AhChat commited on
Commit
5d99eaf
·
verified ·
1 Parent(s): b3f595a

Training in progress, epoch 1

Browse files
Files changed (3) hide show
  1. config.json +360 -0
  2. model.safetensors +3 -0
  3. training_args.bin +3 -0
config.json ADDED
@@ -0,0 +1,360 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "embedding_dropout": 0.0,
16
+ "eos_token_id": 50282,
17
+ "global_attn_every_n_layers": 3,
18
+ "global_rope_theta": 160000.0,
19
+ "gradient_checkpointing": false,
20
+ "hidden_activation": "gelu",
21
+ "hidden_size": 768,
22
+ "id2label": {
23
+ "0": "AGS Script",
24
+ "1": "Ada",
25
+ "10": "C#",
26
+ "100": "Perl",
27
+ "101": "Pickle",
28
+ "102": "PicoLisp",
29
+ "103": "PlantUML",
30
+ "104": "PostScript",
31
+ "105": "PowerShell",
32
+ "106": "Prolog",
33
+ "107": "Protocol Buffer Text Format",
34
+ "108": "PureScript",
35
+ "109": "Python",
36
+ "11": "C++",
37
+ "110": "R",
38
+ "111": "RDoc",
39
+ "112": "REALbasic",
40
+ "113": "Ragel in Ruby Host",
41
+ "114": "Rascal",
42
+ "115": "Raw token data",
43
+ "116": "Redcode",
44
+ "117": "Ruby",
45
+ "118": "Rust",
46
+ "119": "SCSS",
47
+ "12": "CMake",
48
+ "120": "SQF",
49
+ "121": "SQL",
50
+ "122": "Sass",
51
+ "123": "Scala",
52
+ "124": "Scheme",
53
+ "125": "Scilab",
54
+ "126": "Shell",
55
+ "127": "Smali",
56
+ "128": "Solidity",
57
+ "129": "Squirrel",
58
+ "13": "CODEOWNERS",
59
+ "130": "Starlark",
60
+ "131": "Stylus",
61
+ "132": "Swift",
62
+ "133": "TSV",
63
+ "134": "TSX",
64
+ "135": "Thrift",
65
+ "136": "Turtle",
66
+ "137": "Twig",
67
+ "138": "TypeScript",
68
+ "139": "Unity3D Asset",
69
+ "14": "Chapel",
70
+ "140": "Unix Assembly",
71
+ "141": "Unknown",
72
+ "142": "VBScript",
73
+ "143": "VCL",
74
+ "144": "VHDL",
75
+ "145": "Verilog",
76
+ "146": "Vim Script",
77
+ "147": "Vim Snippet",
78
+ "148": "Visual Basic",
79
+ "149": "Visual Basic .NET",
80
+ "15": "CoffeeScript",
81
+ "150": "Vue",
82
+ "151": "Wavefront Object",
83
+ "152": "XML Property List",
84
+ "153": "XS",
85
+ "154": "YANG",
86
+ "16": "ColdFusion",
87
+ "17": "Common Lisp",
88
+ "18": "Crystal",
89
+ "19": "Csound",
90
+ "2": "Adobe Font Metrics",
91
+ "20": "Csound Document",
92
+ "21": "D",
93
+ "22": "DIGITAL Command Language",
94
+ "23": "Dart",
95
+ "24": "Dockerfile",
96
+ "25": "ECL",
97
+ "26": "EJS",
98
+ "27": "Eagle",
99
+ "28": "Edoid",
100
+ "29": "Erlang",
101
+ "3": "ApacheConf",
102
+ "30": "Fish",
103
+ "31": "Fluent",
104
+ "32": "Fortran Free Form",
105
+ "33": "FreeBasic",
106
+ "34": "FreeMarker",
107
+ "35": "G-code",
108
+ "36": "GAP",
109
+ "37": "GAS",
110
+ "38": "GDB",
111
+ "39": "GDScript",
112
+ "4": "Apex",
113
+ "40": "GLSL",
114
+ "41": "Gerber Image",
115
+ "42": "Gettext Catalog",
116
+ "43": "Gherkin",
117
+ "44": "Git Config",
118
+ "45": "Gnuplot",
119
+ "46": "Go",
120
+ "47": "Go Module",
121
+ "48": "Gradle",
122
+ "49": "Graphviz (DOT)",
123
+ "5": "AsciiDoc",
124
+ "50": "Groovy",
125
+ "51": "HCL",
126
+ "52": "HTML+Razor",
127
+ "53": "Hack",
128
+ "54": "Haskell",
129
+ "55": "Haxe",
130
+ "56": "Hoon",
131
+ "57": "Ignore List",
132
+ "58": "ImageJ Macro",
133
+ "59": "Inform 7",
134
+ "6": "Assembly",
135
+ "60": "Inno Setup",
136
+ "61": "Io",
137
+ "62": "Ioke",
138
+ "63": "Isabelle",
139
+ "64": "J",
140
+ "65": "JAR Manifest",
141
+ "66": "JQ",
142
+ "67": "Java",
143
+ "68": "JavaScript",
144
+ "69": "Julia",
145
+ "7": "Blade",
146
+ "70": "Jupyter Notebook",
147
+ "71": "KiCad Layout",
148
+ "72": "Kit",
149
+ "73": "Kotlin",
150
+ "74": "Kvlang",
151
+ "75": "LLVM",
152
+ "76": "Lean",
153
+ "77": "Less",
154
+ "78": "Lex",
155
+ "79": "Logtalk",
156
+ "8": "Bluespec",
157
+ "80": "LookML",
158
+ "81": "Lua",
159
+ "82": "MATLAB",
160
+ "83": "Makefile",
161
+ "84": "Mathematica",
162
+ "85": "Metal",
163
+ "86": "Modelica",
164
+ "87": "Nim",
165
+ "88": "OCaml",
166
+ "89": "ObjDump",
167
+ "9": "C",
168
+ "90": "Objective-C",
169
+ "91": "Objective-C++",
170
+ "92": "Open Policy Agent",
171
+ "93": "OpenEdge ABL",
172
+ "94": "OpenStep Property List",
173
+ "95": "OpenType Feature File",
174
+ "96": "PHP",
175
+ "97": "PLpgSQL",
176
+ "98": "POV-Ray SDL",
177
+ "99": "Pascal"
178
+ },
179
+ "initializer_cutoff_factor": 2.0,
180
+ "initializer_range": 0.02,
181
+ "intermediate_size": 1152,
182
+ "label2id": {
183
+ "AGS Script": 0,
184
+ "Ada": 1,
185
+ "Adobe Font Metrics": 2,
186
+ "ApacheConf": 3,
187
+ "Apex": 4,
188
+ "AsciiDoc": 5,
189
+ "Assembly": 6,
190
+ "Blade": 7,
191
+ "Bluespec": 8,
192
+ "C": 9,
193
+ "C#": 10,
194
+ "C++": 11,
195
+ "CMake": 12,
196
+ "CODEOWNERS": 13,
197
+ "Chapel": 14,
198
+ "CoffeeScript": 15,
199
+ "ColdFusion": 16,
200
+ "Common Lisp": 17,
201
+ "Crystal": 18,
202
+ "Csound": 19,
203
+ "Csound Document": 20,
204
+ "D": 21,
205
+ "DIGITAL Command Language": 22,
206
+ "Dart": 23,
207
+ "Dockerfile": 24,
208
+ "ECL": 25,
209
+ "EJS": 26,
210
+ "Eagle": 27,
211
+ "Edoid": 28,
212
+ "Erlang": 29,
213
+ "Fish": 30,
214
+ "Fluent": 31,
215
+ "Fortran Free Form": 32,
216
+ "FreeBasic": 33,
217
+ "FreeMarker": 34,
218
+ "G-code": 35,
219
+ "GAP": 36,
220
+ "GAS": 37,
221
+ "GDB": 38,
222
+ "GDScript": 39,
223
+ "GLSL": 40,
224
+ "Gerber Image": 41,
225
+ "Gettext Catalog": 42,
226
+ "Gherkin": 43,
227
+ "Git Config": 44,
228
+ "Gnuplot": 45,
229
+ "Go": 46,
230
+ "Go Module": 47,
231
+ "Gradle": 48,
232
+ "Graphviz (DOT)": 49,
233
+ "Groovy": 50,
234
+ "HCL": 51,
235
+ "HTML+Razor": 52,
236
+ "Hack": 53,
237
+ "Haskell": 54,
238
+ "Haxe": 55,
239
+ "Hoon": 56,
240
+ "Ignore List": 57,
241
+ "ImageJ Macro": 58,
242
+ "Inform 7": 59,
243
+ "Inno Setup": 60,
244
+ "Io": 61,
245
+ "Ioke": 62,
246
+ "Isabelle": 63,
247
+ "J": 64,
248
+ "JAR Manifest": 65,
249
+ "JQ": 66,
250
+ "Java": 67,
251
+ "JavaScript": 68,
252
+ "Julia": 69,
253
+ "Jupyter Notebook": 70,
254
+ "KiCad Layout": 71,
255
+ "Kit": 72,
256
+ "Kotlin": 73,
257
+ "Kvlang": 74,
258
+ "LLVM": 75,
259
+ "Lean": 76,
260
+ "Less": 77,
261
+ "Lex": 78,
262
+ "Logtalk": 79,
263
+ "LookML": 80,
264
+ "Lua": 81,
265
+ "MATLAB": 82,
266
+ "Makefile": 83,
267
+ "Mathematica": 84,
268
+ "Metal": 85,
269
+ "Modelica": 86,
270
+ "Nim": 87,
271
+ "OCaml": 88,
272
+ "ObjDump": 89,
273
+ "Objective-C": 90,
274
+ "Objective-C++": 91,
275
+ "Open Policy Agent": 92,
276
+ "OpenEdge ABL": 93,
277
+ "OpenStep Property List": 94,
278
+ "OpenType Feature File": 95,
279
+ "PHP": 96,
280
+ "PLpgSQL": 97,
281
+ "POV-Ray SDL": 98,
282
+ "Pascal": 99,
283
+ "Perl": 100,
284
+ "Pickle": 101,
285
+ "PicoLisp": 102,
286
+ "PlantUML": 103,
287
+ "PostScript": 104,
288
+ "PowerShell": 105,
289
+ "Prolog": 106,
290
+ "Protocol Buffer Text Format": 107,
291
+ "PureScript": 108,
292
+ "Python": 109,
293
+ "R": 110,
294
+ "RDoc": 111,
295
+ "REALbasic": 112,
296
+ "Ragel in Ruby Host": 113,
297
+ "Rascal": 114,
298
+ "Raw token data": 115,
299
+ "Redcode": 116,
300
+ "Ruby": 117,
301
+ "Rust": 118,
302
+ "SCSS": 119,
303
+ "SQF": 120,
304
+ "SQL": 121,
305
+ "Sass": 122,
306
+ "Scala": 123,
307
+ "Scheme": 124,
308
+ "Scilab": 125,
309
+ "Shell": 126,
310
+ "Smali": 127,
311
+ "Solidity": 128,
312
+ "Squirrel": 129,
313
+ "Starlark": 130,
314
+ "Stylus": 131,
315
+ "Swift": 132,
316
+ "TSV": 133,
317
+ "TSX": 134,
318
+ "Thrift": 135,
319
+ "Turtle": 136,
320
+ "Twig": 137,
321
+ "TypeScript": 138,
322
+ "Unity3D Asset": 139,
323
+ "Unix Assembly": 140,
324
+ "Unknown": 141,
325
+ "VBScript": 142,
326
+ "VCL": 143,
327
+ "VHDL": 144,
328
+ "Verilog": 145,
329
+ "Vim Script": 146,
330
+ "Vim Snippet": 147,
331
+ "Visual Basic": 148,
332
+ "Visual Basic .NET": 149,
333
+ "Vue": 150,
334
+ "Wavefront Object": 151,
335
+ "XML Property List": 152,
336
+ "XS": 153,
337
+ "YANG": 154
338
+ },
339
+ "layer_norm_eps": 1e-05,
340
+ "local_attention": 128,
341
+ "local_rope_theta": 10000.0,
342
+ "max_position_embeddings": 8192,
343
+ "mlp_bias": false,
344
+ "mlp_dropout": 0.0,
345
+ "model_type": "modernbert",
346
+ "norm_bias": false,
347
+ "norm_eps": 1e-05,
348
+ "num_attention_heads": 12,
349
+ "num_hidden_layers": 22,
350
+ "pad_token_id": 50283,
351
+ "position_embedding_type": "absolute",
352
+ "problem_type": "single_label_classification",
353
+ "repad_logits_with_grad": false,
354
+ "sep_token_id": 50282,
355
+ "sparse_pred_ignore_index": -100,
356
+ "sparse_prediction": false,
357
+ "torch_dtype": "float32",
358
+ "transformers_version": "4.53.1",
359
+ "vocab_size": 50368
360
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e6e005865274685ab8ab57618c69e4fab963cab2938e284e7025c52f433ba1e4
3
+ size 598910420
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c80b968eb48089b74361b467b0d3a654f9e4c7dc953008148f2e8ad46289562
3
+ size 5841