Hello zabanshenas
Browse files- config.json +501 -0
- merges.txt +0 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +1 -0
- tf_model.h5 +3 -0
- tokenizer.json +0 -0
- tokenizer_config.json +1 -0
- trainer_state.json +310 -0
- vocab.json +0 -0
config.json
ADDED
|
@@ -0,0 +1,501 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/content/checkpoint-84500",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"gradient_checkpointing": false,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "ace",
|
| 15 |
+
"1": "afr",
|
| 16 |
+
"2": "als",
|
| 17 |
+
"3": "amh",
|
| 18 |
+
"4": "ang",
|
| 19 |
+
"5": "ara",
|
| 20 |
+
"6": "arg",
|
| 21 |
+
"7": "arz",
|
| 22 |
+
"8": "asm",
|
| 23 |
+
"9": "ast",
|
| 24 |
+
"10": "ava",
|
| 25 |
+
"11": "aym",
|
| 26 |
+
"12": "azb",
|
| 27 |
+
"13": "aze",
|
| 28 |
+
"14": "bak",
|
| 29 |
+
"15": "bar",
|
| 30 |
+
"16": "bcl",
|
| 31 |
+
"17": "be-tarask",
|
| 32 |
+
"18": "bel",
|
| 33 |
+
"19": "ben",
|
| 34 |
+
"20": "bho",
|
| 35 |
+
"21": "bjn",
|
| 36 |
+
"22": "bod",
|
| 37 |
+
"23": "bos",
|
| 38 |
+
"24": "bpy",
|
| 39 |
+
"25": "bre",
|
| 40 |
+
"26": "bul",
|
| 41 |
+
"27": "bxr",
|
| 42 |
+
"28": "cat",
|
| 43 |
+
"29": "cbk",
|
| 44 |
+
"30": "cdo",
|
| 45 |
+
"31": "ceb",
|
| 46 |
+
"32": "ces",
|
| 47 |
+
"33": "che",
|
| 48 |
+
"34": "chr",
|
| 49 |
+
"35": "chv",
|
| 50 |
+
"36": "ckb",
|
| 51 |
+
"37": "cor",
|
| 52 |
+
"38": "cos",
|
| 53 |
+
"39": "crh",
|
| 54 |
+
"40": "csb",
|
| 55 |
+
"41": "cym",
|
| 56 |
+
"42": "dan",
|
| 57 |
+
"43": "deu",
|
| 58 |
+
"44": "diq",
|
| 59 |
+
"45": "div",
|
| 60 |
+
"46": "dsb",
|
| 61 |
+
"47": "dty",
|
| 62 |
+
"48": "egl",
|
| 63 |
+
"49": "ell",
|
| 64 |
+
"50": "eng",
|
| 65 |
+
"51": "epo",
|
| 66 |
+
"52": "est",
|
| 67 |
+
"53": "eus",
|
| 68 |
+
"54": "ext",
|
| 69 |
+
"55": "fao",
|
| 70 |
+
"56": "fas",
|
| 71 |
+
"57": "fin",
|
| 72 |
+
"58": "fra",
|
| 73 |
+
"59": "frp",
|
| 74 |
+
"60": "fry",
|
| 75 |
+
"61": "fur",
|
| 76 |
+
"62": "gag",
|
| 77 |
+
"63": "gla",
|
| 78 |
+
"64": "gle",
|
| 79 |
+
"65": "glg",
|
| 80 |
+
"66": "glk",
|
| 81 |
+
"67": "glv",
|
| 82 |
+
"68": "grn",
|
| 83 |
+
"69": "guj",
|
| 84 |
+
"70": "hak",
|
| 85 |
+
"71": "hat",
|
| 86 |
+
"72": "hau",
|
| 87 |
+
"73": "hbs",
|
| 88 |
+
"74": "heb",
|
| 89 |
+
"75": "hif",
|
| 90 |
+
"76": "hin",
|
| 91 |
+
"77": "hrv",
|
| 92 |
+
"78": "hsb",
|
| 93 |
+
"79": "hun",
|
| 94 |
+
"80": "hye",
|
| 95 |
+
"81": "ibo",
|
| 96 |
+
"82": "ido",
|
| 97 |
+
"83": "ile",
|
| 98 |
+
"84": "ilo",
|
| 99 |
+
"85": "ina",
|
| 100 |
+
"86": "ind",
|
| 101 |
+
"87": "isl",
|
| 102 |
+
"88": "ita",
|
| 103 |
+
"89": "jam",
|
| 104 |
+
"90": "jav",
|
| 105 |
+
"91": "jbo",
|
| 106 |
+
"92": "jpn",
|
| 107 |
+
"93": "kaa",
|
| 108 |
+
"94": "kab",
|
| 109 |
+
"95": "kan",
|
| 110 |
+
"96": "kat",
|
| 111 |
+
"97": "kaz",
|
| 112 |
+
"98": "kbd",
|
| 113 |
+
"99": "khm",
|
| 114 |
+
"100": "kin",
|
| 115 |
+
"101": "kir",
|
| 116 |
+
"102": "koi",
|
| 117 |
+
"103": "kok",
|
| 118 |
+
"104": "kom",
|
| 119 |
+
"105": "kor",
|
| 120 |
+
"106": "krc",
|
| 121 |
+
"107": "ksh",
|
| 122 |
+
"108": "kur",
|
| 123 |
+
"109": "lad",
|
| 124 |
+
"110": "lao",
|
| 125 |
+
"111": "lat",
|
| 126 |
+
"112": "lav",
|
| 127 |
+
"113": "lez",
|
| 128 |
+
"114": "lij",
|
| 129 |
+
"115": "lim",
|
| 130 |
+
"116": "lin",
|
| 131 |
+
"117": "lit",
|
| 132 |
+
"118": "lmo",
|
| 133 |
+
"119": "lrc",
|
| 134 |
+
"120": "ltg",
|
| 135 |
+
"121": "ltz",
|
| 136 |
+
"122": "lug",
|
| 137 |
+
"123": "lzh",
|
| 138 |
+
"124": "mai",
|
| 139 |
+
"125": "mal",
|
| 140 |
+
"126": "map-bms",
|
| 141 |
+
"127": "mar",
|
| 142 |
+
"128": "mdf",
|
| 143 |
+
"129": "mhr",
|
| 144 |
+
"130": "min",
|
| 145 |
+
"131": "mkd",
|
| 146 |
+
"132": "mlg",
|
| 147 |
+
"133": "mlt",
|
| 148 |
+
"134": "nan",
|
| 149 |
+
"135": "mon",
|
| 150 |
+
"136": "mri",
|
| 151 |
+
"137": "mrj",
|
| 152 |
+
"138": "msa",
|
| 153 |
+
"139": "mwl",
|
| 154 |
+
"140": "mya",
|
| 155 |
+
"141": "myv",
|
| 156 |
+
"142": "mzn",
|
| 157 |
+
"143": "nap",
|
| 158 |
+
"144": "nav",
|
| 159 |
+
"145": "nci",
|
| 160 |
+
"146": "nds",
|
| 161 |
+
"147": "nds-nl",
|
| 162 |
+
"148": "nep",
|
| 163 |
+
"149": "new",
|
| 164 |
+
"150": "nld",
|
| 165 |
+
"151": "nno",
|
| 166 |
+
"152": "nob",
|
| 167 |
+
"153": "nrm",
|
| 168 |
+
"154": "nso",
|
| 169 |
+
"155": "oci",
|
| 170 |
+
"156": "olo",
|
| 171 |
+
"157": "ori",
|
| 172 |
+
"158": "orm",
|
| 173 |
+
"159": "oss",
|
| 174 |
+
"160": "pag",
|
| 175 |
+
"161": "pam",
|
| 176 |
+
"162": "pan",
|
| 177 |
+
"163": "pap",
|
| 178 |
+
"164": "pcd",
|
| 179 |
+
"165": "pdc",
|
| 180 |
+
"166": "pfl",
|
| 181 |
+
"167": "pnb",
|
| 182 |
+
"168": "pol",
|
| 183 |
+
"169": "por",
|
| 184 |
+
"170": "pus",
|
| 185 |
+
"171": "que",
|
| 186 |
+
"172": "roa-tara",
|
| 187 |
+
"173": "roh",
|
| 188 |
+
"174": "ron",
|
| 189 |
+
"175": "rue",
|
| 190 |
+
"176": "rup",
|
| 191 |
+
"177": "rus",
|
| 192 |
+
"178": "sah",
|
| 193 |
+
"179": "san",
|
| 194 |
+
"180": "scn",
|
| 195 |
+
"181": "sco",
|
| 196 |
+
"182": "sgs",
|
| 197 |
+
"183": "sin",
|
| 198 |
+
"184": "slk",
|
| 199 |
+
"185": "slv",
|
| 200 |
+
"186": "sme",
|
| 201 |
+
"187": "sna",
|
| 202 |
+
"188": "snd",
|
| 203 |
+
"189": "som",
|
| 204 |
+
"190": "spa",
|
| 205 |
+
"191": "sqi",
|
| 206 |
+
"192": "srd",
|
| 207 |
+
"193": "srn",
|
| 208 |
+
"194": "srp",
|
| 209 |
+
"195": "stq",
|
| 210 |
+
"196": "sun",
|
| 211 |
+
"197": "swa",
|
| 212 |
+
"198": "swe",
|
| 213 |
+
"199": "szl",
|
| 214 |
+
"200": "tam",
|
| 215 |
+
"201": "tat",
|
| 216 |
+
"202": "tcy",
|
| 217 |
+
"203": "tel",
|
| 218 |
+
"204": "tet",
|
| 219 |
+
"205": "tgk",
|
| 220 |
+
"206": "tgl",
|
| 221 |
+
"207": "tha",
|
| 222 |
+
"208": "ton",
|
| 223 |
+
"209": "tsn",
|
| 224 |
+
"210": "tuk",
|
| 225 |
+
"211": "tur",
|
| 226 |
+
"212": "tyv",
|
| 227 |
+
"213": "udm",
|
| 228 |
+
"214": "uig",
|
| 229 |
+
"215": "ukr",
|
| 230 |
+
"216": "urd",
|
| 231 |
+
"217": "uzb",
|
| 232 |
+
"218": "vec",
|
| 233 |
+
"219": "vep",
|
| 234 |
+
"220": "vie",
|
| 235 |
+
"221": "vls",
|
| 236 |
+
"222": "vol",
|
| 237 |
+
"223": "vro",
|
| 238 |
+
"224": "war",
|
| 239 |
+
"225": "wln",
|
| 240 |
+
"226": "wol",
|
| 241 |
+
"227": "wuu",
|
| 242 |
+
"228": "xho",
|
| 243 |
+
"229": "xmf",
|
| 244 |
+
"230": "yid",
|
| 245 |
+
"231": "yor",
|
| 246 |
+
"232": "zea",
|
| 247 |
+
"233": "zh-yue",
|
| 248 |
+
"234": "zho"
|
| 249 |
+
},
|
| 250 |
+
"initializer_range": 0.02,
|
| 251 |
+
"intermediate_size": 3072,
|
| 252 |
+
"label2id": {
|
| 253 |
+
"ace": 0,
|
| 254 |
+
"afr": 1,
|
| 255 |
+
"als": 2,
|
| 256 |
+
"amh": 3,
|
| 257 |
+
"ang": 4,
|
| 258 |
+
"ara": 5,
|
| 259 |
+
"arg": 6,
|
| 260 |
+
"arz": 7,
|
| 261 |
+
"asm": 8,
|
| 262 |
+
"ast": 9,
|
| 263 |
+
"ava": 10,
|
| 264 |
+
"aym": 11,
|
| 265 |
+
"azb": 12,
|
| 266 |
+
"aze": 13,
|
| 267 |
+
"bak": 14,
|
| 268 |
+
"bar": 15,
|
| 269 |
+
"bcl": 16,
|
| 270 |
+
"be-tarask": 17,
|
| 271 |
+
"bel": 18,
|
| 272 |
+
"ben": 19,
|
| 273 |
+
"bho": 20,
|
| 274 |
+
"bjn": 21,
|
| 275 |
+
"bod": 22,
|
| 276 |
+
"bos": 23,
|
| 277 |
+
"bpy": 24,
|
| 278 |
+
"bre": 25,
|
| 279 |
+
"bul": 26,
|
| 280 |
+
"bxr": 27,
|
| 281 |
+
"cat": 28,
|
| 282 |
+
"cbk": 29,
|
| 283 |
+
"cdo": 30,
|
| 284 |
+
"ceb": 31,
|
| 285 |
+
"ces": 32,
|
| 286 |
+
"che": 33,
|
| 287 |
+
"chr": 34,
|
| 288 |
+
"chv": 35,
|
| 289 |
+
"ckb": 36,
|
| 290 |
+
"cor": 37,
|
| 291 |
+
"cos": 38,
|
| 292 |
+
"crh": 39,
|
| 293 |
+
"csb": 40,
|
| 294 |
+
"cym": 41,
|
| 295 |
+
"dan": 42,
|
| 296 |
+
"deu": 43,
|
| 297 |
+
"diq": 44,
|
| 298 |
+
"div": 45,
|
| 299 |
+
"dsb": 46,
|
| 300 |
+
"dty": 47,
|
| 301 |
+
"egl": 48,
|
| 302 |
+
"ell": 49,
|
| 303 |
+
"eng": 50,
|
| 304 |
+
"epo": 51,
|
| 305 |
+
"est": 52,
|
| 306 |
+
"eus": 53,
|
| 307 |
+
"ext": 54,
|
| 308 |
+
"fao": 55,
|
| 309 |
+
"fas": 56,
|
| 310 |
+
"fin": 57,
|
| 311 |
+
"fra": 58,
|
| 312 |
+
"frp": 59,
|
| 313 |
+
"fry": 60,
|
| 314 |
+
"fur": 61,
|
| 315 |
+
"gag": 62,
|
| 316 |
+
"gla": 63,
|
| 317 |
+
"gle": 64,
|
| 318 |
+
"glg": 65,
|
| 319 |
+
"glk": 66,
|
| 320 |
+
"glv": 67,
|
| 321 |
+
"grn": 68,
|
| 322 |
+
"guj": 69,
|
| 323 |
+
"hak": 70,
|
| 324 |
+
"hat": 71,
|
| 325 |
+
"hau": 72,
|
| 326 |
+
"hbs": 73,
|
| 327 |
+
"heb": 74,
|
| 328 |
+
"hif": 75,
|
| 329 |
+
"hin": 76,
|
| 330 |
+
"hrv": 77,
|
| 331 |
+
"hsb": 78,
|
| 332 |
+
"hun": 79,
|
| 333 |
+
"hye": 80,
|
| 334 |
+
"ibo": 81,
|
| 335 |
+
"ido": 82,
|
| 336 |
+
"ile": 83,
|
| 337 |
+
"ilo": 84,
|
| 338 |
+
"ina": 85,
|
| 339 |
+
"ind": 86,
|
| 340 |
+
"isl": 87,
|
| 341 |
+
"ita": 88,
|
| 342 |
+
"jam": 89,
|
| 343 |
+
"jav": 90,
|
| 344 |
+
"jbo": 91,
|
| 345 |
+
"jpn": 92,
|
| 346 |
+
"kaa": 93,
|
| 347 |
+
"kab": 94,
|
| 348 |
+
"kan": 95,
|
| 349 |
+
"kat": 96,
|
| 350 |
+
"kaz": 97,
|
| 351 |
+
"kbd": 98,
|
| 352 |
+
"khm": 99,
|
| 353 |
+
"kin": 100,
|
| 354 |
+
"kir": 101,
|
| 355 |
+
"koi": 102,
|
| 356 |
+
"kok": 103,
|
| 357 |
+
"kom": 104,
|
| 358 |
+
"kor": 105,
|
| 359 |
+
"krc": 106,
|
| 360 |
+
"ksh": 107,
|
| 361 |
+
"kur": 108,
|
| 362 |
+
"lad": 109,
|
| 363 |
+
"lao": 110,
|
| 364 |
+
"lat": 111,
|
| 365 |
+
"lav": 112,
|
| 366 |
+
"lez": 113,
|
| 367 |
+
"lij": 114,
|
| 368 |
+
"lim": 115,
|
| 369 |
+
"lin": 116,
|
| 370 |
+
"lit": 117,
|
| 371 |
+
"lmo": 118,
|
| 372 |
+
"lrc": 119,
|
| 373 |
+
"ltg": 120,
|
| 374 |
+
"ltz": 121,
|
| 375 |
+
"lug": 122,
|
| 376 |
+
"lzh": 123,
|
| 377 |
+
"mai": 124,
|
| 378 |
+
"mal": 125,
|
| 379 |
+
"map-bms": 126,
|
| 380 |
+
"mar": 127,
|
| 381 |
+
"mdf": 128,
|
| 382 |
+
"mhr": 129,
|
| 383 |
+
"min": 130,
|
| 384 |
+
"mkd": 131,
|
| 385 |
+
"mlg": 132,
|
| 386 |
+
"mlt": 133,
|
| 387 |
+
"nan": 134,
|
| 388 |
+
"mon": 135,
|
| 389 |
+
"mri": 136,
|
| 390 |
+
"mrj": 137,
|
| 391 |
+
"msa": 138,
|
| 392 |
+
"mwl": 139,
|
| 393 |
+
"mya": 140,
|
| 394 |
+
"myv": 141,
|
| 395 |
+
"mzn": 142,
|
| 396 |
+
"nap": 143,
|
| 397 |
+
"nav": 144,
|
| 398 |
+
"nci": 145,
|
| 399 |
+
"nds": 146,
|
| 400 |
+
"nds-nl": 147,
|
| 401 |
+
"nep": 148,
|
| 402 |
+
"new": 149,
|
| 403 |
+
"nld": 150,
|
| 404 |
+
"nno": 151,
|
| 405 |
+
"nob": 152,
|
| 406 |
+
"nrm": 153,
|
| 407 |
+
"nso": 154,
|
| 408 |
+
"oci": 155,
|
| 409 |
+
"olo": 156,
|
| 410 |
+
"ori": 157,
|
| 411 |
+
"orm": 158,
|
| 412 |
+
"oss": 159,
|
| 413 |
+
"pag": 160,
|
| 414 |
+
"pam": 161,
|
| 415 |
+
"pan": 162,
|
| 416 |
+
"pap": 163,
|
| 417 |
+
"pcd": 164,
|
| 418 |
+
"pdc": 165,
|
| 419 |
+
"pfl": 166,
|
| 420 |
+
"pnb": 167,
|
| 421 |
+
"pol": 168,
|
| 422 |
+
"por": 169,
|
| 423 |
+
"pus": 170,
|
| 424 |
+
"que": 171,
|
| 425 |
+
"roa-tara": 172,
|
| 426 |
+
"roh": 173,
|
| 427 |
+
"ron": 174,
|
| 428 |
+
"rue": 175,
|
| 429 |
+
"rup": 176,
|
| 430 |
+
"rus": 177,
|
| 431 |
+
"sah": 178,
|
| 432 |
+
"san": 179,
|
| 433 |
+
"scn": 180,
|
| 434 |
+
"sco": 181,
|
| 435 |
+
"sgs": 182,
|
| 436 |
+
"sin": 183,
|
| 437 |
+
"slk": 184,
|
| 438 |
+
"slv": 185,
|
| 439 |
+
"sme": 186,
|
| 440 |
+
"sna": 187,
|
| 441 |
+
"snd": 188,
|
| 442 |
+
"som": 189,
|
| 443 |
+
"spa": 190,
|
| 444 |
+
"sqi": 191,
|
| 445 |
+
"srd": 192,
|
| 446 |
+
"srn": 193,
|
| 447 |
+
"srp": 194,
|
| 448 |
+
"stq": 195,
|
| 449 |
+
"sun": 196,
|
| 450 |
+
"swa": 197,
|
| 451 |
+
"swe": 198,
|
| 452 |
+
"szl": 199,
|
| 453 |
+
"tam": 200,
|
| 454 |
+
"tat": 201,
|
| 455 |
+
"tcy": 202,
|
| 456 |
+
"tel": 203,
|
| 457 |
+
"tet": 204,
|
| 458 |
+
"tgk": 205,
|
| 459 |
+
"tgl": 206,
|
| 460 |
+
"tha": 207,
|
| 461 |
+
"ton": 208,
|
| 462 |
+
"tsn": 209,
|
| 463 |
+
"tuk": 210,
|
| 464 |
+
"tur": 211,
|
| 465 |
+
"tyv": 212,
|
| 466 |
+
"udm": 213,
|
| 467 |
+
"uig": 214,
|
| 468 |
+
"ukr": 215,
|
| 469 |
+
"urd": 216,
|
| 470 |
+
"uzb": 217,
|
| 471 |
+
"vec": 218,
|
| 472 |
+
"vep": 219,
|
| 473 |
+
"vie": 220,
|
| 474 |
+
"vls": 221,
|
| 475 |
+
"vol": 222,
|
| 476 |
+
"vro": 223,
|
| 477 |
+
"war": 224,
|
| 478 |
+
"wln": 225,
|
| 479 |
+
"wol": 226,
|
| 480 |
+
"wuu": 227,
|
| 481 |
+
"xho": 228,
|
| 482 |
+
"xmf": 229,
|
| 483 |
+
"yid": 230,
|
| 484 |
+
"yor": 231,
|
| 485 |
+
"zea": 232,
|
| 486 |
+
"zh-yue": 233,
|
| 487 |
+
"zho": 234
|
| 488 |
+
},
|
| 489 |
+
"layer_norm_eps": 1e-12,
|
| 490 |
+
"max_position_embeddings": 512,
|
| 491 |
+
"model_type": "roberta",
|
| 492 |
+
"num_attention_heads": 12,
|
| 493 |
+
"num_hidden_layers": 12,
|
| 494 |
+
"pad_token_id": 1,
|
| 495 |
+
"position_embedding_type": "absolute",
|
| 496 |
+
"problem_type": "single_label_classification",
|
| 497 |
+
"transformers_version": "4.8.0.dev0",
|
| 498 |
+
"type_vocab_size": 2,
|
| 499 |
+
"use_cache": true,
|
| 500 |
+
"vocab_size": 32000
|
| 501 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8863bf86a75da23bd15abd55fd68a0d6e2a8a70851389f96e13580dc356301cd
|
| 3 |
+
size 443277549
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": {"content": "<mask>", "single_word": false, "lstrip": true, "rstrip": false, "normalized": false}}
|
tf_model.h5
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:776984d5f858636ec3bdd8021e4a897de447f6c0e8d25b1a3eb0239829f9dabe
|
| 3 |
+
size 443480960
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "add_prefix_space": false, "errors": "replace", "sep_token": "</s>", "cls_token": "<s>", "pad_token": "<pad>", "mask_token": "<mask>", "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "/content/checkpoint-84500", "tokenizer_class": "RobertaTokenizer"}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,310 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 95214,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.12,
|
| 12 |
+
"learning_rate": 3e-05,
|
| 13 |
+
"loss": 2.2074,
|
| 14 |
+
"step": 5000
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.12,
|
| 18 |
+
"eval_accuracy": 0.6701050400733948,
|
| 19 |
+
"eval_loss": 1.3863972425460815,
|
| 20 |
+
"eval_runtime": 1030.5845,
|
| 21 |
+
"eval_samples_per_second": 109.496,
|
| 22 |
+
"eval_steps_per_second": 9.125,
|
| 23 |
+
"step": 5000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.24,
|
| 27 |
+
"learning_rate": 2.877048978688162e-05,
|
| 28 |
+
"loss": 1.2979,
|
| 29 |
+
"step": 10000
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 0.24,
|
| 33 |
+
"eval_accuracy": 0.7089281678199768,
|
| 34 |
+
"eval_loss": 1.1698851585388184,
|
| 35 |
+
"eval_runtime": 1023.3048,
|
| 36 |
+
"eval_samples_per_second": 110.275,
|
| 37 |
+
"eval_steps_per_second": 9.19,
|
| 38 |
+
"step": 10000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.35,
|
| 42 |
+
"learning_rate": 2.7540979573763234e-05,
|
| 43 |
+
"loss": 1.1725,
|
| 44 |
+
"step": 15000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.35,
|
| 48 |
+
"eval_accuracy": 0.723913311958313,
|
| 49 |
+
"eval_loss": 1.0914552211761475,
|
| 50 |
+
"eval_runtime": 1021.0004,
|
| 51 |
+
"eval_samples_per_second": 110.524,
|
| 52 |
+
"eval_steps_per_second": 9.211,
|
| 53 |
+
"step": 15000
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.47,
|
| 57 |
+
"learning_rate": 2.6311223360202048e-05,
|
| 58 |
+
"loss": 1.1066,
|
| 59 |
+
"step": 20000
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.47,
|
| 63 |
+
"eval_accuracy": 0.7364969849586487,
|
| 64 |
+
"eval_loss": 1.0439897775650024,
|
| 65 |
+
"eval_runtime": 1023.0505,
|
| 66 |
+
"eval_samples_per_second": 110.302,
|
| 67 |
+
"eval_steps_per_second": 9.192,
|
| 68 |
+
"step": 20000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.79,
|
| 72 |
+
"learning_rate": 2.3351475380761303e-05,
|
| 73 |
+
"loss": 1.0399,
|
| 74 |
+
"step": 25000
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.79,
|
| 78 |
+
"eval_accuracy": 0.7490540146827698,
|
| 79 |
+
"eval_loss": 0.9834885001182556,
|
| 80 |
+
"eval_runtime": 1112.3584,
|
| 81 |
+
"eval_samples_per_second": 101.447,
|
| 82 |
+
"eval_steps_per_second": 6.341,
|
| 83 |
+
"step": 25000
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 0.95,
|
| 87 |
+
"learning_rate": 2.1689427361606846e-05,
|
| 88 |
+
"loss": 1.0012,
|
| 89 |
+
"step": 30000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.95,
|
| 93 |
+
"eval_accuracy": 0.7581638693809509,
|
| 94 |
+
"eval_loss": 0.9408503770828247,
|
| 95 |
+
"eval_runtime": 1113.6514,
|
| 96 |
+
"eval_samples_per_second": 101.329,
|
| 97 |
+
"eval_steps_per_second": 6.333,
|
| 98 |
+
"step": 30000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.1,
|
| 102 |
+
"learning_rate": 2.0027046799831514e-05,
|
| 103 |
+
"loss": 0.8971,
|
| 104 |
+
"step": 35000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 1.1,
|
| 108 |
+
"eval_accuracy": 0.7663254737854004,
|
| 109 |
+
"eval_loss": 0.9190825819969177,
|
| 110 |
+
"eval_runtime": 1114.1162,
|
| 111 |
+
"eval_samples_per_second": 101.287,
|
| 112 |
+
"eval_steps_per_second": 6.331,
|
| 113 |
+
"step": 35000
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 1.26,
|
| 117 |
+
"learning_rate": 1.8364998780677056e-05,
|
| 118 |
+
"loss": 0.8405,
|
| 119 |
+
"step": 40000
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.26,
|
| 123 |
+
"eval_accuracy": 0.7693473100662231,
|
| 124 |
+
"eval_loss": 0.9057066440582275,
|
| 125 |
+
"eval_runtime": 1056.7123,
|
| 126 |
+
"eval_samples_per_second": 106.789,
|
| 127 |
+
"eval_steps_per_second": 6.674,
|
| 128 |
+
"step": 40000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.42,
|
| 132 |
+
"learning_rate": 1.6702950761522602e-05,
|
| 133 |
+
"loss": 0.834,
|
| 134 |
+
"step": 45000
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 1.42,
|
| 138 |
+
"eval_accuracy": 0.775940477848053,
|
| 139 |
+
"eval_loss": 0.8832579851150513,
|
| 140 |
+
"eval_runtime": 1059.1014,
|
| 141 |
+
"eval_samples_per_second": 106.548,
|
| 142 |
+
"eval_steps_per_second": 6.659,
|
| 143 |
+
"step": 45000
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.58,
|
| 147 |
+
"learning_rate": 1.5040570199747268e-05,
|
| 148 |
+
"loss": 0.8212,
|
| 149 |
+
"step": 50000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.58,
|
| 153 |
+
"eval_accuracy": 0.7792636156082153,
|
| 154 |
+
"eval_loss": 0.866415798664093,
|
| 155 |
+
"eval_runtime": 1054.2119,
|
| 156 |
+
"eval_samples_per_second": 107.042,
|
| 157 |
+
"eval_steps_per_second": 6.69,
|
| 158 |
+
"step": 50000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 1.73,
|
| 162 |
+
"learning_rate": 1.3378854723213692e-05,
|
| 163 |
+
"loss": 0.8133,
|
| 164 |
+
"step": 55000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 1.73,
|
| 168 |
+
"eval_accuracy": 0.7830652594566345,
|
| 169 |
+
"eval_loss": 0.8482676148414612,
|
| 170 |
+
"eval_runtime": 1054.5931,
|
| 171 |
+
"eval_samples_per_second": 107.003,
|
| 172 |
+
"eval_steps_per_second": 6.688,
|
| 173 |
+
"step": 55000
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 1.89,
|
| 177 |
+
"learning_rate": 1.1716474161438358e-05,
|
| 178 |
+
"loss": 0.7947,
|
| 179 |
+
"step": 60000
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 1.89,
|
| 183 |
+
"eval_accuracy": 0.7862643599510193,
|
| 184 |
+
"eval_loss": 0.8344442844390869,
|
| 185 |
+
"eval_runtime": 1058.4611,
|
| 186 |
+
"eval_samples_per_second": 106.612,
|
| 187 |
+
"eval_steps_per_second": 6.663,
|
| 188 |
+
"step": 60000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 2.05,
|
| 192 |
+
"learning_rate": 1.0054758684904782e-05,
|
| 193 |
+
"loss": 0.7446,
|
| 194 |
+
"step": 65000
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 2.05,
|
| 198 |
+
"eval_accuracy": 0.7890114784240723,
|
| 199 |
+
"eval_loss": 0.8323635458946228,
|
| 200 |
+
"eval_runtime": 1054.4779,
|
| 201 |
+
"eval_samples_per_second": 107.015,
|
| 202 |
+
"eval_steps_per_second": 6.689,
|
| 203 |
+
"step": 65000
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 2.21,
|
| 207 |
+
"learning_rate": 8.392378123129448e-06,
|
| 208 |
+
"loss": 0.6463,
|
| 209 |
+
"step": 70000
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 2.21,
|
| 213 |
+
"eval_accuracy": 0.7913243770599365,
|
| 214 |
+
"eval_loss": 0.8241144418716431,
|
| 215 |
+
"eval_runtime": 1057.5768,
|
| 216 |
+
"eval_samples_per_second": 106.701,
|
| 217 |
+
"eval_steps_per_second": 6.669,
|
| 218 |
+
"step": 70000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 2.36,
|
| 222 |
+
"learning_rate": 6.7306626465958724e-06,
|
| 223 |
+
"loss": 0.6461,
|
| 224 |
+
"step": 75000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 2.36,
|
| 228 |
+
"eval_accuracy": 0.792999267578125,
|
| 229 |
+
"eval_loss": 0.8187506794929504,
|
| 230 |
+
"eval_runtime": 1055.2791,
|
| 231 |
+
"eval_samples_per_second": 106.934,
|
| 232 |
+
"eval_steps_per_second": 6.684,
|
| 233 |
+
"step": 75000
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 2.52,
|
| 237 |
+
"learning_rate": 5.068282084820538e-06,
|
| 238 |
+
"loss": 0.6403,
|
| 239 |
+
"step": 80000
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 2.52,
|
| 243 |
+
"eval_accuracy": 0.7952146530151367,
|
| 244 |
+
"eval_loss": 0.8096941709518433,
|
| 245 |
+
"eval_runtime": 1058.933,
|
| 246 |
+
"eval_samples_per_second": 106.565,
|
| 247 |
+
"eval_steps_per_second": 6.66,
|
| 248 |
+
"step": 80000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 2.68,
|
| 252 |
+
"learning_rate": 3.406566608286962e-06,
|
| 253 |
+
"loss": 0.6294,
|
| 254 |
+
"step": 85000
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 2.68,
|
| 258 |
+
"eval_accuracy": 0.7971464991569519,
|
| 259 |
+
"eval_loss": 0.8044777512550354,
|
| 260 |
+
"eval_runtime": 1078.5713,
|
| 261 |
+
"eval_samples_per_second": 104.625,
|
| 262 |
+
"eval_steps_per_second": 6.539,
|
| 263 |
+
"step": 85000
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 2.84,
|
| 267 |
+
"learning_rate": 1.744186046511628e-06,
|
| 268 |
+
"loss": 0.6198,
|
| 269 |
+
"step": 90000
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 2.84,
|
| 273 |
+
"eval_accuracy": 0.7985377907752991,
|
| 274 |
+
"eval_loss": 0.7970817685127258,
|
| 275 |
+
"eval_runtime": 1091.9096,
|
| 276 |
+
"eval_samples_per_second": 103.346,
|
| 277 |
+
"eval_steps_per_second": 6.459,
|
| 278 |
+
"step": 90000
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 2.99,
|
| 282 |
+
"learning_rate": 8.213802735717294e-08,
|
| 283 |
+
"loss": 0.6169,
|
| 284 |
+
"step": 95000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 2.99,
|
| 288 |
+
"eval_accuracy": 0.7989366054534912,
|
| 289 |
+
"eval_loss": 0.794947624206543,
|
| 290 |
+
"eval_runtime": 1091.2301,
|
| 291 |
+
"eval_samples_per_second": 103.411,
|
| 292 |
+
"eval_steps_per_second": 6.463,
|
| 293 |
+
"step": 95000
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 3.0,
|
| 297 |
+
"step": 95214,
|
| 298 |
+
"total_flos": 2.918866335440248e+17,
|
| 299 |
+
"train_loss": 0.06961118817632395,
|
| 300 |
+
"train_runtime": 8353.1581,
|
| 301 |
+
"train_samples_per_second": 364.749,
|
| 302 |
+
"train_steps_per_second": 11.399
|
| 303 |
+
}
|
| 304 |
+
],
|
| 305 |
+
"max_steps": 95214,
|
| 306 |
+
"num_train_epochs": 3,
|
| 307 |
+
"total_flos": 2.918866335440248e+17,
|
| 308 |
+
"trial_name": null,
|
| 309 |
+
"trial_params": null
|
| 310 |
+
}
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|