai",
    "ur": "urdu",
    "hr": "croatian",
    "bg": "bulgarian",
    "lt": "lithuanian",
    "la": "latin",
    "mi": "maori",
    "ml": "malayalam",
    "cy": "welsh",
    "sk": "slovak",
    "te": "telugu",
    "fa": "persian",
    "lv": "latvian",
    "bn": "bengali",
    "sr": "serbian",
    "az": "azerbaijani",
    "sl": "slovenian",
    "kn": "kannada",
    "et": "estonian",
    "mk": "macedonian",
    "br": "breton",
    "eu": "basque",
    "is": "icelandic",
    "hy": "armenian",
    "ne": "nepali",
    "mn": "mongolian",
    "bs": "bosnian",
    "kk": "kazakh",
    "sq": "albanian",
    "sw": "swahili",
    "gl": "galician",
    "mr": "marathi",
    "pa": "punjabi",
    "si": "sinhala",
    "km": "khmer",
    "sn": "shona",
    "yo": "yoruba",
    "so": "somali",
    "af": "afrikaans",
    "oc": "occitan",
    "ka": "georgian",
    "be": "belarusian",
    "tg": "tajik",
    "sd": "sindhi",
    "gu": "gujarati",
    "am": "amharic",
    "yi": "yiddish",
    "lo": "lao",
    "uz": "uzbek",
    "fo": "faroese",
    "ht": "haitian creole",
    "ps": "pashto",
    "tk": "turkmen",
    "nn": "nynorsk",
    "mt": "maltese",
    "sa": "sanskrit",
    "lb": "luxembourgish",
    "my": "myanmar",
    "bo": "tibetan",
    "tl": "tagalog",
    "mg": "malagasy",
    "as": "assamese",
    "tt": "tatar",
    "haw": "hawaiian",
    "ln": "lingala",
    "ha": "hausa",
    "ba": "bashkir",
    "jw": "javanese",
    "su": "sundanese",
    "yue": "cantonese",
}
Downloads last month
5
Inference Examples
Inference API (serverless) does not yet support gliner models for this pipeline type.

Datasets used to train thegenerativegeneration/gliner-finetune-small-v2.5