CesarLeblanc commited on
Commit
51f1003
1 Parent(s): 007bb61
eunis_habitats.xlsx → data/eunis_habitats.xlsx RENAMED
File without changes
plantbert_fill_mask_model/added_tokens.json DELETED
The diff for this file is too large to render. See raw diff
 
plantbert_fill_mask_model/config.json DELETED
@@ -1,26 +0,0 @@
1
- {
2
- "_name_or_path": "../Models/bert-base-uncased/",
3
- "architectures": [
4
- "BertForMaskedLM"
5
- ],
6
- "attention_probs_dropout_prob": 0.1,
7
- "classifier_dropout": null,
8
- "gradient_checkpointing": false,
9
- "hidden_act": "gelu",
10
- "hidden_dropout_prob": 0.1,
11
- "hidden_size": 768,
12
- "initializer_range": 0.02,
13
- "intermediate_size": 3072,
14
- "layer_norm_eps": 1e-12,
15
- "max_position_embeddings": 512,
16
- "model_type": "bert",
17
- "num_attention_heads": 12,
18
- "num_hidden_layers": 12,
19
- "pad_token_id": 0,
20
- "position_embedding_type": "absolute",
21
- "torch_dtype": "float32",
22
- "transformers_version": "4.36.2",
23
- "type_vocab_size": 2,
24
- "use_cache": true,
25
- "vocab_size": 44708
26
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
plantbert_fill_mask_model/generation_config.json DELETED
@@ -1,5 +0,0 @@
1
- {
2
- "_from_model_config": true,
3
- "pad_token_id": 0,
4
- "transformers_version": "4.36.2"
5
- }
 
 
 
 
 
 
plantbert_fill_mask_model/model.safetensors DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:e0fc9a4d97cc9ee7e6416c93681c4fa724df57051066e1bce5ad7cb172f4ef2f
3
- size 481717040
 
 
 
 
plantbert_fill_mask_model/special_tokens_map.json DELETED
@@ -1,37 +0,0 @@
1
- {
2
- "cls_token": {
3
- "content": "[CLS]",
4
- "lstrip": false,
5
- "normalized": false,
6
- "rstrip": false,
7
- "single_word": false
8
- },
9
- "mask_token": {
10
- "content": "[MASK]",
11
- "lstrip": false,
12
- "normalized": false,
13
- "rstrip": false,
14
- "single_word": false
15
- },
16
- "pad_token": {
17
- "content": "[PAD]",
18
- "lstrip": false,
19
- "normalized": false,
20
- "rstrip": false,
21
- "single_word": false
22
- },
23
- "sep_token": {
24
- "content": "[SEP]",
25
- "lstrip": false,
26
- "normalized": false,
27
- "rstrip": false,
28
- "single_word": false
29
- },
30
- "unk_token": {
31
- "content": "[UNK]",
32
- "lstrip": false,
33
- "normalized": false,
34
- "rstrip": false,
35
- "single_word": false
36
- }
37
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
plantbert_fill_mask_model/tokenizer.json DELETED
The diff for this file is too large to render. See raw diff
 
plantbert_fill_mask_model/tokenizer_config.json DELETED
The diff for this file is too large to render. See raw diff
 
plantbert_fill_mask_model/vocab.txt DELETED
The diff for this file is too large to render. See raw diff
 
plantbert_text_classification_model/added_tokens.json DELETED
The diff for this file is too large to render. See raw diff
 
plantbert_text_classification_model/config.json DELETED
@@ -1,485 +0,0 @@
1
- {
2
- "_name_or_path": "../Models/plantbert_fill_mask_model_base-species_2_2e-05/",
3
- "architectures": [
4
- "BertForSequenceClassification"
5
- ],
6
- "attention_probs_dropout_prob": 0.1,
7
- "classifier_dropout": null,
8
- "gradient_checkpointing": false,
9
- "hidden_act": "gelu",
10
- "hidden_dropout_prob": 0.1,
11
- "hidden_size": 768,
12
- "id2label": {
13
- "0": "MA211",
14
- "1": "MA221",
15
- "2": "MA222",
16
- "3": "MA223",
17
- "4": "MA224",
18
- "5": "MA225",
19
- "6": "MA232",
20
- "7": "MA241",
21
- "8": "MA251",
22
- "9": "MA252",
23
- "10": "MA253",
24
- "11": "N11",
25
- "12": "N12",
26
- "13": "N13",
27
- "14": "N14",
28
- "15": "N15",
29
- "16": "N16",
30
- "17": "N17",
31
- "18": "N18",
32
- "19": "N19",
33
- "20": "N1A",
34
- "21": "N1B",
35
- "22": "N1C",
36
- "23": "N1D",
37
- "24": "N1E",
38
- "25": "N1F",
39
- "26": "N1G",
40
- "27": "N1H",
41
- "28": "N1J",
42
- "29": "N21",
43
- "30": "N22",
44
- "31": "N31",
45
- "32": "N32",
46
- "33": "N33",
47
- "34": "N34",
48
- "35": "N35",
49
- "36": "Q11",
50
- "37": "Q12",
51
- "38": "Q21",
52
- "39": "Q22",
53
- "40": "Q23",
54
- "41": "Q24",
55
- "42": "Q25",
56
- "43": "Q41",
57
- "44": "Q42",
58
- "45": "Q43",
59
- "46": "Q44",
60
- "47": "Q45",
61
- "48": "Q46",
62
- "49": "Q51",
63
- "50": "Q52",
64
- "51": "Q53",
65
- "52": "Q54",
66
- "53": "R11",
67
- "54": "R12",
68
- "55": "R13",
69
- "56": "R14",
70
- "57": "R15",
71
- "58": "R16",
72
- "59": "R17",
73
- "60": "R18",
74
- "61": "R19",
75
- "62": "R1A",
76
- "63": "R1B",
77
- "64": "R1C",
78
- "65": "R1D",
79
- "66": "R1E",
80
- "67": "R1F",
81
- "68": "R1G",
82
- "69": "R1H",
83
- "70": "R1J",
84
- "71": "R1K",
85
- "72": "R1M",
86
- "73": "R1P",
87
- "74": "R1Q",
88
- "75": "R1R",
89
- "76": "R1S",
90
- "77": "R21",
91
- "78": "R22",
92
- "79": "R23",
93
- "80": "R24",
94
- "81": "R31",
95
- "82": "R32",
96
- "83": "R33",
97
- "84": "R34",
98
- "85": "R35",
99
- "86": "R36",
100
- "87": "R37",
101
- "88": "R41",
102
- "89": "R42",
103
- "90": "R43",
104
- "91": "R44",
105
- "92": "R45",
106
- "93": "R51",
107
- "94": "R52",
108
- "95": "R53",
109
- "96": "R54",
110
- "97": "R55",
111
- "98": "R56",
112
- "99": "R57",
113
- "100": "R61",
114
- "101": "R62",
115
- "102": "R63",
116
- "103": "R64",
117
- "104": "R65",
118
- "105": "S11",
119
- "106": "S12",
120
- "107": "S21",
121
- "108": "S22",
122
- "109": "S23",
123
- "110": "S24",
124
- "111": "S25",
125
- "112": "S26",
126
- "113": "S31",
127
- "114": "S32",
128
- "115": "S33",
129
- "116": "S34",
130
- "117": "S35",
131
- "118": "S36",
132
- "119": "S37",
133
- "120": "S38",
134
- "121": "S41",
135
- "122": "S42",
136
- "123": "S51",
137
- "124": "S52",
138
- "125": "S53",
139
- "126": "S54",
140
- "127": "S61",
141
- "128": "S62",
142
- "129": "S63",
143
- "130": "S64",
144
- "131": "S65",
145
- "132": "S66",
146
- "133": "S67",
147
- "134": "S68",
148
- "135": "S71",
149
- "136": "S72",
150
- "137": "S73",
151
- "138": "S74",
152
- "139": "S75",
153
- "140": "S76",
154
- "141": "S81",
155
- "142": "S82",
156
- "143": "S91",
157
- "144": "S92",
158
- "145": "S93",
159
- "146": "S94",
160
- "147": "T11",
161
- "148": "T12",
162
- "149": "T13",
163
- "150": "T14",
164
- "151": "T15",
165
- "152": "T16",
166
- "153": "T17",
167
- "154": "T18",
168
- "155": "T19",
169
- "156": "T1A",
170
- "157": "T1B",
171
- "158": "T1C",
172
- "159": "T1D",
173
- "160": "T1E",
174
- "161": "T1F",
175
- "162": "T1G",
176
- "163": "T1H",
177
- "164": "T21",
178
- "165": "T22",
179
- "166": "T23",
180
- "167": "T24",
181
- "168": "T25",
182
- "169": "T27",
183
- "170": "T28",
184
- "171": "T29",
185
- "172": "T31",
186
- "173": "T32",
187
- "174": "T33",
188
- "175": "T34",
189
- "176": "T35",
190
- "177": "T36",
191
- "178": "T37",
192
- "179": "T38",
193
- "180": "T39",
194
- "181": "T3A",
195
- "182": "T3B",
196
- "183": "T3C",
197
- "184": "T3D",
198
- "185": "T3E",
199
- "186": "T3F",
200
- "187": "T3G",
201
- "188": "T3H",
202
- "189": "T3J",
203
- "190": "T3K",
204
- "191": "T3M",
205
- "192": "U21",
206
- "193": "U22",
207
- "194": "U23",
208
- "195": "U24",
209
- "196": "U25",
210
- "197": "U26",
211
- "198": "U27",
212
- "199": "U28",
213
- "200": "U29",
214
- "201": "U2A",
215
- "202": "U32",
216
- "203": "U33",
217
- "204": "U34",
218
- "205": "U35",
219
- "206": "U36",
220
- "207": "U37",
221
- "208": "U38",
222
- "209": "U3A",
223
- "210": "U3B",
224
- "211": "U3C",
225
- "212": "U3D",
226
- "213": "U61",
227
- "214": "U62",
228
- "215": "V11",
229
- "216": "V12",
230
- "217": "V13",
231
- "218": "V14",
232
- "219": "V15",
233
- "220": "V32",
234
- "221": "V33",
235
- "222": "V34",
236
- "223": "V35",
237
- "224": "V37",
238
- "225": "V38",
239
- "226": "V39"
240
- },
241
- "initializer_range": 0.02,
242
- "intermediate_size": 3072,
243
- "label2id": {
244
- "MA211": 0,
245
- "MA221": 1,
246
- "MA222": 2,
247
- "MA223": 3,
248
- "MA224": 4,
249
- "MA225": 5,
250
- "MA232": 6,
251
- "MA241": 7,
252
- "MA251": 8,
253
- "MA252": 9,
254
- "MA253": 10,
255
- "N11": 11,
256
- "N12": 12,
257
- "N13": 13,
258
- "N14": 14,
259
- "N15": 15,
260
- "N16": 16,
261
- "N17": 17,
262
- "N18": 18,
263
- "N19": 19,
264
- "N1A": 20,
265
- "N1B": 21,
266
- "N1C": 22,
267
- "N1D": 23,
268
- "N1E": 24,
269
- "N1F": 25,
270
- "N1G": 26,
271
- "N1H": 27,
272
- "N1J": 28,
273
- "N21": 29,
274
- "N22": 30,
275
- "N31": 31,
276
- "N32": 32,
277
- "N33": 33,
278
- "N34": 34,
279
- "N35": 35,
280
- "Q11": 36,
281
- "Q12": 37,
282
- "Q21": 38,
283
- "Q22": 39,
284
- "Q23": 40,
285
- "Q24": 41,
286
- "Q25": 42,
287
- "Q41": 43,
288
- "Q42": 44,
289
- "Q43": 45,
290
- "Q44": 46,
291
- "Q45": 47,
292
- "Q46": 48,
293
- "Q51": 49,
294
- "Q52": 50,
295
- "Q53": 51,
296
- "Q54": 52,
297
- "R11": 53,
298
- "R12": 54,
299
- "R13": 55,
300
- "R14": 56,
301
- "R15": 57,
302
- "R16": 58,
303
- "R17": 59,
304
- "R18": 60,
305
- "R19": 61,
306
- "R1A": 62,
307
- "R1B": 63,
308
- "R1C": 64,
309
- "R1D": 65,
310
- "R1E": 66,
311
- "R1F": 67,
312
- "R1G": 68,
313
- "R1H": 69,
314
- "R1J": 70,
315
- "R1K": 71,
316
- "R1M": 72,
317
- "R1P": 73,
318
- "R1Q": 74,
319
- "R1R": 75,
320
- "R1S": 76,
321
- "R21": 77,
322
- "R22": 78,
323
- "R23": 79,
324
- "R24": 80,
325
- "R31": 81,
326
- "R32": 82,
327
- "R33": 83,
328
- "R34": 84,
329
- "R35": 85,
330
- "R36": 86,
331
- "R37": 87,
332
- "R41": 88,
333
- "R42": 89,
334
- "R43": 90,
335
- "R44": 91,
336
- "R45": 92,
337
- "R51": 93,
338
- "R52": 94,
339
- "R53": 95,
340
- "R54": 96,
341
- "R55": 97,
342
- "R56": 98,
343
- "R57": 99,
344
- "R61": 100,
345
- "R62": 101,
346
- "R63": 102,
347
- "R64": 103,
348
- "R65": 104,
349
- "S11": 105,
350
- "S12": 106,
351
- "S21": 107,
352
- "S22": 108,
353
- "S23": 109,
354
- "S24": 110,
355
- "S25": 111,
356
- "S26": 112,
357
- "S31": 113,
358
- "S32": 114,
359
- "S33": 115,
360
- "S34": 116,
361
- "S35": 117,
362
- "S36": 118,
363
- "S37": 119,
364
- "S38": 120,
365
- "S41": 121,
366
- "S42": 122,
367
- "S51": 123,
368
- "S52": 124,
369
- "S53": 125,
370
- "S54": 126,
371
- "S61": 127,
372
- "S62": 128,
373
- "S63": 129,
374
- "S64": 130,
375
- "S65": 131,
376
- "S66": 132,
377
- "S67": 133,
378
- "S68": 134,
379
- "S71": 135,
380
- "S72": 136,
381
- "S73": 137,
382
- "S74": 138,
383
- "S75": 139,
384
- "S76": 140,
385
- "S81": 141,
386
- "S82": 142,
387
- "S91": 143,
388
- "S92": 144,
389
- "S93": 145,
390
- "S94": 146,
391
- "T11": 147,
392
- "T12": 148,
393
- "T13": 149,
394
- "T14": 150,
395
- "T15": 151,
396
- "T16": 152,
397
- "T17": 153,
398
- "T18": 154,
399
- "T19": 155,
400
- "T1A": 156,
401
- "T1B": 157,
402
- "T1C": 158,
403
- "T1D": 159,
404
- "T1E": 160,
405
- "T1F": 161,
406
- "T1G": 162,
407
- "T1H": 163,
408
- "T21": 164,
409
- "T22": 165,
410
- "T23": 166,
411
- "T24": 167,
412
- "T25": 168,
413
- "T27": 169,
414
- "T28": 170,
415
- "T29": 171,
416
- "T31": 172,
417
- "T32": 173,
418
- "T33": 174,
419
- "T34": 175,
420
- "T35": 176,
421
- "T36": 177,
422
- "T37": 178,
423
- "T38": 179,
424
- "T39": 180,
425
- "T3A": 181,
426
- "T3B": 182,
427
- "T3C": 183,
428
- "T3D": 184,
429
- "T3E": 185,
430
- "T3F": 186,
431
- "T3G": 187,
432
- "T3H": 188,
433
- "T3J": 189,
434
- "T3K": 190,
435
- "T3M": 191,
436
- "U21": 192,
437
- "U22": 193,
438
- "U23": 194,
439
- "U24": 195,
440
- "U25": 196,
441
- "U26": 197,
442
- "U27": 198,
443
- "U28": 199,
444
- "U29": 200,
445
- "U2A": 201,
446
- "U32": 202,
447
- "U33": 203,
448
- "U34": 204,
449
- "U35": 205,
450
- "U36": 206,
451
- "U37": 207,
452
- "U38": 208,
453
- "U3A": 209,
454
- "U3B": 210,
455
- "U3C": 211,
456
- "U3D": 212,
457
- "U61": 213,
458
- "U62": 214,
459
- "V11": 215,
460
- "V12": 216,
461
- "V13": 217,
462
- "V14": 218,
463
- "V15": 219,
464
- "V32": 220,
465
- "V33": 221,
466
- "V34": 222,
467
- "V35": 223,
468
- "V37": 224,
469
- "V38": 225,
470
- "V39": 226
471
- },
472
- "layer_norm_eps": 1e-12,
473
- "max_position_embeddings": 512,
474
- "model_type": "bert",
475
- "num_attention_heads": 12,
476
- "num_hidden_layers": 12,
477
- "pad_token_id": 0,
478
- "position_embedding_type": "absolute",
479
- "problem_type": "single_label_classification",
480
- "torch_dtype": "float32",
481
- "transformers_version": "4.36.2",
482
- "type_vocab_size": 2,
483
- "use_cache": true,
484
- "vocab_size": 44708
485
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
plantbert_text_classification_model/model.safetensors DELETED
@@ -1,3 +0,0 @@
1
- version https://git-lfs.github.com/spec/v1
2
- oid sha256:eeb123d2b9fbb1c95c954bb2dee5605c6a644affbce21ebd3fec188769f460d6
3
- size 482230156
 
 
 
 
plantbert_text_classification_model/special_tokens_map.json DELETED
@@ -1,37 +0,0 @@
1
- {
2
- "cls_token": {
3
- "content": "[CLS]",
4
- "lstrip": false,
5
- "normalized": false,
6
- "rstrip": false,
7
- "single_word": false
8
- },
9
- "mask_token": {
10
- "content": "[MASK]",
11
- "lstrip": false,
12
- "normalized": false,
13
- "rstrip": false,
14
- "single_word": false
15
- },
16
- "pad_token": {
17
- "content": "[PAD]",
18
- "lstrip": false,
19
- "normalized": false,
20
- "rstrip": false,
21
- "single_word": false
22
- },
23
- "sep_token": {
24
- "content": "[SEP]",
25
- "lstrip": false,
26
- "normalized": false,
27
- "rstrip": false,
28
- "single_word": false
29
- },
30
- "unk_token": {
31
- "content": "[UNK]",
32
- "lstrip": false,
33
- "normalized": false,
34
- "rstrip": false,
35
- "single_word": false
36
- }
37
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
plantbert_text_classification_model/tokenizer.json DELETED
The diff for this file is too large to render. See raw diff
 
plantbert_text_classification_model/tokenizer_config.json DELETED
The diff for this file is too large to render. See raw diff
 
plantbert_text_classification_model/vocab.txt DELETED
The diff for this file is too large to render. See raw diff