carlyou commited on
Commit
3782c26
·
verified ·
1 Parent(s): 94d4f36

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
+ "dtype": "float32",
16
+ "embedding_dropout": 0.0,
17
+ "eos_token_id": 50282,
18
+ "global_attn_every_n_layers": 3,
19
+ "global_rope_theta": 160000.0,
20
+ "gradient_checkpointing": false,
21
+ "hidden_activation": "gelu",
22
+ "hidden_size": 768,
23
+ "id2label": {
24
+ "0": "GAP",
25
+ "1": "Squirrel",
26
+ "2": "Objective-C",
27
+ "3": "Gerber Image",
28
+ "4": "Open Policy Agent",
29
+ "5": "Jupyter Notebook",
30
+ "6": "Rascal",
31
+ "7": "Starlark",
32
+ "8": "ImageJ Macro",
33
+ "9": "PLpgSQL",
34
+ "10": "SCSS",
35
+ "11": "Eagle",
36
+ "12": "Bluespec",
37
+ "13": "C++",
38
+ "14": "PHP",
39
+ "15": "GDScript",
40
+ "16": "ApacheConf",
41
+ "17": "Modelica",
42
+ "18": "Pickle",
43
+ "19": "MATLAB",
44
+ "20": "Kit",
45
+ "21": "Erlang",
46
+ "22": "DIGITAL Command Language",
47
+ "23": "TypeScript",
48
+ "24": "Lua",
49
+ "25": "LLVM",
50
+ "26": "Stylus",
51
+ "27": "Lean",
52
+ "28": "Gradle",
53
+ "29": "Vim Snippet",
54
+ "30": "Fortran Free Form",
55
+ "31": "Vue",
56
+ "32": "Swift",
57
+ "33": "GLSL",
58
+ "34": "TSX",
59
+ "35": "Objective-C++",
60
+ "36": "Prolog",
61
+ "37": "Dockerfile",
62
+ "38": "Twig",
63
+ "39": "Wavefront Object",
64
+ "40": "Go",
65
+ "41": "Makefile",
66
+ "42": "XML Property List",
67
+ "43": "CODEOWNERS",
68
+ "44": "Inform 7",
69
+ "45": "Unix Assembly",
70
+ "46": "Gettext Catalog",
71
+ "47": "CMake",
72
+ "48": "XS",
73
+ "49": "Solidity",
74
+ "50": "Lex",
75
+ "51": "AGS Script",
76
+ "52": "Groovy",
77
+ "53": "C#",
78
+ "54": "Thrift",
79
+ "55": "Edoid",
80
+ "56": "Pascal",
81
+ "57": "Fluent",
82
+ "58": "PowerShell",
83
+ "59": "Java",
84
+ "60": "VBScript",
85
+ "61": "YANG",
86
+ "62": "Raw token data",
87
+ "63": "Haskell",
88
+ "64": "Hoon",
89
+ "65": "Mathematica",
90
+ "66": "SQF",
91
+ "67": "Inno Setup",
92
+ "68": "Ragel in Ruby Host",
93
+ "69": "Go Module",
94
+ "70": "Vim Script",
95
+ "71": "J",
96
+ "72": "ObjDump",
97
+ "73": "PicoLisp",
98
+ "74": "HTML+Razor",
99
+ "75": "Gnuplot",
100
+ "76": "FreeMarker",
101
+ "77": "Assembly",
102
+ "78": "Scala",
103
+ "79": "RDoc",
104
+ "80": "OpenType Feature File",
105
+ "81": "ColdFusion",
106
+ "82": "Julia",
107
+ "83": "OpenStep Property List",
108
+ "84": "Csound",
109
+ "85": "GDB",
110
+ "86": "Dart",
111
+ "87": "Git Config",
112
+ "88": "Gherkin",
113
+ "89": "JAR Manifest",
114
+ "90": "AsciiDoc",
115
+ "91": "Visual Basic .NET",
116
+ "92": "Csound Document",
117
+ "93": "HCL",
118
+ "94": "OCaml",
119
+ "95": "Io",
120
+ "96": "PlantUML",
121
+ "97": "Nim",
122
+ "98": "Rust",
123
+ "99": "VCL",
124
+ "100": "G-code",
125
+ "101": "Metal",
126
+ "102": "Visual Basic",
127
+ "103": "GAS",
128
+ "104": "Adobe Font Metrics",
129
+ "105": "Scilab",
130
+ "106": "Shell",
131
+ "107": "TSV",
132
+ "108": "Ignore List",
133
+ "109": "Ruby",
134
+ "110": "KiCad Layout",
135
+ "111": "LookML",
136
+ "112": "ECL",
137
+ "113": "D",
138
+ "114": "Unity3D Asset",
139
+ "115": "PostScript",
140
+ "116": "R",
141
+ "117": "JQ",
142
+ "118": "Blade",
143
+ "119": "EJS",
144
+ "120": "C",
145
+ "121": "Chapel",
146
+ "122": "Redcode",
147
+ "123": "SQL",
148
+ "124": "PureScript",
149
+ "125": "FreeBasic",
150
+ "126": "CoffeeScript",
151
+ "127": "Scheme",
152
+ "128": "Unknown",
153
+ "129": "Haxe",
154
+ "130": "Sass",
155
+ "131": "Kotlin",
156
+ "132": "Protocol Buffer Text Format",
157
+ "133": "Fish",
158
+ "134": "OpenEdge ABL",
159
+ "135": "Less",
160
+ "136": "Ioke",
161
+ "137": "Common Lisp",
162
+ "138": "Kvlang",
163
+ "139": "Python",
164
+ "140": "REALbasic",
165
+ "141": "Hack",
166
+ "142": "Logtalk",
167
+ "143": "VHDL",
168
+ "144": "Apex",
169
+ "145": "Turtle",
170
+ "146": "POV-Ray SDL",
171
+ "147": "Ada",
172
+ "148": "Crystal",
173
+ "149": "Graphviz (DOT)",
174
+ "150": "Verilog",
175
+ "151": "JavaScript",
176
+ "152": "Smali",
177
+ "153": "Perl",
178
+ "154": "Isabelle"
179
+ },
180
+ "initializer_cutoff_factor": 2.0,
181
+ "initializer_range": 0.02,
182
+ "intermediate_size": 1152,
183
+ "label2id": {
184
+ "AGS Script": 51,
185
+ "Ada": 147,
186
+ "Adobe Font Metrics": 104,
187
+ "ApacheConf": 16,
188
+ "Apex": 144,
189
+ "AsciiDoc": 90,
190
+ "Assembly": 77,
191
+ "Blade": 118,
192
+ "Bluespec": 12,
193
+ "C": 120,
194
+ "C#": 53,
195
+ "C++": 13,
196
+ "CMake": 47,
197
+ "CODEOWNERS": 43,
198
+ "Chapel": 121,
199
+ "CoffeeScript": 126,
200
+ "ColdFusion": 81,
201
+ "Common Lisp": 137,
202
+ "Crystal": 148,
203
+ "Csound": 84,
204
+ "Csound Document": 92,
205
+ "D": 113,
206
+ "DIGITAL Command Language": 22,
207
+ "Dart": 86,
208
+ "Dockerfile": 37,
209
+ "ECL": 112,
210
+ "EJS": 119,
211
+ "Eagle": 11,
212
+ "Edoid": 55,
213
+ "Erlang": 21,
214
+ "Fish": 133,
215
+ "Fluent": 57,
216
+ "Fortran Free Form": 30,
217
+ "FreeBasic": 125,
218
+ "FreeMarker": 76,
219
+ "G-code": 100,
220
+ "GAP": 0,
221
+ "GAS": 103,
222
+ "GDB": 85,
223
+ "GDScript": 15,
224
+ "GLSL": 33,
225
+ "Gerber Image": 3,
226
+ "Gettext Catalog": 46,
227
+ "Gherkin": 88,
228
+ "Git Config": 87,
229
+ "Gnuplot": 75,
230
+ "Go": 40,
231
+ "Go Module": 69,
232
+ "Gradle": 28,
233
+ "Graphviz (DOT)": 149,
234
+ "Groovy": 52,
235
+ "HCL": 93,
236
+ "HTML+Razor": 74,
237
+ "Hack": 141,
238
+ "Haskell": 63,
239
+ "Haxe": 129,
240
+ "Hoon": 64,
241
+ "Ignore List": 108,
242
+ "ImageJ Macro": 8,
243
+ "Inform 7": 44,
244
+ "Inno Setup": 67,
245
+ "Io": 95,
246
+ "Ioke": 136,
247
+ "Isabelle": 154,
248
+ "J": 71,
249
+ "JAR Manifest": 89,
250
+ "JQ": 117,
251
+ "Java": 59,
252
+ "JavaScript": 151,
253
+ "Julia": 82,
254
+ "Jupyter Notebook": 5,
255
+ "KiCad Layout": 110,
256
+ "Kit": 20,
257
+ "Kotlin": 131,
258
+ "Kvlang": 138,
259
+ "LLVM": 25,
260
+ "Lean": 27,
261
+ "Less": 135,
262
+ "Lex": 50,
263
+ "Logtalk": 142,
264
+ "LookML": 111,
265
+ "Lua": 24,
266
+ "MATLAB": 19,
267
+ "Makefile": 41,
268
+ "Mathematica": 65,
269
+ "Metal": 101,
270
+ "Modelica": 17,
271
+ "Nim": 97,
272
+ "OCaml": 94,
273
+ "ObjDump": 72,
274
+ "Objective-C": 2,
275
+ "Objective-C++": 35,
276
+ "Open Policy Agent": 4,
277
+ "OpenEdge ABL": 134,
278
+ "OpenStep Property List": 83,
279
+ "OpenType Feature File": 80,
280
+ "PHP": 14,
281
+ "PLpgSQL": 9,
282
+ "POV-Ray SDL": 146,
283
+ "Pascal": 56,
284
+ "Perl": 153,
285
+ "Pickle": 18,
286
+ "PicoLisp": 73,
287
+ "PlantUML": 96,
288
+ "PostScript": 115,
289
+ "PowerShell": 58,
290
+ "Prolog": 36,
291
+ "Protocol Buffer Text Format": 132,
292
+ "PureScript": 124,
293
+ "Python": 139,
294
+ "R": 116,
295
+ "RDoc": 79,
296
+ "REALbasic": 140,
297
+ "Ragel in Ruby Host": 68,
298
+ "Rascal": 6,
299
+ "Raw token data": 62,
300
+ "Redcode": 122,
301
+ "Ruby": 109,
302
+ "Rust": 98,
303
+ "SCSS": 10,
304
+ "SQF": 66,
305
+ "SQL": 123,
306
+ "Sass": 130,
307
+ "Scala": 78,
308
+ "Scheme": 127,
309
+ "Scilab": 105,
310
+ "Shell": 106,
311
+ "Smali": 152,
312
+ "Solidity": 49,
313
+ "Squirrel": 1,
314
+ "Starlark": 7,
315
+ "Stylus": 26,
316
+ "Swift": 32,
317
+ "TSV": 107,
318
+ "TSX": 34,
319
+ "Thrift": 54,
320
+ "Turtle": 145,
321
+ "Twig": 38,
322
+ "TypeScript": 23,
323
+ "Unity3D Asset": 114,
324
+ "Unix Assembly": 45,
325
+ "Unknown": 128,
326
+ "VBScript": 60,
327
+ "VCL": 99,
328
+ "VHDL": 143,
329
+ "Verilog": 150,
330
+ "Vim Script": 70,
331
+ "Vim Snippet": 29,
332
+ "Visual Basic": 102,
333
+ "Visual Basic .NET": 91,
334
+ "Vue": 31,
335
+ "Wavefront Object": 39,
336
+ "XML Property List": 42,
337
+ "XS": 48,
338
+ "YANG": 61
339
+ },
340
+ "layer_norm_eps": 1e-05,
341
+ "local_attention": 128,
342
+ "local_rope_theta": 10000.0,
343
+ "max_position_embeddings": 8192,
344
+ "mlp_bias": false,
345
+ "mlp_dropout": 0.0,
346
+ "model_type": "modernbert",
347
+ "norm_bias": false,
348
+ "norm_eps": 1e-05,
349
+ "num_attention_heads": 12,
350
+ "num_hidden_layers": 22,
351
+ "pad_token_id": 50283,
352
+ "position_embedding_type": "absolute",
353
+ "problem_type": "single_label_classification",
354
+ "repad_logits_with_grad": false,
355
+ "sep_token_id": 50282,
356
+ "sparse_pred_ignore_index": -100,
357
+ "sparse_prediction": false,
358
+ "transformers_version": "4.57.3",
359
+ "vocab_size": 50368
360
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:621c6c0a0405bdcc127f6f2c0a004b1f068f0a0021535a2de155ec7833288850
3
+ size 598910420
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6d72b0e334cce99dfb2c35b115fbd5c31b3a13498209d341bedfc29b60d7c978
3
+ size 5905