File size: 8,774 Bytes
754c9c1 |
1 |
{ "version": "1.0", "truncation": null, "padding": null, "added_tokens": [ { "id": 0, "content": "[STOP]", "single_word": false, "lstrip": false, "rstrip": false, "normalized": false, "special": true }, { "id": 1, "content": "[UNK]", "single_word": false, "lstrip": false, "rstrip": false, "normalized": false, "special": true }, { "id": 2, "content": "[SPACE]", "single_word": false, "lstrip": false, "rstrip": false, "normalized": false, "special": true } ], "normalizer": null, "pre_tokenizer": { "type": "ByteLevel", "add_prefix_space": true, "trim_offsets": true, "use_regex": true }, "post_processor": null, "decoder": null, "model": { "type": "BPE", "dropout": null, "unk_token": null, "continuing_subword_prefix": null, "end_of_word_suffix": null, "fuse_unk": false, "byte_fallback": false, "vocab": { "[STOP]": 0, "[UNK]": 1, "[SPACE]": 2, "!": 3, "'": 4, "(": 5, ")": 6, ",": 7, "-": 8, ".": 9, "/": 10, ":": 11, ";": 12, "?": 13, "a": 14, "b": 15, "c": 16, "d": 17, "e": 18, "f": 19, "g": 20, "h": 21, "i": 22, "j": 23, "k": 24, "l": 25, "m": 26, "n": 27, "o": 28, "p": 29, "r": 30, "s": 31, "t": 32, "u": 33, "v": 34, "w": 35, "y": 36, "z": 37, "Ġ": 38, "de": 39, "Ġs": 40, "en": 41, "Ġh": 42, "er": 43, "Ġde": 44, "Ġo": 45, "an": 46, "Ġm": 47, "Ġv": 48, "Ġog": 49, "ar": 50, "or": 51, "Ġf": 52, "Ġi": 53, "Ġa": 54, "et": 55, "Ġk": 56, "Ġt": 57, "Ġb": 58, "Ġhan": 59, "te": 60, "ne": 61, "le": 62, "Ġl": 63, "ig": 64, "il": 65, "Ġsa": 66, "Ġg": 67, "Ġp": 68, "un": 69, "ke": 70, "in": 71, "Ġvar": 72, "Ġd": 73, "Ġdet": 74, "Ġen": 75, "Ġn": 76, "Ġfor": 77, "re": 78, "Ġat": 79, "ra": 80, "Ġha": 81, "ed": 82, "gen": 83, "el": 84, "Ġtil": 85, "dt": 86, "kke": 87, "Ġaf": 88, "Ġsk": 89, "Ġpa": 90, "om": 91, "ge": 92, "Ġder": 93, "is": 94, "Ġj": 95, "aer": 96, "Ġst": 97, "ede": 98, "es": 99, "Ġikke": 100, "Ġden": 101, "den": 102, "Ġmen": 103, "ol": 104, "og": 105, "st": 106, "Ġhav": 107, "erne": 108, "Ġsig": 109, "ik": 110, "Ġhv": 111, "Ġla": 112, "Ġmed": 113, "al": 114, "eg": 115, "der": 116, "at": 117, "Ġhavde": 118, "od": 119, "ud": 120, "Ġr": 121, "Ġin": 122, "ul": 123, "Ġal": 124, "ag": 125, "Ġer": 126, "ing": 127, "Ġhun": 128, "Ġkun": 129, "Ġkunne": 130, "ver": 131, "mm": 132, "Ġdu": 133, "on": 134, "ek": 135, "ste": 136, "ere": 137, "and": 138, "ille": 139, "Ġet": 140, "Ġud": 141, "Ġham": 142, "ov": 143, "ende": 144, "lig": 145, "Ġom": 146, "ru": 147, "ter": 148, "Ġma": 149, "Ġkom": 150, "Ġe": 151, "Ġop": 152, "ang": 153, "Ġda": 154, "Ġsp": 155, "la": 156, "lev": 157, "ro": 158, "ten": 159, "Ġind": 160, "Ġsi": 161, "Ġsom": 162, "iv": 163, "idt": 164, "Ġne": 165, "Ġjor": 166, "ler": 167, "pp": 168, "Ġbl": 169, "ene": 170, "Ġdem": 171, "tede": 172, "je": 173, "sk": 174, "Ġno": 175, "age": 176, "yr": 177, "Ġfra": 178, "Ġblev": 179, "ej": 180, "yn": 181, "Ġun": 182, "Ġvaer": 183, "Ġnog": 184, "ch": 185, "Ġover": 186, "Ġman": 187, "gde": 188, "Ġhar": 189, "ad": 190, "Ġhans": 191, "ev": 192, "Ġra": 193, "andre": 194, "Ġsch": 195, "Ġved": 196, "Ġtra": 197, "lede": 198, "Ġjorgen": 199, "Ġka": 200, "Ġhvor": 201, "us": 202, "Ġvi": 203, "Ġned": 204, "Ġandre": 205, "Ġor": 206, "bl": 207, "Ġsl": 208, "Ġigen": 209, "inde": 210, "for": 211, "ket": 212, "tet": 213, "ur": 214, "dan": 215, "ger": 216, "Ġmor": 217, "Ġjeg": 218, "ab": 219, "ker": 220, "Ġel": 221, "Ġsagde": 222, "Ġnu": 223, "Ġskul": 224, "yg": 225, "Ġfa": 226, "fter": 227, "mp": 228, "Ġre": 229, "Ġsamm": 230, "dte": 231, "Ġgr": 232, "Ġtilb": 233, "Ġnoget": 234, "se": 235, "ss": 236, "Ġskov": 237, "Ġefter": 238, "det": 239, "Ġsv": 240, "Ġhol": 241, "Ġhende": 242, "Ġmod": 243, "Ġkan": 244, "Ġlu": 245, "ige": 246, "Ġtilbage": 247, "jor": 248, "Ġva": 249, "Ġville": 250, "Ġti": 251, "Ġlo": 252, "Ġgik": 253, "ning": 254 }, "merges": [ "d e", "Ġ s", "e n", "Ġ h", "e r", "Ġ de", "Ġ o", "a n", "Ġ m", "Ġ v", "Ġo g", "a r", "o r", "Ġ f", "Ġ i", "Ġ a", "e t", "Ġ k", "Ġ t", "Ġ b", "Ġh an", "t e", "n e", "l e", "Ġ l", "i g", "i l", "Ġs a", "Ġ g", "Ġ p", "u n", "k e", "i n", "Ġv ar", "Ġ d", "Ġde t", "Ġ en", "Ġ n", "Ġf or", "r e", "Ġa t", "r a", "Ġh a", "e d", "g en", "e l", "Ġt il", "d t", "k ke", "Ġa f", "Ġs k", "Ġp a", "o m", "g e", "Ġde r", "i s", "Ġ j", "a er", "Ġs t", "e de", "e s", "Ġi kke", "Ġde n", "de n", "Ġm en", "o l", "o g", "s t", "Ġha v", "er ne", "Ġs ig", "i k", "Ġh v", "Ġl a", "Ġm ed", "a l", "e g", "de r", "a t", "Ġhav de", "o d", "u d", "Ġ r", "Ġi n", "u l", "Ġa l", "a g", "Ġ er", "in g", "Ġh un", "Ġk un", "Ġkun ne", "v er", "m m", "Ġd u", "o n", "e k", "s te", "er e", "an d", "il le", "Ġ et", "Ġ ud", "Ġha m", "o v", "en de", "l ig", "Ġo m", "r u", "t er", "Ġm a", "Ġk om", "Ġ e", "Ġo p", "an g", "Ġd a", "Ġs p", "l a", "le v", "r o", "t en", "Ġin d", "Ġs i", "Ġs om", "i v", "i dt", "Ġ ne", "Ġj or", "l er", "p p", "Ġb l", "en e", "Ġde m", "te de", "j e", "s k", "Ġn o", "a ge", "y r", "Ġf ra", "Ġb lev", "e j", "y n", "Ġ un", "Ġv aer", "Ġn og", "c h", "Ġo ver", "Ġm an", "g de", "Ġh ar", "a d", "Ġhan s", "e v", "Ġ ra", "and re", "Ġs ch", "Ġv ed", "Ġt ra", "le de", "Ġjor gen", "Ġk a", "Ġhv or", "u s", "Ġv i", "Ġne d", "Ġ andre", "Ġo r", "b l", "Ġs l", "Ġi gen", "in de", "f or", "k et", "t et", "u r", "d an", "g er", "Ġm or", "Ġj eg", "a b", "k er", "Ġ el", "Ġsa gde", "Ġn u", "Ġsk ul", "y g", "Ġf a", "f ter", "m p", "Ġ re", "Ġsa mm", "d te", "Ġg r", "Ġtil b", "Ġnog et", "s e", "s s", "Ġsk ov", "Ġe fter", "de t", "Ġs v", "Ġh ol", "Ġh ende", "Ġm od", "Ġk an", "Ġl u", "ig e", "Ġtilb age", "j or", "Ġv a", "Ġv ille", "Ġt i", "Ġl o", "Ġg ik", "n ing" ] } } |