zeroshot commited on
Commit
c53df02
1 Parent(s): e3cb7b3

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +544 -0
README.md CHANGED
@@ -1,4 +1,548 @@
1
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2
  license: mit
3
  ---
4
  This is the sparse ONNX variant of the [bge-small-en-v1.5](https://huggingface.co/BAAI/bge-small-en-v1.5) model for embeddings created with [DeepSparse Optimum](https://github.com/neuralmagic/optimum-deepsparse) for ONNX export and Neural Magic's [Sparsify](https://account.neuralmagic.com/signin?client_id=d04a5f0c-983d-11ed-88a6-971073f187d3&return_to=https%3A//accounts.neuralmagic.com/v1/connect/authorize%3Fscope%3Dsparsify%3Aread%2Bsparsify%3Awrite%2Buser%3Aapi-key%3Aread%2Buser%3Aprofile%3Awrite%2Buser%3Aprofile%3Aread%26response_type%3Dcode%26code_challenge_method%3DS256%26redirect_uri%3Dhttps%3A//apps.neuralmagic.com/sparsify/oidc/callback.html%26state%3Da9b466a6193c4a7b92cba469408d2495%26client_id%3Dd04a5f0c-983d-11ed-88a6-971073f187d3%26code_challenge%3DP0EkmKBpplTb7crJOGS8YLSwT8UH-BeuD0wuE4JTORQ%26response_mode%3Dquery) for One-Shot quantization and unstructured pruning (50%).
 
1
  ---
2
+ tags:
3
+ - mteb
4
+ model-index:
5
+ - name: bge-small-en-v1.5-sparse
6
+ results:
7
+ - task:
8
+ type: Classification
9
+ dataset:
10
+ type: mteb/amazon_counterfactual
11
+ name: MTEB AmazonCounterfactualClassification (en)
12
+ config: en
13
+ split: test
14
+ revision: e8379541af4e31359cca9fbcf4b00f2671dba205
15
+ metrics:
16
+ - type: accuracy
17
+ value: 70.71641791044776
18
+ - type: ap
19
+ value: 32.850850647310004
20
+ - type: f1
21
+ value: 64.48101916414805
22
+ - task:
23
+ type: Classification
24
+ dataset:
25
+ type: mteb/amazon_polarity
26
+ name: MTEB AmazonPolarityClassification
27
+ config: default
28
+ split: test
29
+ revision: e2d317d38cd51312af73b3d32a06d1a08b442046
30
+ metrics:
31
+ - type: accuracy
32
+ value: 83.33962500000001
33
+ - type: ap
34
+ value: 78.28706349240106
35
+ - type: f1
36
+ value: 83.27426715603062
37
+ - task:
38
+ type: Classification
39
+ dataset:
40
+ type: mteb/amazon_reviews_multi
41
+ name: MTEB AmazonReviewsClassification (en)
42
+ config: en
43
+ split: test
44
+ revision: 1399c76144fd37290681b995c656ef9b2e06e26d
45
+ metrics:
46
+ - type: accuracy
47
+ value: 40.988
48
+ - type: f1
49
+ value: 40.776679545648506
50
+ - task:
51
+ type: STS
52
+ dataset:
53
+ type: mteb/biosses-sts
54
+ name: MTEB BIOSSES
55
+ config: default
56
+ split: test
57
+ revision: d3fb88f8f02e40887cd149695127462bbcf29b4a
58
+ metrics:
59
+ - type: cos_sim_pearson
60
+ value: 79.64892774481326
61
+ - type: cos_sim_spearman
62
+ value: 78.953003817029
63
+ - type: euclidean_pearson
64
+ value: 78.92456838230683
65
+ - type: euclidean_spearman
66
+ value: 78.56504316985354
67
+ - type: manhattan_pearson
68
+ value: 79.21436359014227
69
+ - type: manhattan_spearman
70
+ value: 78.66263575501259
71
+ - task:
72
+ type: Classification
73
+ dataset:
74
+ type: mteb/banking77
75
+ name: MTEB Banking77Classification
76
+ config: default
77
+ split: test
78
+ revision: 0fd18e25b25c072e09e0d92ab615fda904d66300
79
+ metrics:
80
+ - type: accuracy
81
+ value: 81.25
82
+ - type: f1
83
+ value: 81.20841448916138
84
+ - task:
85
+ type: Classification
86
+ dataset:
87
+ type: mteb/emotion
88
+ name: MTEB EmotionClassification
89
+ config: default
90
+ split: test
91
+ revision: 4f58c6b202a23cf9a4da393831edf4f9183cad37
92
+ metrics:
93
+ - type: accuracy
94
+ value: 41.665
95
+ - type: f1
96
+ value: 37.601137843331244
97
+ - task:
98
+ type: Classification
99
+ dataset:
100
+ type: mteb/imdb
101
+ name: MTEB ImdbClassification
102
+ config: default
103
+ split: test
104
+ revision: 3d86128a09e091d6018b6d26cad27f2739fc2db7
105
+ metrics:
106
+ - type: accuracy
107
+ value: 74.8052
108
+ - type: ap
109
+ value: 68.92588517572685
110
+ - type: f1
111
+ value: 74.66801685854456
112
+ - task:
113
+ type: Classification
114
+ dataset:
115
+ type: mteb/mtop_domain
116
+ name: MTEB MTOPDomainClassification (en)
117
+ config: en
118
+ split: test
119
+ revision: d80d48c1eb48d3562165c59d59d0034df9fff0bf
120
+ metrics:
121
+ - type: accuracy
122
+ value: 91.2220702234382
123
+ - type: f1
124
+ value: 90.81687856852439
125
+ - task:
126
+ type: Classification
127
+ dataset:
128
+ type: mteb/mtop_intent
129
+ name: MTEB MTOPIntentClassification (en)
130
+ config: en
131
+ split: test
132
+ revision: ae001d0e6b1228650b7bd1c2c65fb50ad11a8aba
133
+ metrics:
134
+ - type: accuracy
135
+ value: 69.39124487004105
136
+ - type: f1
137
+ value: 51.8350043424968
138
+ - task:
139
+ type: Classification
140
+ dataset:
141
+ type: mteb/amazon_massive_intent
142
+ name: MTEB MassiveIntentClassification (en)
143
+ config: en
144
+ split: test
145
+ revision: 31efe3c427b0bae9c22cbb560b8f15491cc6bed7
146
+ metrics:
147
+ - type: accuracy
148
+ value: 69.80497646267652
149
+ - type: f1
150
+ value: 67.34213899244814
151
+ - task:
152
+ type: Classification
153
+ dataset:
154
+ type: mteb/amazon_massive_scenario
155
+ name: MTEB MassiveScenarioClassification (en)
156
+ config: en
157
+ split: test
158
+ revision: 7d571f92784cd94a019292a1f45445077d0ef634
159
+ metrics:
160
+ - type: accuracy
161
+ value: 74.54270342972428
162
+ - type: f1
163
+ value: 74.02802500235784
164
+ - task:
165
+ type: STS
166
+ dataset:
167
+ type: mteb/sickr-sts
168
+ name: MTEB SICK-R
169
+ config: default
170
+ split: test
171
+ revision: a6ea5a8cab320b040a23452cc28066d9beae2cee
172
+ metrics:
173
+ - type: cos_sim_pearson
174
+ value: 80.89215288990194
175
+ - type: cos_sim_spearman
176
+ value: 74.386413188675
177
+ - type: euclidean_pearson
178
+ value: 78.83679563989534
179
+ - type: euclidean_spearman
180
+ value: 74.29328198771996
181
+ - type: manhattan_pearson
182
+ value: 78.77968796707641
183
+ - type: manhattan_spearman
184
+ value: 74.20887429784696
185
+ - task:
186
+ type: STS
187
+ dataset:
188
+ type: mteb/sts12-sts
189
+ name: MTEB STS12
190
+ config: default
191
+ split: test
192
+ revision: a0d554a64d88156834ff5ae9920b964011b16384
193
+ metrics:
194
+ - type: cos_sim_pearson
195
+ value: 78.31858821914498
196
+ - type: cos_sim_spearman
197
+ value: 72.2217008523832
198
+ - type: euclidean_pearson
199
+ value: 75.38901061978429
200
+ - type: euclidean_spearman
201
+ value: 71.81255767675184
202
+ - type: manhattan_pearson
203
+ value: 75.49472202181288
204
+ - type: manhattan_spearman
205
+ value: 71.96322588726144
206
+ - task:
207
+ type: STS
208
+ dataset:
209
+ type: mteb/sts13-sts
210
+ name: MTEB STS13
211
+ config: default
212
+ split: test
213
+ revision: 7e90230a92c190f1bf69ae9002b8cea547a64cca
214
+ metrics:
215
+ - type: cos_sim_pearson
216
+ value: 79.48334648997455
217
+ - type: cos_sim_spearman
218
+ value: 80.99654029572798
219
+ - type: euclidean_pearson
220
+ value: 80.46546523970035
221
+ - type: euclidean_spearman
222
+ value: 80.90646216980744
223
+ - type: manhattan_pearson
224
+ value: 80.35474057857608
225
+ - type: manhattan_spearman
226
+ value: 80.8141299909659
227
+ - task:
228
+ type: STS
229
+ dataset:
230
+ type: mteb/sts14-sts
231
+ name: MTEB STS14
232
+ config: default
233
+ split: test
234
+ revision: 6031580fec1f6af667f0bd2da0a551cf4f0b2375
235
+ metrics:
236
+ - type: cos_sim_pearson
237
+ value: 79.73826970784727
238
+ - type: cos_sim_spearman
239
+ value: 76.9926870133034
240
+ - type: euclidean_pearson
241
+ value: 79.6386542120984
242
+ - type: euclidean_spearman
243
+ value: 77.05041986942253
244
+ - type: manhattan_pearson
245
+ value: 79.61799508502459
246
+ - type: manhattan_spearman
247
+ value: 77.07169617647067
248
+ - task:
249
+ type: STS
250
+ dataset:
251
+ type: mteb/sts15-sts
252
+ name: MTEB STS15
253
+ config: default
254
+ split: test
255
+ revision: ae752c7c21bf194d8b67fd573edf7ae58183cbe3
256
+ metrics:
257
+ - type: cos_sim_pearson
258
+ value: 83.93999019426069
259
+ - type: cos_sim_spearman
260
+ value: 85.21166521594695
261
+ - type: euclidean_pearson
262
+ value: 84.97207676326357
263
+ - type: euclidean_spearman
264
+ value: 85.40726578482739
265
+ - type: manhattan_pearson
266
+ value: 85.0386693192183
267
+ - type: manhattan_spearman
268
+ value: 85.49230945586409
269
+ - task:
270
+ type: STS
271
+ dataset:
272
+ type: mteb/sts16-sts
273
+ name: MTEB STS16
274
+ config: default
275
+ split: test
276
+ revision: 4d8694f8f0e0100860b497b999b3dbed754a0513
277
+ metrics:
278
+ - type: cos_sim_pearson
279
+ value: 80.8133974034008
280
+ - type: cos_sim_spearman
281
+ value: 82.82919022688844
282
+ - type: euclidean_pearson
283
+ value: 81.92587923760179
284
+ - type: euclidean_spearman
285
+ value: 82.86629450518863
286
+ - type: manhattan_pearson
287
+ value: 81.98232365999253
288
+ - type: manhattan_spearman
289
+ value: 82.94313939920296
290
+ - task:
291
+ type: STS
292
+ dataset:
293
+ type: mteb/sts17-crosslingual-sts
294
+ name: MTEB STS17 (en-en)
295
+ config: en-en
296
+ split: test
297
+ revision: af5e6fb845001ecf41f4c1e033ce921939a2a68d
298
+ metrics:
299
+ - type: cos_sim_pearson
300
+ value: 86.12872422642363
301
+ - type: cos_sim_spearman
302
+ value: 87.77672179979807
303
+ - type: euclidean_pearson
304
+ value: 87.76172961705947
305
+ - type: euclidean_spearman
306
+ value: 87.9891393339215
307
+ - type: manhattan_pearson
308
+ value: 87.78863663568221
309
+ - type: manhattan_spearman
310
+ value: 88.08297053203866
311
+ - task:
312
+ type: STS
313
+ dataset:
314
+ type: mteb/sts22-crosslingual-sts
315
+ name: MTEB STS22 (en)
316
+ config: en
317
+ split: test
318
+ revision: 6d1ba47164174a496b7fa5d3569dae26a6813b80
319
+ metrics:
320
+ - type: cos_sim_pearson
321
+ value: 58.82824030232733
322
+ - type: cos_sim_spearman
323
+ value: 64.17079382633538
324
+ - type: euclidean_pearson
325
+ value: 61.31505225602925
326
+ - type: euclidean_spearman
327
+ value: 64.05080034530694
328
+ - type: manhattan_pearson
329
+ value: 61.77095758943306
330
+ - type: manhattan_spearman
331
+ value: 64.14475973774933
332
+ - task:
333
+ type: STS
334
+ dataset:
335
+ type: mteb/stsbenchmark-sts
336
+ name: MTEB STSBenchmark
337
+ config: default
338
+ split: test
339
+ revision: b0fddb56ed78048fa8b90373c8a3cfc37b684831
340
+ metrics:
341
+ - type: cos_sim_pearson
342
+ value: 81.39239803497064
343
+ - type: cos_sim_spearman
344
+ value: 81.76637354520439
345
+ - type: euclidean_pearson
346
+ value: 82.98008209033587
347
+ - type: euclidean_spearman
348
+ value: 82.18662536188657
349
+ - type: manhattan_pearson
350
+ value: 82.9630328314908
351
+ - type: manhattan_spearman
352
+ value: 82.13726553603003
353
+ - task:
354
+ type: PairClassification
355
+ dataset:
356
+ type: mteb/sprintduplicatequestions-pairclassification
357
+ name: MTEB SprintDuplicateQuestions
358
+ config: default
359
+ split: test
360
+ revision: d66bd1f72af766a5cc4b0ca5e00c162f89e8cc46
361
+ metrics:
362
+ - type: cos_sim_accuracy
363
+ value: 99.8019801980198
364
+ - type: cos_sim_ap
365
+ value: 94.58629018512772
366
+ - type: cos_sim_f1
367
+ value: 89.84771573604061
368
+ - type: cos_sim_precision
369
+ value: 91.23711340206185
370
+ - type: cos_sim_recall
371
+ value: 88.5
372
+ - type: dot_accuracy
373
+ value: 99.74950495049505
374
+ - type: dot_ap
375
+ value: 92.5761214576951
376
+ - type: dot_f1
377
+ value: 87.09841917389087
378
+ - type: dot_precision
379
+ value: 88.86576482830385
380
+ - type: dot_recall
381
+ value: 85.39999999999999
382
+ - type: euclidean_accuracy
383
+ value: 99.80495049504951
384
+ - type: euclidean_ap
385
+ value: 94.56231673602272
386
+ - type: euclidean_f1
387
+ value: 90.02531645569621
388
+ - type: euclidean_precision
389
+ value: 91.17948717948718
390
+ - type: euclidean_recall
391
+ value: 88.9
392
+ - type: manhattan_accuracy
393
+ value: 99.8009900990099
394
+ - type: manhattan_ap
395
+ value: 94.5775591647447
396
+ - type: manhattan_f1
397
+ value: 89.86384266263238
398
+ - type: manhattan_precision
399
+ value: 90.64089521871821
400
+ - type: manhattan_recall
401
+ value: 89.1
402
+ - type: max_accuracy
403
+ value: 99.80495049504951
404
+ - type: max_ap
405
+ value: 94.58629018512772
406
+ - type: max_f1
407
+ value: 90.02531645569621
408
+ - task:
409
+ type: Classification
410
+ dataset:
411
+ type: mteb/toxic_conversations_50k
412
+ name: MTEB ToxicConversationsClassification
413
+ config: default
414
+ split: test
415
+ revision: d7c0de2777da35d6aae2200a62c6e0e5af397c4c
416
+ metrics:
417
+ - type: accuracy
418
+ value: 67.43820000000001
419
+ - type: ap
420
+ value: 12.899489312331003
421
+ - type: f1
422
+ value: 52.03468121072981
423
+ - task:
424
+ type: Classification
425
+ dataset:
426
+ type: mteb/tweet_sentiment_extraction
427
+ name: MTEB TweetSentimentExtractionClassification
428
+ config: default
429
+ split: test
430
+ revision: d604517c81ca91fe16a244d1248fc021f9ecee7a
431
+ metrics:
432
+ - type: accuracy
433
+ value: 57.475947934352
434
+ - type: f1
435
+ value: 57.77676730676238
436
+ - task:
437
+ type: PairClassification
438
+ dataset:
439
+ type: mteb/twittersemeval2015-pairclassification
440
+ name: MTEB TwitterSemEval2015
441
+ config: default
442
+ split: test
443
+ revision: 70970daeab8776df92f5ea462b6173c0b46fd2d1
444
+ metrics:
445
+ - type: cos_sim_accuracy
446
+ value: 83.94230196101806
447
+ - type: cos_sim_ap
448
+ value: 67.00916556336148
449
+ - type: cos_sim_f1
450
+ value: 63.046014257939085
451
+ - type: cos_sim_precision
452
+ value: 61.961783439490446
453
+ - type: cos_sim_recall
454
+ value: 64.16886543535621
455
+ - type: dot_accuracy
456
+ value: 83.18531322644095
457
+ - type: dot_ap
458
+ value: 63.112896030267066
459
+ - type: dot_f1
460
+ value: 59.06565656565657
461
+ - type: dot_precision
462
+ value: 56.63438256658596
463
+ - type: dot_recall
464
+ value: 61.715039577836414
465
+ - type: euclidean_accuracy
466
+ value: 83.94230196101806
467
+ - type: euclidean_ap
468
+ value: 67.19856676674463
469
+ - type: euclidean_f1
470
+ value: 63.08428413691571
471
+ - type: euclidean_precision
472
+ value: 58.9543682641596
473
+ - type: euclidean_recall
474
+ value: 67.83641160949868
475
+ - type: manhattan_accuracy
476
+ value: 83.91845979614949
477
+ - type: manhattan_ap
478
+ value: 66.9845327263072
479
+ - type: manhattan_f1
480
+ value: 62.693323274236135
481
+ - type: manhattan_precision
482
+ value: 59.884698534710544
483
+ - type: manhattan_recall
484
+ value: 65.77836411609499
485
+ - type: max_accuracy
486
+ value: 83.94230196101806
487
+ - type: max_ap
488
+ value: 67.19856676674463
489
+ - type: max_f1
490
+ value: 63.08428413691571
491
+ - task:
492
+ type: PairClassification
493
+ dataset:
494
+ type: mteb/twitterurlcorpus-pairclassification
495
+ name: MTEB TwitterURLCorpus
496
+ config: default
497
+ split: test
498
+ revision: 8b6510b0b1fa4e4c4f879467980e9be563ec1cdf
499
+ metrics:
500
+ - type: cos_sim_accuracy
501
+ value: 88.0777738968448
502
+ - type: cos_sim_ap
503
+ value: 84.19747786536
504
+ - type: cos_sim_f1
505
+ value: 75.91830995817077
506
+ - type: cos_sim_precision
507
+ value: 69.84671107949033
508
+ - type: cos_sim_recall
509
+ value: 83.14598090545118
510
+ - type: dot_accuracy
511
+ value: 87.14246904955951
512
+ - type: dot_ap
513
+ value: 82.37528804640529
514
+ - type: dot_f1
515
+ value: 74.40963166732163
516
+ - type: dot_precision
517
+ value: 69.4127841098447
518
+ - type: dot_recall
519
+ value: 80.18170619032954
520
+ - type: euclidean_accuracy
521
+ value: 88.08359529630924
522
+ - type: euclidean_ap
523
+ value: 84.22633217661986
524
+ - type: euclidean_f1
525
+ value: 76.09190339866403
526
+ - type: euclidean_precision
527
+ value: 72.70304390517605
528
+ - type: euclidean_recall
529
+ value: 79.81213427779488
530
+ - type: manhattan_accuracy
531
+ value: 88.08359529630924
532
+ - type: manhattan_ap
533
+ value: 84.18362004611083
534
+ - type: manhattan_f1
535
+ value: 76.08789625360231
536
+ - type: manhattan_precision
537
+ value: 71.49336582724072
538
+ - type: manhattan_recall
539
+ value: 81.3135201724669
540
+ - type: max_accuracy
541
+ value: 88.08359529630924
542
+ - type: max_ap
543
+ value: 84.22633217661986
544
+ - type: max_f1
545
+ value: 76.09190339866403
546
  license: mit
547
  ---
548
  This is the sparse ONNX variant of the [bge-small-en-v1.5](https://huggingface.co/BAAI/bge-small-en-v1.5) model for embeddings created with [DeepSparse Optimum](https://github.com/neuralmagic/optimum-deepsparse) for ONNX export and Neural Magic's [Sparsify](https://account.neuralmagic.com/signin?client_id=d04a5f0c-983d-11ed-88a6-971073f187d3&return_to=https%3A//accounts.neuralmagic.com/v1/connect/authorize%3Fscope%3Dsparsify%3Aread%2Bsparsify%3Awrite%2Buser%3Aapi-key%3Aread%2Buser%3Aprofile%3Awrite%2Buser%3Aprofile%3Aread%26response_type%3Dcode%26code_challenge_method%3DS256%26redirect_uri%3Dhttps%3A//apps.neuralmagic.com/sparsify/oidc/callback.html%26state%3Da9b466a6193c4a7b92cba469408d2495%26client_id%3Dd04a5f0c-983d-11ed-88a6-971073f187d3%26code_challenge%3DP0EkmKBpplTb7crJOGS8YLSwT8UH-BeuD0wuE4JTORQ%26response_mode%3Dquery) for One-Shot quantization and unstructured pruning (50%).