dotan1111 commited on
Commit
d2f1d34
·
verified ·
1 Parent(s): e385323

Upload TextClassificationPipeline

Browse files
Files changed (5) hide show
  1. config.json +811 -0
  2. model.safetensors +3 -0
  3. special_tokens_map.json +37 -0
  4. tokenizer_config.json +53 -0
  5. vocab.txt +33 -0
config.json ADDED
@@ -0,0 +1,811 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/groups/pupko/edodotan/PretrainAccelerator/validators/cell_location/models/results/facebook_esm2_t30_150M_UR50D_long/checkpoint-1200000/",
3
+ "architectures": [
4
+ "EsmForSequenceClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "classifier_dropout": null,
8
+ "emb_layer_norm_before": false,
9
+ "esmfold_config": null,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.0,
12
+ "hidden_size": 640,
13
+ "id2label": {
14
+ "0": "a band",
15
+ "1": "acidocalcisome membrane",
16
+ "2": "acrosome",
17
+ "3": "acrosome inner membrane",
18
+ "4": "acrosome lumen",
19
+ "5": "acrosome membrane",
20
+ "6": "actin patch",
21
+ "7": "adherens junction",
22
+ "8": "aleurone grain",
23
+ "9": "aleurone grain membrane",
24
+ "10": "amyloplast",
25
+ "11": "amyloplast membrane",
26
+ "12": "anammoxosome",
27
+ "13": "apical cell membrane",
28
+ "14": "apicolateral cell membrane",
29
+ "15": "apicoplast",
30
+ "16": "apoplast",
31
+ "17": "archaeal flagellum",
32
+ "18": "ascus epiplasm",
33
+ "19": "attachment organelle",
34
+ "20": "attachment organelle membrane",
35
+ "21": "autolysosome membrane",
36
+ "22": "autophagosome",
37
+ "23": "autophagosome membrane",
38
+ "24": "axon",
39
+ "25": "bacterial flagellum",
40
+ "26": "bacterial flagellum basal body",
41
+ "27": "bacterial microcompartment",
42
+ "28": "basal cell membrane",
43
+ "29": "basement membrane",
44
+ "30": "basolateral cell membrane",
45
+ "31": "bud",
46
+ "32": "bud membrane",
47
+ "33": "bud neck",
48
+ "34": "bud tip",
49
+ "35": "cajal body",
50
+ "36": "calyx",
51
+ "37": "capsule",
52
+ "38": "carboxysome",
53
+ "39": "caveola",
54
+ "40": "cell cortex",
55
+ "41": "cell envelope",
56
+ "42": "cell inner membrane",
57
+ "43": "cell junction",
58
+ "44": "cell membrane",
59
+ "45": "cell outer membrane",
60
+ "46": "cell projection",
61
+ "47": "cell septum",
62
+ "48": "cell surface",
63
+ "49": "cell tip",
64
+ "50": "cellular chromatophore membrane",
65
+ "51": "cellular thylakoid lumen",
66
+ "52": "cellular thylakoid membrane",
67
+ "53": "cell wall",
68
+ "54": "centriolar satellite",
69
+ "55": "centriole",
70
+ "56": "centromere",
71
+ "57": "centrosome",
72
+ "58": "chloroplast",
73
+ "59": "chloroplast envelope",
74
+ "60": "chloroplast inner membrane",
75
+ "61": "chloroplast intermembrane space",
76
+ "62": "chloroplast membrane",
77
+ "63": "chloroplast nucleoid",
78
+ "64": "chloroplast outer membrane",
79
+ "65": "chloroplast stroma",
80
+ "66": "chloroplast thylakoid",
81
+ "67": "chloroplast thylakoid lumen",
82
+ "68": "chloroplast thylakoid membrane",
83
+ "69": "chlorosome",
84
+ "70": "chlorosome envelope",
85
+ "71": "chromaffin granule",
86
+ "72": "chromaffin granule lumen",
87
+ "73": "chromaffin granule membrane",
88
+ "74": "chromoplast",
89
+ "75": "chromoplast stroma",
90
+ "76": "chromosome",
91
+ "77": "cilium",
92
+ "78": "cilium axoneme",
93
+ "79": "cilium basal body",
94
+ "80": "cilium membrane",
95
+ "81": "cis-golgi network",
96
+ "82": "cis-golgi network membrane",
97
+ "83": "clathrin-coated pit",
98
+ "84": "clathrin-coated vesicle",
99
+ "85": "clathrin-coated vesicle membrane",
100
+ "86": "cleavage furrow",
101
+ "87": "coated pit",
102
+ "88": "contractile vacuole",
103
+ "89": "contractile vacuole membrane",
104
+ "90": "copi-coated vesicle",
105
+ "91": "copi-coated vesicle membrane",
106
+ "92": "copii-coated vesicle",
107
+ "93": "copii-coated vesicle membrane",
108
+ "94": "cortical granule",
109
+ "95": "cvt vesicle membrane",
110
+ "96": "cyanelle",
111
+ "97": "cyanelle stroma",
112
+ "98": "cyanelle thylakoid lumen",
113
+ "99": "cyanelle thylakoid membrane",
114
+ "100": "cytolytic granule",
115
+ "101": "cytoplasm",
116
+ "102": "cytoplasmic granule",
117
+ "103": "cytoplasmic granule lumen",
118
+ "104": "cytoplasmic granule membrane",
119
+ "105": "cytoplasmic ribonucleoprotein granule",
120
+ "106": "cytoplasmic vesicle",
121
+ "107": "cytoplasmic vesicle lumen",
122
+ "108": "cytoplasmic vesicle membrane",
123
+ "109": "cytoskeleton",
124
+ "110": "cytosol",
125
+ "111": "dendrite",
126
+ "112": "dendritic spine",
127
+ "113": "dendritic spine membrane",
128
+ "114": "desmosome",
129
+ "115": "dynein axonemal particle",
130
+ "116": "early endosome",
131
+ "117": "early endosome membrane",
132
+ "118": "encapsulin nanocompartment",
133
+ "119": "endomembrane system",
134
+ "120": "endoplasmic reticulum",
135
+ "121": "endoplasmic reticulum-golgi intermediate compartment",
136
+ "122": "endoplasmic reticulum-golgi intermediate compartment membrane",
137
+ "123": "endoplasmic reticulum lumen",
138
+ "124": "endoplasmic reticulum membrane",
139
+ "125": "endosome",
140
+ "126": "endosome membrane",
141
+ "127": "etioplast membrane",
142
+ "128": "extracellular exosome",
143
+ "129": "extracellular matrix",
144
+ "130": "extracellular space",
145
+ "131": "extracellular vesicle",
146
+ "132": "extracellular vesicle membrane",
147
+ "133": "filopodium",
148
+ "134": "filopodium membrane",
149
+ "135": "fimbrium",
150
+ "136": "flagellar pocket",
151
+ "137": "flagellum",
152
+ "138": "flagellum axoneme",
153
+ "139": "flagellum basal body",
154
+ "140": "flagellum membrane",
155
+ "141": "focal adhesion",
156
+ "142": "forespore",
157
+ "143": "forespore inner membrane",
158
+ "144": "forespore intermembrane space",
159
+ "145": "forespore membrane",
160
+ "146": "forespore outer membrane",
161
+ "147": "gap junction",
162
+ "148": "gas vesicle",
163
+ "149": "gas vesicle shell",
164
+ "150": "gem",
165
+ "151": "glycosome",
166
+ "152": "glyoxysome",
167
+ "153": "glyoxysome membrane",
168
+ "154": "golgi apparatus",
169
+ "155": "golgi apparatus lumen",
170
+ "156": "golgi apparatus membrane",
171
+ "157": "golgi outpost",
172
+ "158": "golgi stack",
173
+ "159": "golgi stack membrane",
174
+ "160": "growth cone",
175
+ "161": "hemidesmosome",
176
+ "162": "host cell inner membrane",
177
+ "163": "host cell junction",
178
+ "164": "host cell membrane",
179
+ "165": "host cell outer membrane",
180
+ "166": "host cell surface",
181
+ "167": "host cellular thylakoid membrane",
182
+ "168": "host cell wall",
183
+ "169": "host chloroplast envelope",
184
+ "170": "host cytoplasm",
185
+ "171": "host cytoplasmic vesicle",
186
+ "172": "host cytoplasmic vesicle membrane",
187
+ "173": "host cytoskeleton",
188
+ "174": "host cytosol",
189
+ "175": "host endoplasmic reticulum",
190
+ "176": "host endoplasmic reticulum-golgi intermediate compartment",
191
+ "177": "host endoplasmic reticulum lumen",
192
+ "178": "host endoplasmic reticulum membrane",
193
+ "179": "host endosome",
194
+ "180": "host golgi apparatus",
195
+ "181": "host golgi apparatus membrane",
196
+ "182": "host membrane",
197
+ "183": "host mitochondrion",
198
+ "184": "host mitochondrion inner membrane",
199
+ "185": "host mitochondrion membrane",
200
+ "186": "host mitochondrion outer membrane",
201
+ "187": "host multivesicular body",
202
+ "188": "host nucleolus",
203
+ "189": "host nucleoplasm",
204
+ "190": "host nucleus",
205
+ "191": "host nucleus envelope",
206
+ "192": "host nucleus inner membrane",
207
+ "193": "host nucleus matrix",
208
+ "194": "host nucleus membrane",
209
+ "195": "host perinuclear region",
210
+ "196": "host periplasm",
211
+ "197": "host plasmodesma",
212
+ "198": "host rough endoplasmic reticulum",
213
+ "199": "host rough endoplasmic reticulum membrane",
214
+ "200": "hydrogenosome",
215
+ "201": "hydrogenosome lumen",
216
+ "202": "i band",
217
+ "203": "inflammasome",
218
+ "204": "inner membrane complex",
219
+ "205": "interphotoreceptor matrix",
220
+ "206": "invadopodium membrane",
221
+ "207": "kinetochore",
222
+ "208": "kinetoplast",
223
+ "209": "lamellipodium",
224
+ "210": "lamellipodium membrane",
225
+ "211": "late endosome",
226
+ "212": "late endosome lumen",
227
+ "213": "late endosome membrane",
228
+ "214": "lateral cell membrane",
229
+ "215": "lipid droplet",
230
+ "216": "lysosome",
231
+ "217": "lysosome lumen",
232
+ "218": "lysosome membrane",
233
+ "219": "lytic vacuole",
234
+ "220": "magnetosome",
235
+ "221": "magnetosome membrane",
236
+ "222": "melanosome",
237
+ "223": "melanosome lumen",
238
+ "224": "melanosome membrane",
239
+ "225": "membrane",
240
+ "226": "membrane raft",
241
+ "227": "microneme",
242
+ "228": "microneme membrane",
243
+ "229": "microsome",
244
+ "230": "microsome membrane",
245
+ "231": "microtubule organizing center",
246
+ "232": "microvillus",
247
+ "233": "microvillus membrane",
248
+ "234": "midbody",
249
+ "235": "midbody ring",
250
+ "236": "mitochondrion",
251
+ "237": "mitochondrion envelope",
252
+ "238": "mitochondrion inner membrane",
253
+ "239": "mitochondrion intermembrane space",
254
+ "240": "mitochondrion matrix",
255
+ "241": "mitochondrion membrane",
256
+ "242": "mitochondrion nucleoid",
257
+ "243": "mitochondrion outer membrane",
258
+ "244": "mitosome",
259
+ "245": "mitosome matrix",
260
+ "246": "mitosome membrane",
261
+ "247": "m line",
262
+ "248": "multivesicular body",
263
+ "249": "multivesicular body membrane",
264
+ "250": "myelin membrane",
265
+ "251": "myofibril",
266
+ "252": "nematocyst",
267
+ "253": "neuronal dense core vesicle membrane",
268
+ "254": "neuron projection",
269
+ "255": "nuclear body",
270
+ "256": "nuclear pore complex",
271
+ "257": "nucleoid",
272
+ "258": "nucleolus",
273
+ "259": "nucleoplasm",
274
+ "260": "nucleus",
275
+ "261": "nucleus envelope",
276
+ "262": "nucleus inner membrane",
277
+ "263": "nucleus lamina",
278
+ "264": "nucleus matrix",
279
+ "265": "nucleus membrane",
280
+ "266": "nucleus outer membrane",
281
+ "267": "nucleus speckle",
282
+ "268": "organellar chromatophore",
283
+ "269": "organellar chromatophore thylakoid membrane",
284
+ "270": "parasitophorous vacuole",
285
+ "271": "parasitophorous vacuole lumen",
286
+ "272": "parasitophorous vacuole membrane",
287
+ "273": "p-body",
288
+ "274": "peribacteroid membrane",
289
+ "275": "peribacteroid space",
290
+ "276": "perikaryon",
291
+ "277": "perinuclear region",
292
+ "278": "perinuclear theca",
293
+ "279": "periplasm",
294
+ "280": "periplasmic flagellum",
295
+ "281": "perispore",
296
+ "282": "peroxisome",
297
+ "283": "peroxisome matrix",
298
+ "284": "peroxisome membrane",
299
+ "285": "phagocytic cup",
300
+ "286": "phagosome",
301
+ "287": "phagosome membrane",
302
+ "288": "photoreceptor inner segment",
303
+ "289": "photoreceptor inner segment membrane",
304
+ "290": "photoreceptor outer segment",
305
+ "291": "photoreceptor outer segment membrane",
306
+ "292": "phragmoplast",
307
+ "293": "plasmodesma",
308
+ "294": "plastid",
309
+ "295": "plastid inner membrane",
310
+ "296": "plastid membrane",
311
+ "297": "plastid outer membrane",
312
+ "298": "plastid stroma",
313
+ "299": "plastid thylakoid membrane",
314
+ "300": "plastoglobule",
315
+ "301": "pml body",
316
+ "302": "podosome",
317
+ "303": "pollen coat",
318
+ "304": "postsynapse",
319
+ "305": "postsynaptic cell membrane",
320
+ "306": "postsynaptic density",
321
+ "307": "postsynaptic density membrane",
322
+ "308": "postsynaptic recycling endosome membrane",
323
+ "309": "preautophagosomal structure",
324
+ "310": "preautophagosomal structure membrane",
325
+ "311": "presynapse",
326
+ "312": "presynaptic active zone",
327
+ "313": "presynaptic cell membrane",
328
+ "314": "prevacuolar compartment",
329
+ "315": "prevacuolar compartment membrane",
330
+ "316": "primary cell wall",
331
+ "317": "prospore",
332
+ "318": "prospore membrane",
333
+ "319": "protein storage vacuole",
334
+ "320": "protein storage vacuole lumen",
335
+ "321": "protein storage vacuole membrane",
336
+ "322": "pseudopodium",
337
+ "323": "recycling endosome",
338
+ "324": "recycling endosome membrane",
339
+ "325": "rhoptry membrane",
340
+ "326": "rough endoplasmic reticulum",
341
+ "327": "rough endoplasmic reticulum lumen",
342
+ "328": "rough endoplasmic reticulum membrane",
343
+ "329": "ruffle",
344
+ "330": "ruffle membrane",
345
+ "331": "sarcolemma",
346
+ "332": "sarcomere",
347
+ "333": "sarcoplasmic reticulum",
348
+ "334": "sarcoplasmic reticulum lumen",
349
+ "335": "sarcoplasmic reticulum membrane",
350
+ "336": "secreted",
351
+ "337": "secretory vesicle",
352
+ "338": "secretory vesicle lumen",
353
+ "339": "secretory vesicle membrane",
354
+ "340": "septate junction",
355
+ "341": "s-layer",
356
+ "342": "smooth endoplasmic reticulum membrane",
357
+ "343": "spindle",
358
+ "344": "spindle pole",
359
+ "345": "spindle pole body",
360
+ "346": "spore",
361
+ "347": "spore coat",
362
+ "348": "spore core",
363
+ "349": "spore cortex",
364
+ "350": "spore membrane",
365
+ "351": "spore outer membrane",
366
+ "352": "spore polar tube",
367
+ "353": "spore wall",
368
+ "354": "stereocilium",
369
+ "355": "stereocilium membrane",
370
+ "356": "stress fiber",
371
+ "357": "stress granule",
372
+ "358": "surface film",
373
+ "359": "symbiosome",
374
+ "360": "synapse",
375
+ "361": "synaptic cell membrane",
376
+ "362": "synaptic cleft",
377
+ "363": "synaptic vesicle",
378
+ "364": "synaptic vesicle membrane",
379
+ "365": "synaptosome",
380
+ "366": "target cell",
381
+ "367": "target cell cytoplasm",
382
+ "368": "target cell membrane",
383
+ "369": "tegument membrane",
384
+ "370": "telomere",
385
+ "371": "thylakoid",
386
+ "372": "tight junction",
387
+ "373": "trans-golgi network",
388
+ "374": "trans-golgi network membrane",
389
+ "375": "trichocyst",
390
+ "376": "t-tubule",
391
+ "377": "vacuole",
392
+ "378": "vacuole lumen",
393
+ "379": "vacuole membrane",
394
+ "380": "vesicle",
395
+ "381": "virion",
396
+ "382": "virion membrane",
397
+ "383": "virion tegument",
398
+ "384": "z line",
399
+ "385": "zona pellucida",
400
+ "386": "zymogen granule",
401
+ "387": "zymogen granule membrane"
402
+ },
403
+ "initializer_range": 0.02,
404
+ "intermediate_size": 2560,
405
+ "is_folding_model": false,
406
+ "label2id": {
407
+ "a band": 0,
408
+ "acidocalcisome membrane": 1,
409
+ "acrosome": 2,
410
+ "acrosome inner membrane": 3,
411
+ "acrosome lumen": 4,
412
+ "acrosome membrane": 5,
413
+ "actin patch": 6,
414
+ "adherens junction": 7,
415
+ "aleurone grain": 8,
416
+ "aleurone grain membrane": 9,
417
+ "amyloplast": 10,
418
+ "amyloplast membrane": 11,
419
+ "anammoxosome": 12,
420
+ "apical cell membrane": 13,
421
+ "apicolateral cell membrane": 14,
422
+ "apicoplast": 15,
423
+ "apoplast": 16,
424
+ "archaeal flagellum": 17,
425
+ "ascus epiplasm": 18,
426
+ "attachment organelle": 19,
427
+ "attachment organelle membrane": 20,
428
+ "autolysosome membrane": 21,
429
+ "autophagosome": 22,
430
+ "autophagosome membrane": 23,
431
+ "axon": 24,
432
+ "bacterial flagellum": 25,
433
+ "bacterial flagellum basal body": 26,
434
+ "bacterial microcompartment": 27,
435
+ "basal cell membrane": 28,
436
+ "basement membrane": 29,
437
+ "basolateral cell membrane": 30,
438
+ "bud": 31,
439
+ "bud membrane": 32,
440
+ "bud neck": 33,
441
+ "bud tip": 34,
442
+ "cajal body": 35,
443
+ "calyx": 36,
444
+ "capsule": 37,
445
+ "carboxysome": 38,
446
+ "caveola": 39,
447
+ "cell cortex": 40,
448
+ "cell envelope": 41,
449
+ "cell inner membrane": 42,
450
+ "cell junction": 43,
451
+ "cell membrane": 44,
452
+ "cell outer membrane": 45,
453
+ "cell projection": 46,
454
+ "cell septum": 47,
455
+ "cell surface": 48,
456
+ "cell tip": 49,
457
+ "cell wall": 53,
458
+ "cellular chromatophore membrane": 50,
459
+ "cellular thylakoid lumen": 51,
460
+ "cellular thylakoid membrane": 52,
461
+ "centriolar satellite": 54,
462
+ "centriole": 55,
463
+ "centromere": 56,
464
+ "centrosome": 57,
465
+ "chloroplast": 58,
466
+ "chloroplast envelope": 59,
467
+ "chloroplast inner membrane": 60,
468
+ "chloroplast intermembrane space": 61,
469
+ "chloroplast membrane": 62,
470
+ "chloroplast nucleoid": 63,
471
+ "chloroplast outer membrane": 64,
472
+ "chloroplast stroma": 65,
473
+ "chloroplast thylakoid": 66,
474
+ "chloroplast thylakoid lumen": 67,
475
+ "chloroplast thylakoid membrane": 68,
476
+ "chlorosome": 69,
477
+ "chlorosome envelope": 70,
478
+ "chromaffin granule": 71,
479
+ "chromaffin granule lumen": 72,
480
+ "chromaffin granule membrane": 73,
481
+ "chromoplast": 74,
482
+ "chromoplast stroma": 75,
483
+ "chromosome": 76,
484
+ "cilium": 77,
485
+ "cilium axoneme": 78,
486
+ "cilium basal body": 79,
487
+ "cilium membrane": 80,
488
+ "cis-golgi network": 81,
489
+ "cis-golgi network membrane": 82,
490
+ "clathrin-coated pit": 83,
491
+ "clathrin-coated vesicle": 84,
492
+ "clathrin-coated vesicle membrane": 85,
493
+ "cleavage furrow": 86,
494
+ "coated pit": 87,
495
+ "contractile vacuole": 88,
496
+ "contractile vacuole membrane": 89,
497
+ "copi-coated vesicle": 90,
498
+ "copi-coated vesicle membrane": 91,
499
+ "copii-coated vesicle": 92,
500
+ "copii-coated vesicle membrane": 93,
501
+ "cortical granule": 94,
502
+ "cvt vesicle membrane": 95,
503
+ "cyanelle": 96,
504
+ "cyanelle stroma": 97,
505
+ "cyanelle thylakoid lumen": 98,
506
+ "cyanelle thylakoid membrane": 99,
507
+ "cytolytic granule": 100,
508
+ "cytoplasm": 101,
509
+ "cytoplasmic granule": 102,
510
+ "cytoplasmic granule lumen": 103,
511
+ "cytoplasmic granule membrane": 104,
512
+ "cytoplasmic ribonucleoprotein granule": 105,
513
+ "cytoplasmic vesicle": 106,
514
+ "cytoplasmic vesicle lumen": 107,
515
+ "cytoplasmic vesicle membrane": 108,
516
+ "cytoskeleton": 109,
517
+ "cytosol": 110,
518
+ "dendrite": 111,
519
+ "dendritic spine": 112,
520
+ "dendritic spine membrane": 113,
521
+ "desmosome": 114,
522
+ "dynein axonemal particle": 115,
523
+ "early endosome": 116,
524
+ "early endosome membrane": 117,
525
+ "encapsulin nanocompartment": 118,
526
+ "endomembrane system": 119,
527
+ "endoplasmic reticulum": 120,
528
+ "endoplasmic reticulum lumen": 123,
529
+ "endoplasmic reticulum membrane": 124,
530
+ "endoplasmic reticulum-golgi intermediate compartment": 121,
531
+ "endoplasmic reticulum-golgi intermediate compartment membrane": 122,
532
+ "endosome": 125,
533
+ "endosome membrane": 126,
534
+ "etioplast membrane": 127,
535
+ "extracellular exosome": 128,
536
+ "extracellular matrix": 129,
537
+ "extracellular space": 130,
538
+ "extracellular vesicle": 131,
539
+ "extracellular vesicle membrane": 132,
540
+ "filopodium": 133,
541
+ "filopodium membrane": 134,
542
+ "fimbrium": 135,
543
+ "flagellar pocket": 136,
544
+ "flagellum": 137,
545
+ "flagellum axoneme": 138,
546
+ "flagellum basal body": 139,
547
+ "flagellum membrane": 140,
548
+ "focal adhesion": 141,
549
+ "forespore": 142,
550
+ "forespore inner membrane": 143,
551
+ "forespore intermembrane space": 144,
552
+ "forespore membrane": 145,
553
+ "forespore outer membrane": 146,
554
+ "gap junction": 147,
555
+ "gas vesicle": 148,
556
+ "gas vesicle shell": 149,
557
+ "gem": 150,
558
+ "glycosome": 151,
559
+ "glyoxysome": 152,
560
+ "glyoxysome membrane": 153,
561
+ "golgi apparatus": 154,
562
+ "golgi apparatus lumen": 155,
563
+ "golgi apparatus membrane": 156,
564
+ "golgi outpost": 157,
565
+ "golgi stack": 158,
566
+ "golgi stack membrane": 159,
567
+ "growth cone": 160,
568
+ "hemidesmosome": 161,
569
+ "host cell inner membrane": 162,
570
+ "host cell junction": 163,
571
+ "host cell membrane": 164,
572
+ "host cell outer membrane": 165,
573
+ "host cell surface": 166,
574
+ "host cell wall": 168,
575
+ "host cellular thylakoid membrane": 167,
576
+ "host chloroplast envelope": 169,
577
+ "host cytoplasm": 170,
578
+ "host cytoplasmic vesicle": 171,
579
+ "host cytoplasmic vesicle membrane": 172,
580
+ "host cytoskeleton": 173,
581
+ "host cytosol": 174,
582
+ "host endoplasmic reticulum": 175,
583
+ "host endoplasmic reticulum lumen": 177,
584
+ "host endoplasmic reticulum membrane": 178,
585
+ "host endoplasmic reticulum-golgi intermediate compartment": 176,
586
+ "host endosome": 179,
587
+ "host golgi apparatus": 180,
588
+ "host golgi apparatus membrane": 181,
589
+ "host membrane": 182,
590
+ "host mitochondrion": 183,
591
+ "host mitochondrion inner membrane": 184,
592
+ "host mitochondrion membrane": 185,
593
+ "host mitochondrion outer membrane": 186,
594
+ "host multivesicular body": 187,
595
+ "host nucleolus": 188,
596
+ "host nucleoplasm": 189,
597
+ "host nucleus": 190,
598
+ "host nucleus envelope": 191,
599
+ "host nucleus inner membrane": 192,
600
+ "host nucleus matrix": 193,
601
+ "host nucleus membrane": 194,
602
+ "host perinuclear region": 195,
603
+ "host periplasm": 196,
604
+ "host plasmodesma": 197,
605
+ "host rough endoplasmic reticulum": 198,
606
+ "host rough endoplasmic reticulum membrane": 199,
607
+ "hydrogenosome": 200,
608
+ "hydrogenosome lumen": 201,
609
+ "i band": 202,
610
+ "inflammasome": 203,
611
+ "inner membrane complex": 204,
612
+ "interphotoreceptor matrix": 205,
613
+ "invadopodium membrane": 206,
614
+ "kinetochore": 207,
615
+ "kinetoplast": 208,
616
+ "lamellipodium": 209,
617
+ "lamellipodium membrane": 210,
618
+ "late endosome": 211,
619
+ "late endosome lumen": 212,
620
+ "late endosome membrane": 213,
621
+ "lateral cell membrane": 214,
622
+ "lipid droplet": 215,
623
+ "lysosome": 216,
624
+ "lysosome lumen": 217,
625
+ "lysosome membrane": 218,
626
+ "lytic vacuole": 219,
627
+ "m line": 247,
628
+ "magnetosome": 220,
629
+ "magnetosome membrane": 221,
630
+ "melanosome": 222,
631
+ "melanosome lumen": 223,
632
+ "melanosome membrane": 224,
633
+ "membrane": 225,
634
+ "membrane raft": 226,
635
+ "microneme": 227,
636
+ "microneme membrane": 228,
637
+ "microsome": 229,
638
+ "microsome membrane": 230,
639
+ "microtubule organizing center": 231,
640
+ "microvillus": 232,
641
+ "microvillus membrane": 233,
642
+ "midbody": 234,
643
+ "midbody ring": 235,
644
+ "mitochondrion": 236,
645
+ "mitochondrion envelope": 237,
646
+ "mitochondrion inner membrane": 238,
647
+ "mitochondrion intermembrane space": 239,
648
+ "mitochondrion matrix": 240,
649
+ "mitochondrion membrane": 241,
650
+ "mitochondrion nucleoid": 242,
651
+ "mitochondrion outer membrane": 243,
652
+ "mitosome": 244,
653
+ "mitosome matrix": 245,
654
+ "mitosome membrane": 246,
655
+ "multivesicular body": 248,
656
+ "multivesicular body membrane": 249,
657
+ "myelin membrane": 250,
658
+ "myofibril": 251,
659
+ "nematocyst": 252,
660
+ "neuron projection": 254,
661
+ "neuronal dense core vesicle membrane": 253,
662
+ "nuclear body": 255,
663
+ "nuclear pore complex": 256,
664
+ "nucleoid": 257,
665
+ "nucleolus": 258,
666
+ "nucleoplasm": 259,
667
+ "nucleus": 260,
668
+ "nucleus envelope": 261,
669
+ "nucleus inner membrane": 262,
670
+ "nucleus lamina": 263,
671
+ "nucleus matrix": 264,
672
+ "nucleus membrane": 265,
673
+ "nucleus outer membrane": 266,
674
+ "nucleus speckle": 267,
675
+ "organellar chromatophore": 268,
676
+ "organellar chromatophore thylakoid membrane": 269,
677
+ "p-body": 273,
678
+ "parasitophorous vacuole": 270,
679
+ "parasitophorous vacuole lumen": 271,
680
+ "parasitophorous vacuole membrane": 272,
681
+ "peribacteroid membrane": 274,
682
+ "peribacteroid space": 275,
683
+ "perikaryon": 276,
684
+ "perinuclear region": 277,
685
+ "perinuclear theca": 278,
686
+ "periplasm": 279,
687
+ "periplasmic flagellum": 280,
688
+ "perispore": 281,
689
+ "peroxisome": 282,
690
+ "peroxisome matrix": 283,
691
+ "peroxisome membrane": 284,
692
+ "phagocytic cup": 285,
693
+ "phagosome": 286,
694
+ "phagosome membrane": 287,
695
+ "photoreceptor inner segment": 288,
696
+ "photoreceptor inner segment membrane": 289,
697
+ "photoreceptor outer segment": 290,
698
+ "photoreceptor outer segment membrane": 291,
699
+ "phragmoplast": 292,
700
+ "plasmodesma": 293,
701
+ "plastid": 294,
702
+ "plastid inner membrane": 295,
703
+ "plastid membrane": 296,
704
+ "plastid outer membrane": 297,
705
+ "plastid stroma": 298,
706
+ "plastid thylakoid membrane": 299,
707
+ "plastoglobule": 300,
708
+ "pml body": 301,
709
+ "podosome": 302,
710
+ "pollen coat": 303,
711
+ "postsynapse": 304,
712
+ "postsynaptic cell membrane": 305,
713
+ "postsynaptic density": 306,
714
+ "postsynaptic density membrane": 307,
715
+ "postsynaptic recycling endosome membrane": 308,
716
+ "preautophagosomal structure": 309,
717
+ "preautophagosomal structure membrane": 310,
718
+ "presynapse": 311,
719
+ "presynaptic active zone": 312,
720
+ "presynaptic cell membrane": 313,
721
+ "prevacuolar compartment": 314,
722
+ "prevacuolar compartment membrane": 315,
723
+ "primary cell wall": 316,
724
+ "prospore": 317,
725
+ "prospore membrane": 318,
726
+ "protein storage vacuole": 319,
727
+ "protein storage vacuole lumen": 320,
728
+ "protein storage vacuole membrane": 321,
729
+ "pseudopodium": 322,
730
+ "recycling endosome": 323,
731
+ "recycling endosome membrane": 324,
732
+ "rhoptry membrane": 325,
733
+ "rough endoplasmic reticulum": 326,
734
+ "rough endoplasmic reticulum lumen": 327,
735
+ "rough endoplasmic reticulum membrane": 328,
736
+ "ruffle": 329,
737
+ "ruffle membrane": 330,
738
+ "s-layer": 341,
739
+ "sarcolemma": 331,
740
+ "sarcomere": 332,
741
+ "sarcoplasmic reticulum": 333,
742
+ "sarcoplasmic reticulum lumen": 334,
743
+ "sarcoplasmic reticulum membrane": 335,
744
+ "secreted": 336,
745
+ "secretory vesicle": 337,
746
+ "secretory vesicle lumen": 338,
747
+ "secretory vesicle membrane": 339,
748
+ "septate junction": 340,
749
+ "smooth endoplasmic reticulum membrane": 342,
750
+ "spindle": 343,
751
+ "spindle pole": 344,
752
+ "spindle pole body": 345,
753
+ "spore": 346,
754
+ "spore coat": 347,
755
+ "spore core": 348,
756
+ "spore cortex": 349,
757
+ "spore membrane": 350,
758
+ "spore outer membrane": 351,
759
+ "spore polar tube": 352,
760
+ "spore wall": 353,
761
+ "stereocilium": 354,
762
+ "stereocilium membrane": 355,
763
+ "stress fiber": 356,
764
+ "stress granule": 357,
765
+ "surface film": 358,
766
+ "symbiosome": 359,
767
+ "synapse": 360,
768
+ "synaptic cell membrane": 361,
769
+ "synaptic cleft": 362,
770
+ "synaptic vesicle": 363,
771
+ "synaptic vesicle membrane": 364,
772
+ "synaptosome": 365,
773
+ "t-tubule": 376,
774
+ "target cell": 366,
775
+ "target cell cytoplasm": 367,
776
+ "target cell membrane": 368,
777
+ "tegument membrane": 369,
778
+ "telomere": 370,
779
+ "thylakoid": 371,
780
+ "tight junction": 372,
781
+ "trans-golgi network": 373,
782
+ "trans-golgi network membrane": 374,
783
+ "trichocyst": 375,
784
+ "vacuole": 377,
785
+ "vacuole lumen": 378,
786
+ "vacuole membrane": 379,
787
+ "vesicle": 380,
788
+ "virion": 381,
789
+ "virion membrane": 382,
790
+ "virion tegument": 383,
791
+ "z line": 384,
792
+ "zona pellucida": 385,
793
+ "zymogen granule": 386,
794
+ "zymogen granule membrane": 387
795
+ },
796
+ "layer_norm_eps": 1e-05,
797
+ "mask_token_id": 32,
798
+ "max_position_embeddings": 1026,
799
+ "model_type": "esm",
800
+ "num_attention_heads": 20,
801
+ "num_hidden_layers": 30,
802
+ "pad_token_id": 1,
803
+ "position_embedding_type": "rotary",
804
+ "problem_type": "multi_label_classification",
805
+ "token_dropout": true,
806
+ "torch_dtype": "float32",
807
+ "transformers_version": "4.43.3",
808
+ "use_cache": true,
809
+ "vocab_list": null,
810
+ "vocab_size": 33
811
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8bd2710b9aad03a31df8f552cdaaf2cd3112bd51fc3a6f17be8464b1f843aaf0
3
+ size 596238916
special_tokens_map.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": {
3
+ "content": "<cls>",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "eos_token": {
10
+ "content": "<eos>",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "mask_token": {
17
+ "content": "<mask>",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "pad_token": {
24
+ "content": "<pad>",
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
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "<cls>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "1": {
12
+ "content": "<pad>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "2": {
20
+ "content": "<eos>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "3": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "32": {
36
+ "content": "<mask>",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "additional_special_tokens": [],
45
+ "clean_up_tokenization_spaces": true,
46
+ "cls_token": "<cls>",
47
+ "eos_token": "<eos>",
48
+ "mask_token": "<mask>",
49
+ "model_max_length": 1000000000000000019884624838656,
50
+ "pad_token": "<pad>",
51
+ "tokenizer_class": "EsmTokenizer",
52
+ "unk_token": "<unk>"
53
+ }
vocab.txt ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <cls>
2
+ <pad>
3
+ <eos>
4
+ <unk>
5
+ L
6
+ A
7
+ G
8
+ V
9
+ S
10
+ E
11
+ R
12
+ T
13
+ I
14
+ D
15
+ P
16
+ K
17
+ Q
18
+ N
19
+ F
20
+ Y
21
+ M
22
+ H
23
+ W
24
+ C
25
+ X
26
+ B
27
+ U
28
+ Z
29
+ O
30
+ .
31
+ -
32
+ <null_1>
33
+ <mask>