{ "version": "1.0", "truncation": null, "padding": null, "added_tokens": [ { "id": 0, "content": "<|pad|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": false, "special": true }, { "id": 1, "content": "<|start_story|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": false, "special": true }, { "id": 2, "content": "<|end_story|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": false, "special": true }, { "id": 3, "content": "<|unk|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": false, "special": true }, { "id": 4, "content": "[PERSON_NAME_1]", "single_word": false, "lstrip": false, "rstrip": false, "normalized": false, "special": true }, { "id": 5, "content": "[PERSON_NAME_2]", "single_word": false, "lstrip": false, "rstrip": false, "normalized": false, "special": true }, { "id": 6, "content": "[PERSON_NAME_3]", "single_word": false, "lstrip": false, "rstrip": false, "normalized": false, "special": true } ], "normalizer": null, "pre_tokenizer": { "type": "Sequence", "pretokenizers": [ { "type": "Split", "pattern": { "Regex": "\\s+(?=\\[PERSON_NAME_[0-9]+\\])" }, "behavior": "Isolated", "invert": false }, { "type": "Split", "pattern": { "Regex": "(?<=\\])(?=['’]s)" }, "behavior": "Isolated", "invert": false }, { "type": "ByteLevel", "add_prefix_space": false, "trim_offsets": true, "use_regex": true } ] }, "post_processor": null, "decoder": { "type": "ByteLevel", "add_prefix_space": true, "trim_offsets": true, "use_regex": true }, "model": { "type": "WordLevel", "vocab": { "<|pad|>": 0, "<|start_story|>": 1, "<|end_story|>": 2, "<|unk|>": 3, "[PERSON_NAME_1]": 4, "[PERSON_NAME_2]": 5, "[PERSON_NAME_3]": 6, "_": 7, ".": 8, "Ġthe": 9, "NAME": 10, "PERSON": 11, "[": 12, "Ġ": 13, "]": 14, "1": 15, "Ġand": 16, "Ġa": 17, "ĠThey": 18, "Ġto": 19, "ĠThe": 20, "Ġwas": 21, "2": 22, "Ġin": 23, "Ġball": 24, "Ġred": 25, "Ġof": 26, "Ġflower": 27, "Ġlooked": 28, "Ġat": 29, "Ġwater": 30, "Ġkite": 31, "Ġwalked": 32, ",": 33, "Ġbucket": 34, "Ġsmall": 35, "Ġstick": 36, "Ġbench": 37, "Ġinto": 38, "Ġit": 39, "Ġbread": 40, "Ġwanted": 41, "Ġsmiled": 42, "Ġdry": 43, "Ġgrabbed": 44, "Ġreached": 45, "Ġcup": 46, "Ġbook": 47, "Ġtoward": 48, "Ġon": 49, "Ġbirdhouse": 50, "Ġagain": 51, "Ġacross": 52, "Ġboard": 53, "].": 54, "Ġfor": 55, "Ġpaper": 56, "Ġpath": 57, "Ġcarefully": 58, "Ġbasket": 59, "Ġheld": 60, "Ġsaid": 61, "Ġmoved": 62, "Ġvillage": 63, "Ġunder": 64, "Ġover": 65, "Ġonto": 66, "Ġhad": 67, "Ġplaced": 68, "Ġsaw": 69, "Ġyard": 70, "Ġfrom": 71, "Ġstones": 72, "Ġgrass": 73, "Ġcloth": 74, "Ġhappy": 75, "Ġlarge": 76, "Ġgarden": 77, "Ġpatch": 78, "I": 79, "Ġmud": 80, "Ġgap": 81, "Ġwooden": 82, "Ġseveral": 83, "Ġrolled": 84, "Ġstone": 85, "Ġtear": 86, "Ġbright": 87, "Ġvery": 88, "Ġflew": 89, "Ġpoured": 90, "Ġglue": 91, "Ġheavy": 92, "ĠIt": 93, "Ġnails": 94, "Ġsoil": 95, "\"": 96, "Ġfirmly": 97, "Ġfresh": 98, "Ġhome": 99, "Ġstepped": 100, "Ġout": 101, "Ġdirt": 102, "Ġnear": 103, "Ġsteady": 104, "Ġtightly": 105, "!\"": 106, "Ġknelt": 107, "Ġwith": 108, "Ġhouse": 109, "Ġwere": 110, "Ġground": 111, "ĠFinally": 112, "Ġfelt": 113, "Ġwent": 114, "Ġback": 115, "Ġpushed": 116, "Ġtrips": 117, "Ġbecause": 118, "Ġpressed": 119, "Ġtoo": 120, "Ġaround": 121, "Ġgot": 122, "Ġgripped": 123, "Ġran": 124, "Ġup": 125, ",\"": 126, "The": 127, "Ġone": 128, "Ġsun": 129, "Ġarrived": 130, "Ġbut": 131, "Ġlong": 132, "Ġnoticed": 133, "Ġfinished": 134, "Ġthrough": 135, "Ġcarried": 136, "Ġtree": 137, "Ġpark": 138, "Ġenough": 139, "Ġdipped": 140, "Ġsafely": 141, "Ġwood": 142, "Ġ\"": 143, "Ġbakery": 144, "Ġfound": 145, "Ġsky": 146, "Ġhigh": 147, "Ġmade": 148, "Ġtogether": 149, "Ġsat": 150, "Ġhole": 151, "Ġmuddy": 152, "Ġdark": 153, "Ġdoor": 154, "Ġkept": 155, "Ġbag": 156, "Ġedge": 157, "Ġschool": 158, "Ġtheir": 159, "Ġflat": 160, "Ġpiece": 161, "Ġtapped": 162, "Ġwant": 163, "Ġwatched": 164, "Ġnot": 165, "Ġsome": 166, "Ġwind": 167, "Ġfull": 168, "Ġpetals": 169, "Ġbalanced": 170, "Ġbuild": 171, "Ġreceived": 172, "Ġthirsty": 173, "Ġworked": 174, "Ġfix": 175, "Ġdown": 176, "Ġfly": 177, "Ġsad": 178, "Ġcity": 179, "Ġhelp": 180, "Ġsturdy": 181, "Ġtook": 182, "Ġhammer": 183, "Ġreach": 184, "Ġthick": 185, "Ġthis": 186, "Ġbottom": 187, "Ġfirst": 188, "Ġspace": 189, "Ġuntil": 190, "Ġwaited": 191, "Ġwrapped": 192, "Ġalong": 193, "Ġmany": 194, "Ġside": 195, "Ġnearby": 196, "Ġreturn": 197, "Ġfar": 198, "Ġplant": 199, "Ġsighed": 200, "Ġwarm": 201, "Ġclean": 202, "Ġhanded": 203, "Ġmiddle": 204, "Ġaway": 205, "Ġfalling": 206, "Ġinside": 207, "Ġspotted": 208, "Ġtried": 209, "Ġstopped": 210, "Ġslowly": 211, "Ġstring": 212, "Ġwoods": 213, "Ġagainst": 214, "Ġcolorful": 215, "Ġother": 216, "Ġblue": 217, "Ġby": 218, "ĠA": 219, "Ġlift": 220, "Ġpieces": 221, "Ġwork": 222, "Ġfall": 223, "Ġpicked": 224, "ĠHowever": 225, "Ġforward": 226, "Ġhands": 227, "Ġbrought": 228, "Ġloaves": 229, "Ġthem": 230, "Ġthey": 231, "Ġanother": 232, "Ġcan": 233, "Ġcarry": 234, "Ġstayed": 235, "Ġreturned": 236, "Ġshouted": 237, "Ġstrong": 238, "Ġready": 239, "Ġtime": 240, "Ġbirds": 241, "Ġbring": 242, "Ġnarrow": 243, "Ġgood": 244, "Ġkeep": 245, "Ġneeded": 246, "Ġnext": 247, "Ġopened": 248, "Ġdid": 249, "Ġput": 250, "You": 251, "Ġforth": 252, "Ġmore": 253, "Ġstood": 254, "Ġboth": 255, "Ġfence": 256, "Ġwould": 257, "Ġcould": 258, "Ġperfectly": 259, "Ġsecure": 260, "ĠThen": 261, "Ġadded": 262, "Ġdeep": 263, "Ġloaf": 264, "Ġmy": 265, "Ġfriend": 266, "ĠSuddenly": 267, "Ġgrip": 268, "Ġstep": 269, "Ġplay": 270, "Ġrepeated": 271, "Ġyou": 272, "Ġcaught": 273, "Ġfixed": 274, "Ġtimes": 275, "Ġway": 276, "Ġbird": 277, "Ġthen": 278, "Ġtight": 279, "Ġtorn": 280, "'s": 281, "Ġget": 282, "Ġnail": 283, "Ġopen": 284, "ĠAfter": 285, "Ġcups": 286, "Ġfell": 287, "Ġhave": 288, "Ġhold": 289, "Ġlifted": 290, "Ġmake": 291, "A": 292, "We": 293, "Ġ\"[": 294, "Ġempty": 295, "Ġstuck": 296, "Ġwell": 297, "Ġfilled": 298, "Ġhelped": 299, "Ġstay": 300, "Ġdone": 301, "Ġfabric": 302, "Ġlittle": 303, "Ġnew": 304, "Ġplace": 305, "Ġproud": 306, "Ġquickly": 307, "Ġbalance": 308, "Ġcheered": 309, "Ġspread": 310, "Ġsteps": 311, "Ġgave": 312, "Ġthat": 313, "Ġthere": 314, "],": 315, "Ġafter": 316, "Ġbetween": 317, "Ġbrown": 318, "Ġedges": 319, "Ġnow": 320, "Ġsee": 321, "Ġwhile": 322, "Ġas": 323, "Ġdo": 324, "Ġdrop": 325, "Ġis": 326, "Ġlonger": 327, "Ġmove": 328, "Ġno": 329, "Ġstill": 330, "It": 331, "ĠI": 332, "Ġfind": 333, "Ġits": 334, "Ġme": 335, "Ġneed": 336, "Ġroll": 337, "Ġwhen": 338, "!": 339, ".\"": 340, "?\"": 341, "Yes": 342, "Ġfinally": 343, "Ġgive": 344, "Ġyour": 345, "Look": 346, "Ġdinner": 347, "Ġevery": 348, "Ġgo": 349, "Ġlying": 350, "Ġstand": 351, "Ġwalk": 352, "Can": 353, "Is": 354, "Where": 355, "ĠBut": 356, "ĠIts": 357, "ĠMany": 358, "ĠNow": 359, "ĠOne": 360, "ĠSlowly": 361, "ĠThere": 362, "ĠThis": 363, "Ġam": 364, "Ġare": 365, "Ġbe": 366, "Ġbefore": 367, "Ġbegan": 368, "Ġlook": 369, "Ġmuch": 370, "Ġpush": 371, "Ġsmile": 372, "Ġso": 373, "Ġtap": 374, "Ġthemselves": 375, "Ġwatered": 376, "Ġworks": 377 }, "unk_token": "<|unk|>" } }