Training in progress, epoch 1
Browse files- config.json +392 -0
- model.safetensors +3 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,392 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"ModernBertForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_bias": false,
|
| 6 |
+
"attention_dropout": 0.0,
|
| 7 |
+
"bos_token_id": 50281,
|
| 8 |
+
"classifier_activation": "gelu",
|
| 9 |
+
"classifier_bias": false,
|
| 10 |
+
"classifier_dropout": 0.0,
|
| 11 |
+
"classifier_pooling": "mean",
|
| 12 |
+
"cls_token_id": 50281,
|
| 13 |
+
"decoder_bias": true,
|
| 14 |
+
"deterministic_flash_attn": false,
|
| 15 |
+
"dtype": "float32",
|
| 16 |
+
"embedding_dropout": 0.0,
|
| 17 |
+
"eos_token_id": 50282,
|
| 18 |
+
"global_attn_every_n_layers": 3,
|
| 19 |
+
"gradient_checkpointing": false,
|
| 20 |
+
"hidden_activation": "gelu",
|
| 21 |
+
"hidden_size": 768,
|
| 22 |
+
"id2label": {
|
| 23 |
+
"0": "Nim",
|
| 24 |
+
"1": "Vue",
|
| 25 |
+
"10": "Stylus",
|
| 26 |
+
"100": "Bluespec",
|
| 27 |
+
"101": "Apex",
|
| 28 |
+
"102": "Hoon",
|
| 29 |
+
"103": "TSV",
|
| 30 |
+
"104": "VCL",
|
| 31 |
+
"105": "Gradle",
|
| 32 |
+
"106": "Ioke",
|
| 33 |
+
"107": "Lex",
|
| 34 |
+
"108": "Visual Basic .NET",
|
| 35 |
+
"109": "Gnuplot",
|
| 36 |
+
"11": "ColdFusion",
|
| 37 |
+
"110": "Common Lisp",
|
| 38 |
+
"111": "JQ",
|
| 39 |
+
"112": "Dockerfile",
|
| 40 |
+
"113": "Fish",
|
| 41 |
+
"114": "SQL",
|
| 42 |
+
"115": "Lean",
|
| 43 |
+
"116": "Kvlang",
|
| 44 |
+
"117": "Ada",
|
| 45 |
+
"118": "YANG",
|
| 46 |
+
"119": "Java",
|
| 47 |
+
"12": "AGS Script",
|
| 48 |
+
"120": "Haskell",
|
| 49 |
+
"121": "Less",
|
| 50 |
+
"122": "Vim Script",
|
| 51 |
+
"123": "PureScript",
|
| 52 |
+
"124": "Scilab",
|
| 53 |
+
"125": "RDoc",
|
| 54 |
+
"126": "HTML+Razor",
|
| 55 |
+
"127": "Ragel in Ruby Host",
|
| 56 |
+
"128": "Go",
|
| 57 |
+
"129": "EJS",
|
| 58 |
+
"13": "Raw token data",
|
| 59 |
+
"130": "PostScript",
|
| 60 |
+
"131": "GLSL",
|
| 61 |
+
"132": "Makefile",
|
| 62 |
+
"133": "FreeMarker",
|
| 63 |
+
"134": "Io",
|
| 64 |
+
"135": "GAS",
|
| 65 |
+
"136": "PLpgSQL",
|
| 66 |
+
"137": "Unity3D Asset",
|
| 67 |
+
"138": "CMake",
|
| 68 |
+
"139": "CoffeeScript",
|
| 69 |
+
"14": "Csound",
|
| 70 |
+
"140": "J",
|
| 71 |
+
"141": "HCL",
|
| 72 |
+
"142": "Edoid",
|
| 73 |
+
"143": "TSX",
|
| 74 |
+
"144": "GDScript",
|
| 75 |
+
"145": "Thrift",
|
| 76 |
+
"146": "DIGITAL Command Language",
|
| 77 |
+
"147": "Kotlin",
|
| 78 |
+
"148": "Visual Basic",
|
| 79 |
+
"149": "Objective-C",
|
| 80 |
+
"15": "Sass",
|
| 81 |
+
"150": "Smali",
|
| 82 |
+
"151": "Fluent",
|
| 83 |
+
"152": "Git Config",
|
| 84 |
+
"153": "Assembly",
|
| 85 |
+
"16": "C",
|
| 86 |
+
"17": "Jupyter Notebook",
|
| 87 |
+
"18": "CODEOWNERS",
|
| 88 |
+
"19": "Rascal",
|
| 89 |
+
"2": "Squirrel",
|
| 90 |
+
"20": "Blade",
|
| 91 |
+
"21": "Eagle",
|
| 92 |
+
"22": "Objective-C++",
|
| 93 |
+
"23": "PHP",
|
| 94 |
+
"24": "Crystal",
|
| 95 |
+
"25": "OCaml",
|
| 96 |
+
"26": "Scheme",
|
| 97 |
+
"27": "Modelica",
|
| 98 |
+
"28": "Rust",
|
| 99 |
+
"29": "Unix Assembly",
|
| 100 |
+
"3": "Gherkin",
|
| 101 |
+
"30": "Shell",
|
| 102 |
+
"31": "JAR Manifest",
|
| 103 |
+
"32": "XML Property List",
|
| 104 |
+
"33": "Erlang",
|
| 105 |
+
"34": "Prolog",
|
| 106 |
+
"35": "ECL",
|
| 107 |
+
"36": "Csound Document",
|
| 108 |
+
"37": "Perl",
|
| 109 |
+
"38": "LookML",
|
| 110 |
+
"39": "Swift",
|
| 111 |
+
"4": "Logtalk",
|
| 112 |
+
"40": "C#",
|
| 113 |
+
"41": "Go Module",
|
| 114 |
+
"42": "Scala",
|
| 115 |
+
"43": "Gettext Catalog",
|
| 116 |
+
"44": "FreeBasic",
|
| 117 |
+
"45": "Adobe Font Metrics",
|
| 118 |
+
"46": "PowerShell",
|
| 119 |
+
"47": "Pascal",
|
| 120 |
+
"48": "Wavefront Object",
|
| 121 |
+
"49": "AsciiDoc",
|
| 122 |
+
"5": "Redcode",
|
| 123 |
+
"50": "Python",
|
| 124 |
+
"51": "Julia",
|
| 125 |
+
"52": "POV-Ray SDL",
|
| 126 |
+
"53": "Starlark",
|
| 127 |
+
"54": "Metal",
|
| 128 |
+
"55": "GAP",
|
| 129 |
+
"56": "Kit",
|
| 130 |
+
"57": "Graphviz (DOT)",
|
| 131 |
+
"58": "TypeScript",
|
| 132 |
+
"59": "Gerber Image",
|
| 133 |
+
"6": "ImageJ Macro",
|
| 134 |
+
"60": "Mathematica",
|
| 135 |
+
"61": "REALbasic",
|
| 136 |
+
"62": "SQF",
|
| 137 |
+
"63": "Solidity",
|
| 138 |
+
"64": "OpenType Feature File",
|
| 139 |
+
"65": "Unknown",
|
| 140 |
+
"66": "Protocol Buffer Text Format",
|
| 141 |
+
"67": "VHDL",
|
| 142 |
+
"68": "Haxe",
|
| 143 |
+
"69": "Twig",
|
| 144 |
+
"7": "Hack",
|
| 145 |
+
"70": "Isabelle",
|
| 146 |
+
"71": "Lua",
|
| 147 |
+
"72": "Groovy",
|
| 148 |
+
"73": "Ignore List",
|
| 149 |
+
"74": "LLVM",
|
| 150 |
+
"75": "JavaScript",
|
| 151 |
+
"76": "MATLAB",
|
| 152 |
+
"77": "Dart",
|
| 153 |
+
"78": "Inform 7",
|
| 154 |
+
"79": "VBScript",
|
| 155 |
+
"8": "PlantUML",
|
| 156 |
+
"80": "KiCad Layout",
|
| 157 |
+
"81": "OpenEdge ABL",
|
| 158 |
+
"82": "Turtle",
|
| 159 |
+
"83": "Open Policy Agent",
|
| 160 |
+
"84": "D",
|
| 161 |
+
"85": "Inno Setup",
|
| 162 |
+
"86": "ApacheConf",
|
| 163 |
+
"87": "C++",
|
| 164 |
+
"88": "SCSS",
|
| 165 |
+
"89": "Ruby",
|
| 166 |
+
"9": "Verilog",
|
| 167 |
+
"90": "Chapel",
|
| 168 |
+
"91": "OpenStep Property List",
|
| 169 |
+
"92": "Fortran Free Form",
|
| 170 |
+
"93": "ObjDump",
|
| 171 |
+
"94": "G-code",
|
| 172 |
+
"95": "PicoLisp",
|
| 173 |
+
"96": "XS",
|
| 174 |
+
"97": "Vim Snippet",
|
| 175 |
+
"98": "R",
|
| 176 |
+
"99": "GDB"
|
| 177 |
+
},
|
| 178 |
+
"initializer_cutoff_factor": 2.0,
|
| 179 |
+
"initializer_range": 0.02,
|
| 180 |
+
"intermediate_size": 1152,
|
| 181 |
+
"label2id": {
|
| 182 |
+
"AGS Script": "12",
|
| 183 |
+
"Ada": "117",
|
| 184 |
+
"Adobe Font Metrics": "45",
|
| 185 |
+
"ApacheConf": "86",
|
| 186 |
+
"Apex": "101",
|
| 187 |
+
"AsciiDoc": "49",
|
| 188 |
+
"Assembly": "153",
|
| 189 |
+
"Blade": "20",
|
| 190 |
+
"Bluespec": "100",
|
| 191 |
+
"C": "16",
|
| 192 |
+
"C#": "40",
|
| 193 |
+
"C++": "87",
|
| 194 |
+
"CMake": "138",
|
| 195 |
+
"CODEOWNERS": "18",
|
| 196 |
+
"Chapel": "90",
|
| 197 |
+
"CoffeeScript": "139",
|
| 198 |
+
"ColdFusion": "11",
|
| 199 |
+
"Common Lisp": "110",
|
| 200 |
+
"Crystal": "24",
|
| 201 |
+
"Csound": "14",
|
| 202 |
+
"Csound Document": "36",
|
| 203 |
+
"D": "84",
|
| 204 |
+
"DIGITAL Command Language": "146",
|
| 205 |
+
"Dart": "77",
|
| 206 |
+
"Dockerfile": "112",
|
| 207 |
+
"ECL": "35",
|
| 208 |
+
"EJS": "129",
|
| 209 |
+
"Eagle": "21",
|
| 210 |
+
"Edoid": "142",
|
| 211 |
+
"Erlang": "33",
|
| 212 |
+
"Fish": "113",
|
| 213 |
+
"Fluent": "151",
|
| 214 |
+
"Fortran Free Form": "92",
|
| 215 |
+
"FreeBasic": "44",
|
| 216 |
+
"FreeMarker": "133",
|
| 217 |
+
"G-code": "94",
|
| 218 |
+
"GAP": "55",
|
| 219 |
+
"GAS": "135",
|
| 220 |
+
"GDB": "99",
|
| 221 |
+
"GDScript": "144",
|
| 222 |
+
"GLSL": "131",
|
| 223 |
+
"Gerber Image": "59",
|
| 224 |
+
"Gettext Catalog": "43",
|
| 225 |
+
"Gherkin": "3",
|
| 226 |
+
"Git Config": "152",
|
| 227 |
+
"Gnuplot": "109",
|
| 228 |
+
"Go": "128",
|
| 229 |
+
"Go Module": "41",
|
| 230 |
+
"Gradle": "105",
|
| 231 |
+
"Graphviz (DOT)": "57",
|
| 232 |
+
"Groovy": "72",
|
| 233 |
+
"HCL": "141",
|
| 234 |
+
"HTML+Razor": "126",
|
| 235 |
+
"Hack": "7",
|
| 236 |
+
"Haskell": "120",
|
| 237 |
+
"Haxe": "68",
|
| 238 |
+
"Hoon": "102",
|
| 239 |
+
"Ignore List": "73",
|
| 240 |
+
"ImageJ Macro": "6",
|
| 241 |
+
"Inform 7": "78",
|
| 242 |
+
"Inno Setup": "85",
|
| 243 |
+
"Io": "134",
|
| 244 |
+
"Ioke": "106",
|
| 245 |
+
"Isabelle": "70",
|
| 246 |
+
"J": "140",
|
| 247 |
+
"JAR Manifest": "31",
|
| 248 |
+
"JQ": "111",
|
| 249 |
+
"Java": "119",
|
| 250 |
+
"JavaScript": "75",
|
| 251 |
+
"Julia": "51",
|
| 252 |
+
"Jupyter Notebook": "17",
|
| 253 |
+
"KiCad Layout": "80",
|
| 254 |
+
"Kit": "56",
|
| 255 |
+
"Kotlin": "147",
|
| 256 |
+
"Kvlang": "116",
|
| 257 |
+
"LLVM": "74",
|
| 258 |
+
"Lean": "115",
|
| 259 |
+
"Less": "121",
|
| 260 |
+
"Lex": "107",
|
| 261 |
+
"Logtalk": "4",
|
| 262 |
+
"LookML": "38",
|
| 263 |
+
"Lua": "71",
|
| 264 |
+
"MATLAB": "76",
|
| 265 |
+
"Makefile": "132",
|
| 266 |
+
"Mathematica": "60",
|
| 267 |
+
"Metal": "54",
|
| 268 |
+
"Modelica": "27",
|
| 269 |
+
"Nim": "0",
|
| 270 |
+
"OCaml": "25",
|
| 271 |
+
"ObjDump": "93",
|
| 272 |
+
"Objective-C": "149",
|
| 273 |
+
"Objective-C++": "22",
|
| 274 |
+
"Open Policy Agent": "83",
|
| 275 |
+
"OpenEdge ABL": "81",
|
| 276 |
+
"OpenStep Property List": "91",
|
| 277 |
+
"OpenType Feature File": "64",
|
| 278 |
+
"PHP": "23",
|
| 279 |
+
"PLpgSQL": "136",
|
| 280 |
+
"POV-Ray SDL": "52",
|
| 281 |
+
"Pascal": "47",
|
| 282 |
+
"Perl": "37",
|
| 283 |
+
"PicoLisp": "95",
|
| 284 |
+
"PlantUML": "8",
|
| 285 |
+
"PostScript": "130",
|
| 286 |
+
"PowerShell": "46",
|
| 287 |
+
"Prolog": "34",
|
| 288 |
+
"Protocol Buffer Text Format": "66",
|
| 289 |
+
"PureScript": "123",
|
| 290 |
+
"Python": "50",
|
| 291 |
+
"R": "98",
|
| 292 |
+
"RDoc": "125",
|
| 293 |
+
"REALbasic": "61",
|
| 294 |
+
"Ragel in Ruby Host": "127",
|
| 295 |
+
"Rascal": "19",
|
| 296 |
+
"Raw token data": "13",
|
| 297 |
+
"Redcode": "5",
|
| 298 |
+
"Ruby": "89",
|
| 299 |
+
"Rust": "28",
|
| 300 |
+
"SCSS": "88",
|
| 301 |
+
"SQF": "62",
|
| 302 |
+
"SQL": "114",
|
| 303 |
+
"Sass": "15",
|
| 304 |
+
"Scala": "42",
|
| 305 |
+
"Scheme": "26",
|
| 306 |
+
"Scilab": "124",
|
| 307 |
+
"Shell": "30",
|
| 308 |
+
"Smali": "150",
|
| 309 |
+
"Solidity": "63",
|
| 310 |
+
"Squirrel": "2",
|
| 311 |
+
"Starlark": "53",
|
| 312 |
+
"Stylus": "10",
|
| 313 |
+
"Swift": "39",
|
| 314 |
+
"TSV": "103",
|
| 315 |
+
"TSX": "143",
|
| 316 |
+
"Thrift": "145",
|
| 317 |
+
"Turtle": "82",
|
| 318 |
+
"Twig": "69",
|
| 319 |
+
"TypeScript": "58",
|
| 320 |
+
"Unity3D Asset": "137",
|
| 321 |
+
"Unix Assembly": "29",
|
| 322 |
+
"Unknown": "65",
|
| 323 |
+
"VBScript": "79",
|
| 324 |
+
"VCL": "104",
|
| 325 |
+
"VHDL": "67",
|
| 326 |
+
"Verilog": "9",
|
| 327 |
+
"Vim Script": "122",
|
| 328 |
+
"Vim Snippet": "97",
|
| 329 |
+
"Visual Basic": "148",
|
| 330 |
+
"Visual Basic .NET": "108",
|
| 331 |
+
"Vue": "1",
|
| 332 |
+
"Wavefront Object": "48",
|
| 333 |
+
"XML Property List": "32",
|
| 334 |
+
"XS": "96",
|
| 335 |
+
"YANG": "118"
|
| 336 |
+
},
|
| 337 |
+
"layer_norm_eps": 1e-05,
|
| 338 |
+
"layer_types": [
|
| 339 |
+
"full_attention",
|
| 340 |
+
"sliding_attention",
|
| 341 |
+
"sliding_attention",
|
| 342 |
+
"full_attention",
|
| 343 |
+
"sliding_attention",
|
| 344 |
+
"sliding_attention",
|
| 345 |
+
"full_attention",
|
| 346 |
+
"sliding_attention",
|
| 347 |
+
"sliding_attention",
|
| 348 |
+
"full_attention",
|
| 349 |
+
"sliding_attention",
|
| 350 |
+
"sliding_attention",
|
| 351 |
+
"full_attention",
|
| 352 |
+
"sliding_attention",
|
| 353 |
+
"sliding_attention",
|
| 354 |
+
"full_attention",
|
| 355 |
+
"sliding_attention",
|
| 356 |
+
"sliding_attention",
|
| 357 |
+
"full_attention",
|
| 358 |
+
"sliding_attention",
|
| 359 |
+
"sliding_attention",
|
| 360 |
+
"full_attention"
|
| 361 |
+
],
|
| 362 |
+
"local_attention": 128,
|
| 363 |
+
"max_position_embeddings": 8192,
|
| 364 |
+
"mlp_bias": false,
|
| 365 |
+
"mlp_dropout": 0.0,
|
| 366 |
+
"model_type": "modernbert",
|
| 367 |
+
"norm_bias": false,
|
| 368 |
+
"norm_eps": 1e-05,
|
| 369 |
+
"num_attention_heads": 12,
|
| 370 |
+
"num_hidden_layers": 22,
|
| 371 |
+
"pad_token_id": 50283,
|
| 372 |
+
"position_embedding_type": "absolute",
|
| 373 |
+
"problem_type": "single_label_classification",
|
| 374 |
+
"repad_logits_with_grad": false,
|
| 375 |
+
"rope_parameters": {
|
| 376 |
+
"full_attention": {
|
| 377 |
+
"rope_theta": 160000.0,
|
| 378 |
+
"rope_type": "default"
|
| 379 |
+
},
|
| 380 |
+
"sliding_attention": {
|
| 381 |
+
"rope_theta": 10000.0,
|
| 382 |
+
"rope_type": "default"
|
| 383 |
+
}
|
| 384 |
+
},
|
| 385 |
+
"sep_token_id": 50282,
|
| 386 |
+
"sparse_pred_ignore_index": -100,
|
| 387 |
+
"sparse_prediction": false,
|
| 388 |
+
"tie_word_embeddings": true,
|
| 389 |
+
"transformers_version": "5.0.0",
|
| 390 |
+
"use_cache": false,
|
| 391 |
+
"vocab_size": 50368
|
| 392 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fa9ed8cb62eea64c544112cb7f00a2fed2f57b3a1393d0bee7a430002c87e096
|
| 3 |
+
size 598907344
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1f562983391125d3ac0ace4d191d7d312a0bed2ab3d25272af12aac6734cf231
|
| 3 |
+
size 5265
|