sileod commited on
Commit
6132e97
·
verified ·
1 Parent(s): 17c1f43

Add new SentenceTransformer model

Browse files
Files changed (2) hide show
  1. README.md +160 -128
  2. model.safetensors +1 -1
README.md CHANGED
@@ -7,7 +7,7 @@ tags:
7
  - feature-extraction
8
  - dense
9
  - generated_from_trainer
10
- - dataset_size:7176192
11
  - loss:AnglELoss
12
  - loss:CoSENTLoss
13
  - loss:CachedMultipleNegativesRankingLoss
@@ -37,43 +37,46 @@ widget:
37
  \ pediatrician, or paediatrician. The word pediatrics and its cognates mean healer\
38
  \ of children; they derive from two Greek words: Ï\x80αá¿\x96Ï\x82 (pais child)\
39
  \ and ἰαÏ\x84Ï\x81Ï\x8CÏ\x82 (iatros doctor, healer)."
40
- - source_sentence: For example , Elizabeth Coffin , daughter of a wealthy merchant
41
- from Nantucket , was mother of the prominent Massachusetts industrialists Henry
42
- Coffin Nevins and David Nevins , Jr..
43
  sentences:
44
- - Born in the middle of the Great Depression , Carl Carl Spitz was trained and owned
45
- by Terry .
46
- - For example , Elizabeth Coffin , a daughter of a wealthy merchant from Nantucket
47
- , was the mother of the prominent Massachusetts industrialists Henry Coffin Nevins
48
- and David Nevins , Jr ...
49
- - The couple had their first child , Shalk Jr , in August 2012 , and in March 2014
50
- his second son Nicol was born .
51
- - source_sentence: UN Chief Finds His Voice, Remains Cautious on China
52
  sentences:
53
- - 'Insight: U.N. chief finds his voice, but remains cautious on China'
54
- - kashmir is claimed by both india and pakistan.
55
- - Death toll in Kenya bus attack rises to six
56
- - source_sentence: Mayor Michael R. Bloomberg said yesterday that the men's behavior
57
- "was a disgrace, and totally inappropriate for city employees."
58
  sentences:
59
- - SOCIAL ECONOMICAL RESEARCH STUDY
60
- - I assume you arrive from you apartment in Old Town, Selina Kyle.
61
- - The way the men acted "was a disgrace, and totally inappropriate for city employees"
62
- according to Michael R. Bloomberg's comments yesterday
63
- - source_sentence: what are the three subatomic particles called?
64
  sentences:
65
- - Subatomic particles include electrons, the negatively charged, almost massless
66
- particles that nevertheless account for most of the size of the atom, and they
67
- include the heavier building blocks of the small but very dense nucleus of the
68
- atom, the positively charged protons and the electrically neutral neutrons.
69
- - Your body needs cholesterol to build healthy cells, but high levels of cholesterol
70
- can increase your risk of heart disease. With high cholesterol, you can develop
71
- fatty deposits in your blood vessels. Eventually, these deposits grow, making
72
- it difficult for enough blood to flow through your arteries.
73
- - 'If you experience any of the following symptoms, stop taking ibuprofen and call
74
- your doctor: stomach pain, heartburn, vomit that is bloody or looks like coffee
75
- grounds, blood in the stool, or black and tarry stools. Keep all appointments
76
- with your doctor and the laboratory.'
 
 
 
 
77
  datasets:
78
  - google-research-datasets/paws
79
  - nyu-mll/glue
@@ -88,7 +91,7 @@ library_name: sentence-transformers
88
 
89
  # SentenceTransformer based on jhu-clsp/ettin-encoder-17m
90
 
91
- This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [jhu-clsp/ettin-encoder-17m](https://huggingface.co/jhu-clsp/ettin-encoder-17m) on 19 datasets. It maps sentences & paragraphs to a 256-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
92
 
93
  ## Model Details
94
 
@@ -118,6 +121,7 @@ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [j
118
  - sentence-transformers/s2orc
119
  - sentence-transformers/codesearchnet
120
  - sentence-transformers/stackexchange-duplicates
 
121
  - **Language:** en
122
  <!-- - **License:** Unknown -->
123
 
@@ -155,12 +159,12 @@ from sentence_transformers import SentenceTransformer
155
  model = SentenceTransformer("tasksource/ettin-17m-embed")
156
  # Run inference
157
  queries = [
158
- "what are the three subatomic particles called?",
159
  ]
160
  documents = [
161
- 'Subatomic particles include electrons, the negatively charged, almost massless particles that nevertheless account for most of the size of the atom, and they include the heavier building blocks of the small but very dense nucleus of the atom, the positively charged protons and the electrically neutral neutrons.',
162
- 'Your body needs cholesterol to build healthy cells, but high levels of cholesterol can increase your risk of heart disease. With high cholesterol, you can develop fatty deposits in your blood vessels. Eventually, these deposits grow, making it difficult for enough blood to flow through your arteries.',
163
- 'If you experience any of the following symptoms, stop taking ibuprofen and call your doctor: stomach pain, heartburn, vomit that is bloody or looks like coffee grounds, blood in the stool, or black and tarry stools. Keep all appointments with your doctor and the laboratory.',
164
  ]
165
  query_embeddings = model.encode_query(queries)
166
  document_embeddings = model.encode_document(documents)
@@ -170,7 +174,7 @@ print(query_embeddings.shape, document_embeddings.shape)
170
  # Get the similarity scores for the embeddings
171
  similarities = model.similarity(query_embeddings, document_embeddings)
172
  print(similarities)
173
- # tensor([[ 0.7121, -0.0953, 0.0591]])
174
  ```
175
 
176
  <!--
@@ -223,13 +227,13 @@ You can finetune this model on your own dataset.
223
  | | sentence1 | sentence2 | label |
224
  |:--------|:-----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:------------------------------------------------|
225
  | type | string | string | int |
226
- | details | <ul><li>min: 11 tokens</li><li>mean: 27.69 tokens</li><li>max: 57 tokens</li></ul> | <ul><li>min: 11 tokens</li><li>mean: 27.57 tokens</li><li>max: 56 tokens</li></ul> | <ul><li>0: ~53.10%</li><li>1: ~46.90%</li></ul> |
227
  * Samples:
228
- | sentence1 | sentence2 | label |
229
- |:-------------------------------------------------------------------------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
230
- | <code>He currently lives in Ridgefield , Connecticut , with his wife Cynthia . He has four children : Lindsay , Heather , Jason , and Rebecca .</code> | <code>He lives in Ridgefield , Connecticut with his wife Cynthia , and has four children : Jason , Heather , Lindsay and Rebecca .</code> | <code>1</code> |
231
- | <code>It is located close to Mandurriao at 113 R. Mapa Street in Old Iloilo Airport 's Iloilo City district .</code> | <code>It is located close to the Old Iloilo Airport at 113 R. Mapa Street in the Mandurriao district of Iloilo City .</code> | <code>0</code> |
232
- | <code>Alaric informs Caroline that Stefan ( Paul Wesley ) stopped looking for a way to bring back Damon and Bonnie .</code> | <code>Stefan informs Caroline that Alaric ( Paul Wesley ) has stopped looking for a way to get Damon and Bonnie back .</code> | <code>0</code> |
233
  * Loss: [<code>AnglELoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#angleloss) with these parameters:
234
  ```json
235
  {
@@ -249,13 +253,13 @@ You can finetune this model on your own dataset.
249
  | | sentence1 | sentence2 | label |
250
  |:--------|:-----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:------------------------------------------------|
251
  | type | string | string | int |
252
- | details | <ul><li>min: 10 tokens</li><li>mean: 27.86 tokens</li><li>max: 45 tokens</li></ul> | <ul><li>min: 12 tokens</li><li>mean: 28.03 tokens</li><li>max: 46 tokens</li></ul> | <ul><li>0: ~30.20%</li><li>1: ~69.80%</li></ul> |
253
  * Samples:
254
- | sentence1 | sentence2 | label |
255
- |:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
256
- | <code>The transaction will expand the revenue from Callebaut 's consumer products business to 45 percent of total sales from 23 percent .</code> | <code>The transaction will expand Callebaut 's sales revenues from its consumer products business to 45 percent from 23 percent .</code> | <code>1</code> |
257
- | <code>AutoAdvice is available as a one-year subscription at $ 400 per CPU , scaling from one to 50,000 CPUs .</code> | <code>SAN Architect will run approximately $ 2,400 while AutoAdvice is available with coverage from one to 50,000 CPUs .</code> | <code>0</code> |
258
- | <code>" The flu season has had an earlier onset than we 've seen in many years , " said Julie Gerberding , director of the Centers for Disease Control and Prevention in Atlanta .</code> | <code>We 're seeing some very high levels of widespread flu infections in some places , " said Dr. Julie Gerberding , director of the federal Centers for Disease Control and Prevention .</code> | <code>0</code> |
259
  * Loss: [<code>AnglELoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#angleloss) with these parameters:
260
  ```json
261
  {
@@ -275,13 +279,13 @@ You can finetune this model on your own dataset.
275
  | | sentence1 | sentence2 | label |
276
  |:--------|:----------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------|:------------------------------------------------|
277
  | type | string | string | int |
278
- | details | <ul><li>min: 6 tokens</li><li>mean: 13.61 tokens</li><li>max: 66 tokens</li></ul> | <ul><li>min: 12 tokens</li><li>mean: 317.54 tokens</li><li>max: 1024 tokens</li></ul> | <ul><li>0: ~30.00%</li><li>1: ~70.00%</li></ul> |
279
  * Samples:
280
- | sentence1 | sentence2 | label |
281
- |:--------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
282
- | <code>Lauren Bacall worked in Applause.</code> | <code>Dr. V. N. Tiwari -LRB- 1936 -- 1984 -RRB- was an Indian author and parliamentarian .. He wrote books in Punjabi , English and Hindi .. He was nominated as member of the Rajya Sabha in 1982 and served till 1984 .. Rajya Sabha. Rajya Sabha</code> | <code>1</code> |
283
- | <code>Jon Snow does not negotiate an alliance between the Night's Watch and the wildlings.</code> | <code>Court Square Place is a 16 story office building located at Hunter and 25th Streets , near Court Square , Long Island City , Queens just outside Manhattan and is `` the tallest building constructed in the area since 1989 . ''. office. office. Long Island City. Long Island City. Queens. Queens, New York. Manhattan. Manhattan. It is owned and operated by the United Nations Federal Credit Union -LRB- UNFCU -RRB- .. United Nations Federal Credit Union. United Nations Federal Credit Union. It was completed in 2006 based on the designs from architects of HLW International with Tishman Construction Corporation handling the major construction .. Tishman Construction Corporation. Tishman International Companies. Langan Engineering and Environmental Services , Inc. , Eurotech Construction , Permasteelisa USA and Cives Steel Company also provided construction support .. The New York Times states that the building `` rivals new construction in Manhattan . ''. Manhattan. Manhattan. New York Times. ...</code> | <code>1</code> |
284
- | <code>Zeus is the son of Titans.</code> | <code>The Brantford Smoke were a professional ice hockey team in the Colonial Hockey League , later the United Hockey League , now known as the International Hockey League .. Brantford. Brantford. United Hockey League. United Hockey League. ice hockey. ice hockey. Colonial Hockey League. Colonial Hockey League. International Hockey League. International Hockey League ( 2007- ). They played in Brantford , Ontario from 1991-92 -LRB- the league 's inaugural season -RRB- until 1997-98 , playing home games at the Brantford Civic Centre .. Brantford. Brantford. Ontario. Ontario. Brantford Civic Centre. Brantford Civic Centre. They won the Colonial Cup in 1993 , beating the St. Thomas Wildcats in a series that included a bench brawl in a game in St. Thomas .. St. Thomas Wildcats. St. Thomas Wildcats. In 1998 they moved to Asheville , North Carolina and became the Asheville Smoke .. Asheville Smoke. Asheville Smoke. Category : Defunct ice hockey teams in Canada. ice hockey. ice hockey. Category : ...</code> | <code>1</code> |
285
  * Loss: [<code>AnglELoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#angleloss) with these parameters:
286
  ```json
287
  {
@@ -298,16 +302,16 @@ You can finetune this model on your own dataset.
298
  * Size: 22,650 training samples
299
  * Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
300
  * Approximate statistics based on the first 1000 samples:
301
- | | sentence1 | sentence2 | label |
302
- |:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:------------------------------------------------|
303
- | type | string | string | int |
304
- | details | <ul><li>min: 6 tokens</li><li>mean: 22.51 tokens</li><li>max: 61 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 21.94 tokens</li><li>max: 49 tokens</li></ul> | <ul><li>0: ~53.30%</li><li>1: ~46.70%</li></ul> |
305
  * Samples:
306
- | sentence1 | sentence2 | label |
307
- |:--------------------------------------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
308
- | <code>it solves subset of problem and then use that info to solve the more difficult original process</code> | <code>a coding approach that aims to eliminate repeated calculations basic idea: store results that might need to be repeatedly calculated (memoisation)</code> | <code>0</code> |
309
- | <code>cryptographic algorithms that provide message integrity by producing a condensed representation of a message called a message digest</code> | <code>mathematical algorithms that generate a message summary or digest in order to confirm message identity and integrity</code> | <code>1</code> |
310
- | <code>interaction among people in virtual communities where they can share information and ideas facilitates b2b and c2c e-commerce</code> | <code>refers to the enabling technologies for social interaction among people in which allows for the creation and exchange of user-generated-content.</code> | <code>0</code> |
311
  * Loss: [<code>AnglELoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#angleloss) with these parameters:
312
  ```json
313
  {
@@ -324,16 +328,16 @@ You can finetune this model on your own dataset.
324
  * Size: 10,047 training samples
325
  * Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
326
  * Approximate statistics based on the first 1000 samples:
327
- | | sentence1 | sentence2 | label |
328
- |:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:------------------------------------------------|
329
- | type | string | string | int |
330
- | details | <ul><li>min: 4 tokens</li><li>mean: 17.33 tokens</li><li>max: 59 tokens</li></ul> | <ul><li>min: 3 tokens</li><li>mean: 17.23 tokens</li><li>max: 70 tokens</li></ul> | <ul><li>0: ~34.30%</li><li>1: ~65.70%</li></ul> |
331
  * Samples:
332
- | sentence1 | sentence2 | label |
333
- |:------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------|
334
- | <code>That provision now reads as follows:</code> | <code>Following the new stipulation that reads:</code> | <code>1</code> |
335
- | <code>We intend to appeal vigorously and still expect to be vindicated ultimately.</code> | <code>We want to appeal seriously, but expect it to be proven in the end.</code> | <code>1</code> |
336
- | <code>Gilberto, you're the fucking man.</code> | <code>amazing</code> | <code>0</code> |
337
  * Loss: [<code>AnglELoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#angleloss) with these parameters:
338
  ```json
339
  {
@@ -353,13 +357,13 @@ You can finetune this model on your own dataset.
353
  | | sentence1 | sentence2 | label |
354
  |:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------|
355
  | type | string | string | float |
356
- | details | <ul><li>min: 6 tokens</li><li>mean: 14.68 tokens</li><li>max: 45 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 14.79 tokens</li><li>max: 68 tokens</li></ul> | <ul><li>min: 0.0</li><li>mean: 2.76</li><li>max: 5.0</li></ul> |
357
  * Samples:
358
- | sentence1 | sentence2 | label |
359
- |:--------------------------------------------------------------------------|:------------------------------------------------------------|:-------------------------------|
360
- | <code>Updated - Two explosions near finish line of Boston Marathon</code> | <code>Two explosions at Boston Marathon finish line</code> | <code>5.0</code> |
361
- | <code>Pakistan condemns US drone strike in Shawal Area</code> | <code>Pakistan condemns US drone strike in Miranshah</code> | <code>3.200000047683716</code> |
362
- | <code>A woman is frying something in the pan.</code> | <code>A man is playing his guitar.</code> | <code>0.0</code> |
363
  * Loss: [<code>CoSENTLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cosentloss) with these parameters:
364
  ```json
365
  {
@@ -376,16 +380,16 @@ You can finetune this model on your own dataset.
376
  * Size: 13,317 training samples
377
  * Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
378
  * Approximate statistics based on the first 1000 samples:
379
- | | sentence1 | sentence2 | label |
380
- |:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------|
381
- | type | string | string | float |
382
- | details | <ul><li>min: 6 tokens</li><li>mean: 12.21 tokens</li><li>max: 30 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 11.96 tokens</li><li>max: 30 tokens</li></ul> | <ul><li>min: 1.0</li><li>mean: 3.44</li><li>max: 5.0</li></ul> |
383
  * Samples:
384
- | sentence1 | sentence2 | label |
385
- |:-------------------------------------------------------------------------------|:-----------------------------------------------------------------------|:--------------------------------|
386
- | <code>A woman is cutting garlic</code> | <code>A woman is slicing an onion</code> | <code>2.9000000953674316</code> |
387
- | <code>A man is jumping onto a low wall</code> | <code>A man is standing in front of a wall</code> | <code>3.0</code> |
388
- | <code>A small boy in a shirt, which is yellow, is laughing on the beach</code> | <code>A little boy is wearing a yellow tank top and is laughing</code> | <code>4.264999866485596</code> |
389
  * Loss: [<code>CoSENTLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cosentloss) with these parameters:
390
  ```json
391
  {
@@ -402,16 +406,16 @@ You can finetune this model on your own dataset.
402
  * Size: 14,280 training samples
403
  * Columns: <code>label</code>, <code>sentence1</code>, and <code>sentence2</code>
404
  * Approximate statistics based on the first 1000 samples:
405
- | | label | sentence1 | sentence2 |
406
- |:--------|:---------------------------------------------------------------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|
407
- | type | float | string | string |
408
- | details | <ul><li>min: 0.0</li><li>mean: 3.19</li><li>max: 5.0</li></ul> | <ul><li>min: 4 tokens</li><li>mean: 19.21 tokens</li><li>max: 80 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 17.7 tokens</li><li>max: 82 tokens</li></ul> |
409
  * Samples:
410
- | label | sentence1 | sentence2 |
411
- |:------------------|:--------------------------------------------------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------------------------------------------------------|
412
- | <code>2.67</code> | <code>and they definitely shouldn't be forcing anyone to pay for someone else's contraception & abortions.</code> | <code>and there will not be any stain on that money abetting abortions.</code> |
413
- | <code>5.0</code> | <code>someone or something that is the agent of fulfilling desired expectations</code> | <code>someone (or something) on which expectations are centered.</code> |
414
- | <code>4.4</code> | <code>We have heard that the previous Council Presidency first restructured and then scrapped the Ministry for Equality.</code> | <code>We heard that the previous presidency initially restructured then dismantled the ministry for the equal opportunity.</code> |
415
  * Loss: [<code>CoSENTLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cosentloss) with these parameters:
416
  ```json
417
  {
@@ -726,6 +730,34 @@ You can finetune this model on your own dataset.
726
  }
727
  ```
728
  </details>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
729
 
730
  ### Training Hyperparameters
731
  #### Non-Default Hyperparameters
@@ -864,34 +896,34 @@ You can finetune this model on your own dataset.
864
  ### Training Logs
865
  | Epoch | Step | Training Loss |
866
  |:------:|:-----:|:-------------:|
867
- | 0.0357 | 500 | 4.7629 |
868
- | 0.0713 | 1000 | 3.8023 |
869
- | 0.1070 | 1500 | 3.3211 |
870
- | 0.1426 | 2000 | 3.5102 |
871
- | 0.1783 | 2500 | 3.5391 |
872
- | 0.2139 | 3000 | 2.9845 |
873
- | 0.2496 | 3500 | 2.9055 |
874
- | 0.2852 | 4000 | 3.5688 |
875
- | 0.3209 | 4500 | 3.1019 |
876
- | 0.3565 | 5000 | 3.3654 |
877
- | 0.3922 | 5500 | 3.0016 |
878
- | 0.4278 | 6000 | 2.9012 |
879
- | 0.4635 | 6500 | 3.0244 |
880
- | 0.4991 | 7000 | 3.0991 |
881
- | 0.5348 | 7500 | 3.0689 |
882
- | 0.5704 | 8000 | 2.8734 |
883
- | 0.6061 | 8500 | 3.3082 |
884
- | 0.6417 | 9000 | 3.121 |
885
- | 0.6774 | 9500 | 2.7939 |
886
- | 0.7130 | 10000 | 2.7496 |
887
- | 0.7487 | 10500 | 2.9717 |
888
- | 0.7843 | 11000 | 2.8985 |
889
- | 0.8200 | 11500 | 3.044 |
890
- | 0.8556 | 12000 | 3.0341 |
891
- | 0.8913 | 12500 | 2.4321 |
892
- | 0.9269 | 13000 | 2.4815 |
893
- | 0.9626 | 13500 | 2.47 |
894
- | 0.9982 | 14000 | 2.6064 |
895
 
896
 
897
  ### Framework Versions
 
7
  - feature-extraction
8
  - dense
9
  - generated_from_trainer
10
+ - dataset_size:7376192
11
  - loss:AnglELoss
12
  - loss:CoSENTLoss
13
  - loss:CachedMultipleNegativesRankingLoss
 
37
  \ pediatrician, or paediatrician. The word pediatrics and its cognates mean healer\
38
  \ of children; they derive from two Greek words: Ï\x80αá¿\x96Ï\x82 (pais child)\
39
  \ and ἰαÏ\x84Ï\x81Ï\x8CÏ\x82 (iatros doctor, healer)."
40
+ - source_sentence: These TBMs used were then buried to provide an electrical mass
41
+ .
 
42
  sentences:
43
+ - As a result of this success , other multinational companies such as Unilever ,
44
+ Microsoft , Digital Equipment , Schlumberger or Lazard have approached the services
45
+ of Leclercq 's company .
46
+ - These buried TBMs were then used to provide an electrical earth .
47
+ - Huge fields along the zone were discovered in 1920 at Huntington Beach Oil Field
48
+ and at Long Beach Oil Field in 1921 .
49
+ - source_sentence: Eric Gagne earned his 17th save in as many opportunities as he
50
+ struck out three in the ninth and allowed only an infield single by Greg Norton.
51
  sentences:
52
+ - Closer Eric Gagne earned his 17th save in as many opportunities as he struck out
53
+ three of the four batters he faced in the ninth.
54
+ - Syrian soldiers killed in bomb attack
55
+ - Two puppies play with a red chew toy in a field.
56
+ - source_sentence: I think I want to learn French.
57
  sentences:
58
+ - GOOD
59
+ - In my experience, I have observed them not achieve their goals very quickly but
60
+ I have also witnessed them succeed.
61
+ - my way of thinking I want to learn French
62
+ - source_sentence: who said better to reign in hell than serve in heaven
63
  sentences:
64
+ - House of Cards (season 6) The sixth and final season of the American political
65
+ drama web television series House of Cards was confirmed by Netflix on December
66
+ 4, 2017, and is scheduled to be released in late 2018.[1] Unlike previous seasons
67
+ that consisted of thirteen episodes each, the sixth season will consist of only
68
+ eight. The season will not include former lead actor Kevin Spacey, who was fired
69
+ from the show due to sexual misconduct allegations.
70
+ - 'Maggie Elizabeth Jones Maggie Elizabeth Jones (born October 10, 2003) is an American
71
+ child actress, best known for her roles in We Bought a Zoo, the Fox sitcom Ben
72
+ and Kate,[1] and as Lea Clark in American Girl: Lea to the Rescue. [2]'
73
+ - 'Paradise Lost Satan, formerly called Lucifer, is the first major character introduced
74
+ in the poem. He was once the most beautiful of all angels, and is a tragic figure
75
+ who famously declares: "Better to reign in Hell than serve in Heaven." Following
76
+ his failed rebellion against God, he is cast out from Heaven and condemned to
77
+ Hell. Satan''s desire to rebel against his creator stems from his unwillingness
78
+ to be subjugated by God and his Son, claiming that angels are "self-begot, self-raised,"[13]
79
+ and thereby denying God''s authority over them as their creator.'
80
  datasets:
81
  - google-research-datasets/paws
82
  - nyu-mll/glue
 
91
 
92
  # SentenceTransformer based on jhu-clsp/ettin-encoder-17m
93
 
94
+ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [jhu-clsp/ettin-encoder-17m](https://huggingface.co/jhu-clsp/ettin-encoder-17m) on 20 datasets. It maps sentences & paragraphs to a 256-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
95
 
96
  ## Model Details
97
 
 
121
  - sentence-transformers/s2orc
122
  - sentence-transformers/codesearchnet
123
  - sentence-transformers/stackexchange-duplicates
124
+ - tasksource/flan
125
  - **Language:** en
126
  <!-- - **License:** Unknown -->
127
 
 
159
  model = SentenceTransformer("tasksource/ettin-17m-embed")
160
  # Run inference
161
  queries = [
162
+ "who said better to reign in hell than serve in heaven",
163
  ]
164
  documents = [
165
+ 'Paradise Lost Satan, formerly called Lucifer, is the first major character introduced in the poem. He was once the most beautiful of all angels, and is a tragic figure who famously declares: "Better to reign in Hell than serve in Heaven." Following his failed rebellion against God, he is cast out from Heaven and condemned to Hell. Satan\'s desire to rebel against his creator stems from his unwillingness to be subjugated by God and his Son, claiming that angels are "self-begot, self-raised,"[13] and thereby denying God\'s authority over them as their creator.',
166
+ 'Maggie Elizabeth Jones Maggie Elizabeth Jones (born October 10, 2003) is an American child actress, best known for her roles in We Bought a Zoo, the Fox sitcom Ben and Kate,[1] and as Lea Clark in American Girl: Lea to the Rescue. [2]',
167
+ 'House of Cards (season 6) The sixth and final season of the American political drama web television series House of Cards was confirmed by Netflix on December 4, 2017, and is scheduled to be released in late 2018.[1] Unlike previous seasons that consisted of thirteen episodes each, the sixth season will consist of only eight. The season will not include former lead actor Kevin Spacey, who was fired from the show due to sexual misconduct allegations.',
168
  ]
169
  query_embeddings = model.encode_query(queries)
170
  document_embeddings = model.encode_document(documents)
 
174
  # Get the similarity scores for the embeddings
175
  similarities = model.similarity(query_embeddings, document_embeddings)
176
  print(similarities)
177
+ # tensor([[ 0.6853, -0.1319, -0.0476]])
178
  ```
179
 
180
  <!--
 
227
  | | sentence1 | sentence2 | label |
228
  |:--------|:-----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:------------------------------------------------|
229
  | type | string | string | int |
230
+ | details | <ul><li>min: 11 tokens</li><li>mean: 27.89 tokens</li><li>max: 58 tokens</li></ul> | <ul><li>min: 11 tokens</li><li>mean: 27.86 tokens</li><li>max: 57 tokens</li></ul> | <ul><li>0: ~55.30%</li><li>1: ~44.70%</li></ul> |
231
  * Samples:
232
+ | sentence1 | sentence2 | label |
233
+ |:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
234
+ | <code>Wesley Enoch , the eldest son of Doug and Lyn Enoch from Stradbroke Island , grew up in Brisbane and is the brother of the Minister of Queensland , Leeanne Enoch .</code> | <code>Wesley Enoch , the eldest son of Doug and Lyn Enoch from Brisbane , grew up in Stradbroke Island and is the brother of the Minister of Queensland , Leeanne Enoch .</code> | <code>0</code> |
235
+ | <code>The idea of the ensemble is further discussed in the article Statistical Ensemble ( Mathematical Physics ) .</code> | <code>The idea of the ensemble is further discussed in the mathematical ensemble ( statistical physics ) article .</code> | <code>0</code> |
236
+ | <code>Clatsop County comprises the Astoria , OR Micropolitan Statistical Area and is located in the northwest - Oregon .</code> | <code>Clatsop County comprises the Astoria , OR Micropolitan Statistical Area and is located in Northwest Oregon .</code> | <code>1</code> |
237
  * Loss: [<code>AnglELoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#angleloss) with these parameters:
238
  ```json
239
  {
 
253
  | | sentence1 | sentence2 | label |
254
  |:--------|:-----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:------------------------------------------------|
255
  | type | string | string | int |
256
+ | details | <ul><li>min: 10 tokens</li><li>mean: 27.03 tokens</li><li>max: 52 tokens</li></ul> | <ul><li>min: 12 tokens</li><li>mean: 27.13 tokens</li><li>max: 50 tokens</li></ul> | <ul><li>0: ~32.90%</li><li>1: ~67.10%</li></ul> |
257
  * Samples:
258
+ | sentence1 | sentence2 | label |
259
+ |:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
260
+ | <code>Colorado Attorney-General Ken Salazar later said his office has also filed suit against Invesco , charging it with violations of the state 's consumer protection act .</code> | <code>Colorado Attorney General Ken Salazar also filed a lawsuit Tuesday against Invesco , accusing it of violating the state 's Consumer Protection Act .</code> | <code>1</code> |
261
+ | <code>At first , the animals performance declined compared to the sessions on the joystick .</code> | <code>At first , the animals ' performance declined compared with the joystick sessions .</code> | <code>1</code> |
262
+ | <code>For the first quarter , HP pulled in $ 2.94 billion and captured 27.9 percent of the market .</code> | <code>H-P came in second with nearly $ 3.32 billion in sales and 26 percent of the market 's revenue .</code> | <code>1</code> |
263
  * Loss: [<code>AnglELoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#angleloss) with these parameters:
264
  ```json
265
  {
 
279
  | | sentence1 | sentence2 | label |
280
  |:--------|:----------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------|:------------------------------------------------|
281
  | type | string | string | int |
282
+ | details | <ul><li>min: 6 tokens</li><li>mean: 13.48 tokens</li><li>max: 32 tokens</li></ul> | <ul><li>min: 28 tokens</li><li>mean: 331.34 tokens</li><li>max: 1024 tokens</li></ul> | <ul><li>0: ~32.20%</li><li>1: ~67.80%</li></ul> |
283
  * Samples:
284
+ | sentence1 | sentence2 | label |
285
+ |:------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
286
+ | <code>Dopamine is released by neurons.</code> | <code>Dopamine -LRB- DA , contracted from 3,4-dihydroxyphenethylamine -RRB- is an organic chemical of the catecholamine and phenethylamine families that plays several important roles in the brain and body .. organic chemical. organic compound. catecholamine. catecholamine. phenethylamine. phenethylamine. brain. brain. Dopamine. Dopamine ( medication ). It is an amine synthesized by removing a carboxyl group from a molecule of its precursor chemical L-DOPA , which is synthesized in the brain and kidneys .. L-DOPA. L-DOPA. amine. amine. carboxyl group. C-terminus. precursor chemical. precursor ( chemistry ). synthesized. biosynthesis. brain. brain. Dopamine is also synthesized in plants and most animals .. synthesized. biosynthesis. Dopamine. Dopamine ( medication ). In the brain , dopamine functions as a neurotransmitter -- a chemical released by neurons -LRB- nerve cells -RRB- to send signals to other nerve cells .. brain. brain. neurotransmitter. neurotransmitter. The brain includes several...</code> | <code>0</code> |
287
+ | <code>Tanzania shares borders with several landlocked countries.</code> | <code>Mah-Adhur Gushnasp , also known by the Arabicized form of Mahadharjushnas , was an Iranian nobleman who served as the wuzurg framadar -LRB- vizier or prime minister -RRB- of the Sasanian Empire during the reign of the child ruler Ardashir III -LRB- r. 628 -- 629 -RRB- .. Arabicized. Arabicized. Iranian. Iranian people. wuzurg framadar. Wuzurg framadar. vizier. vizier. prime minister. prime minister. Sasanian Empire. Sasanian Empire. Ardashir III. Ardashir III</code> | <code>1</code> |
288
+ | <code>Naga Chaitanya was in Dhada.</code> | <code>BMW 320 TC is a racing car built under Super 2000 specifications , which is currently competing in the FIA World Touring Car Championship .. World Touring Car Championship. World Touring Car Championship. BMW. BMW. Super 2000. Super 2000. FIA. FIA</code> | <code>1</code> |
289
  * Loss: [<code>AnglELoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#angleloss) with these parameters:
290
  ```json
291
  {
 
302
  * Size: 22,650 training samples
303
  * Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
304
  * Approximate statistics based on the first 1000 samples:
305
+ | | sentence1 | sentence2 | label |
306
+ |:--------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:------------------------------------------------|
307
+ | type | string | string | int |
308
+ | details | <ul><li>min: 6 tokens</li><li>mean: 22.13 tokens</li><li>max: 60 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 22.1 tokens</li><li>max: 48 tokens</li></ul> | <ul><li>0: ~53.00%</li><li>1: ~47.00%</li></ul> |
309
  * Samples:
310
+ | sentence1 | sentence2 | label |
311
+ |:----------------------------------------------------------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
312
+ | <code>a statistical statement of how likely it is that an obtained result occurred by chance (p value should be less than or equal to 3)</code> | <code>refers to the cutoff point (i.e., critical value); any value that exceeds the cutoff point will be noted as statistically significant</code> | <code>0</code> |
313
+ | <code>a website that finds webpages that match a word of phase of a given search expression</code> | <code>search tool that allows you to find specific documents through keyword searches and menu choices, in contrast to directories, which are lists of websites classified by topic.</code> | <code>1</code> |
314
+ | <code>explains a phenomenon using assumptions and a philosophical stance. similar to theory but more abstract, generally not testable. connects concepts</code> | <code>broadly presents an understanding of a phenomena and reflects the assumptions of he model's designer</code> | <code>1</code> |
315
  * Loss: [<code>AnglELoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#angleloss) with these parameters:
316
  ```json
317
  {
 
328
  * Size: 10,047 training samples
329
  * Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
330
  * Approximate statistics based on the first 1000 samples:
331
+ | | sentence1 | sentence2 | label |
332
+ |:--------|:----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:------------------------------------------------|
333
+ | type | string | string | int |
334
+ | details | <ul><li>min: 4 tokens</li><li>mean: 17.38 tokens</li><li>max: 82 tokens</li></ul> | <ul><li>min: 3 tokens</li><li>mean: 17.38 tokens</li><li>max: 143 tokens</li></ul> | <ul><li>0: ~36.60%</li><li>1: ~63.40%</li></ul> |
335
  * Samples:
336
+ | sentence1 | sentence2 | label |
337
+ |:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
338
+ | <code>The Embraer jets are scheduled to be delivered by September 2006.</code> | <code>Joseph Biden Jr. is an American politician and the president-elect of the United States.</code> | <code>0</code> |
339
+ | <code>Attackers detonated a second roadside bomb later Sunday as a U.S. convoy was traveling near Fallujah, killing an American soldier and wounding three others.</code> | <code>ATTAACKERS DETONATED A SECOND ROADSIDE BOMB LATER SUNDAY AS A U S CONVOY WAS TRAVELING NEAR FALLUJAH, AND AMERICAN SOLDER WOUNDING OTHERS</code> | <code>0</code> |
340
+ | <code>You want something to eat?</code> | <code>Do you want to go get some food</code> | <code>0</code> |
341
  * Loss: [<code>AnglELoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#angleloss) with these parameters:
342
  ```json
343
  {
 
357
  | | sentence1 | sentence2 | label |
358
  |:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------|
359
  | type | string | string | float |
360
+ | details | <ul><li>min: 6 tokens</li><li>mean: 14.85 tokens</li><li>max: 50 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 14.79 tokens</li><li>max: 51 tokens</li></ul> | <ul><li>min: 0.0</li><li>mean: 2.68</li><li>max: 5.0</li></ul> |
361
  * Samples:
362
+ | sentence1 | sentence2 | label |
363
+ |:-------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------|:-------------------------------|
364
+ | <code>New Syria opposition chief wants no-strings aid</code> | <code>Syria opposition unites as Israel fires warning shots</code> | <code>1.0</code> |
365
+ | <code>A spokesman said: "Since November, we have co-operated fully with the police.</code> | <code>It added it had "co-operated fully" with police since November.</code> | <code>4.25</code> |
366
+ | <code>A small boy in a bathrobe is sitting in a metal chair.</code> | <code>A boy in a robe sits in a chair.</code> | <code>4.199999809265137</code> |
367
  * Loss: [<code>CoSENTLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cosentloss) with these parameters:
368
  ```json
369
  {
 
380
  * Size: 13,317 training samples
381
  * Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
382
  * Approximate statistics based on the first 1000 samples:
383
+ | | sentence1 | sentence2 | label |
384
+ |:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:--------------------------------------------------------------|
385
+ | type | string | string | float |
386
+ | details | <ul><li>min: 5 tokens</li><li>mean: 12.26 tokens</li><li>max: 30 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 12.22 tokens</li><li>max: 30 tokens</li></ul> | <ul><li>min: 1.0</li><li>mean: 3.5</li><li>max: 5.0</li></ul> |
387
  * Samples:
388
+ | sentence1 | sentence2 | label |
389
+ |:----------------------------------------------------|:---------------------------------------------------------|:-------------------------------|
390
+ | <code>A woman is cooking eggs</code> | <code>The woman is cooking something</code> | <code>4.199999809265137</code> |
391
+ | <code>A great dog is climbing a steep hill</code> | <code>A great dog is wildly climbing a steep hill</code> | <code>4.5</code> |
392
+ | <code>A man is passionately playing a guitar</code> | <code>A person is singing and playing a guitar</code> | <code>4.199999809265137</code> |
393
  * Loss: [<code>CoSENTLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cosentloss) with these parameters:
394
  ```json
395
  {
 
406
  * Size: 14,280 training samples
407
  * Columns: <code>label</code>, <code>sentence1</code>, and <code>sentence2</code>
408
  * Approximate statistics based on the first 1000 samples:
409
+ | | label | sentence1 | sentence2 |
410
+ |:--------|:---------------------------------------------------------------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|
411
+ | type | float | string | string |
412
+ | details | <ul><li>min: 0.0</li><li>mean: 3.13</li><li>max: 5.0</li></ul> | <ul><li>min: 4 tokens</li><li>mean: 18.59 tokens</li><li>max: 80 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 17.19 tokens</li><li>max: 77 tokens</li></ul> |
413
  * Samples:
414
+ | label | sentence1 | sentence2 |
415
+ |:------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
416
+ | <code>0.0</code> | <code>they "have" fox because fox news reports on newsworthy events.</code> | <code>i saved up for it because i knew i was going to need it.</code> |
417
+ | <code>4.25</code> | <code>Each time, radical improvements in technology made the threat evaporate.</code> | <code>Each time, of radical technological progress leave these threats.</code> |
418
+ | <code>3.4</code> | <code>I call on Prague to respond to this signal, this challenge, this plea for dialogue, to grant our request and to ensure, together with our House, that this legacy of a nationalist era can be consigned to the past.</code> | <code>I invite Prague to seize this, this invitation, this request for dialogue, to respond and to ensure that, together with this House, these relics of a nationalist age can be abandoned.</code> |
419
  * Loss: [<code>CoSENTLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cosentloss) with these parameters:
420
  ```json
421
  {
 
730
  }
731
  ```
732
  </details>
733
+ <details><summary>tasksource/flan</summary>
734
+
735
+ #### tasksource/flan
736
+
737
+ * Dataset: tasksource/flan
738
+ * Size: 200,000 training samples
739
+ * Columns: <code>anchor</code>, <code>positive</code>, <code>rejected</code>, <code>task</code>, and <code>template_type</code>
740
+ * Approximate statistics based on the first 1000 samples:
741
+ | | anchor | positive | rejected | task | template_type |
742
+ |:--------|:--------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:--------------------------------------------------------------------------------|
743
+ | type | string | string | string | string | string |
744
+ | details | <ul><li>min: 72 tokens</li><li>mean: 513.33 tokens</li><li>max: 1024 tokens</li></ul> | <ul><li>min: 3 tokens</li><li>mean: 8.69 tokens</li><li>max: 74 tokens</li></ul> | <ul><li>min: 3 tokens</li><li>mean: 7.59 tokens</li><li>max: 74 tokens</li></ul> | <ul><li>min: 14 tokens</li><li>mean: 16.38 tokens</li><li>max: 19 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 5.97 tokens</li><li>max: 7 tokens</li></ul> |
745
+ * Samples:
746
+ | anchor | positive | rejected | task | template_type |
747
+ |:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:----------------------|:------------------------------------------|:---------------------------------------------------------|:--------------------|
748
+ | <code>Please answer the following question: Given the following passage "A healthy, and legal, publishing industry existed throughout Europe, although established publishers and book sellers occasionally ran afoul of the law. The Encyclopédie, for example, condemned not only by the King but also by Clement XII, nevertheless found its way into print with the help of the aforementioned Malesherbes and creative use of French censorship law. But many works were sold without running into any legal trouble at all. Borrowing records from libraries in England, Germany and North America indicate that more than 70 percent of books borrowed were novels. Less than 1 percent of the books were of a religious nature, indicating the general trend of declining religiosity.", answer the following question. Note that the answer is present within the text. Question: What country was omitted from the borrowing records?<br>Answer:</code> | <code>French</code> | <code>US</code> | <code>adversarial_qa_dbert_answer_the_following_q</code> | <code>zs_opt</code> |
749
+ | <code>Problem: Given the question: Given the following passage "On 3 December, Chopin complained about his bad health and the incompetence of the doctors in Majorca: "Three doctors have visited me ... The first said I was dead; the second said I was dying; and the third said I was about to die." He also had problems having his Pleyel piano sent to him. It finally arrived from Paris in December. Chopin wrote to Pleyel in January 1839: "I am sending you my Preludes [(Op. 28)]. I finished them on your little piano, which arrived in the best possible condition in spite of the sea, the bad weather and the Palma customs." Chopin was also able to undertake work on his Ballade No. 2, Op. 38; two Polonaises, Op. 40; and the Scherzo No. 3, Op. 39.", answer the following question. Note that the answer is present within the text. Question: Which doctor did not think Chopin was dead?<br>++++++++++++++++++++++++++++++++<br>The answer is:<br>the third<br><br><br>Problem: Given the question: Given the following passage "...</code> | <code>Jehovah</code> | <code>Ashkenazi and Sephardic Jews</code> | <code>adversarial_qa_dbert_answer_the_following_q</code> | <code>fs_opt</code> |
750
+ | <code>input: Please answer the following: Given the following passage "Jehovah's Witnesses are perhaps best known for their efforts to spread their beliefs, most notably by visiting people from house to house, distributing literature published by the Watch Tower Society in 700 languages. The objective is to start a regular "Bible study" with any person who is not already a member, with the intention that the student be baptized as a member of the group; Witnesses are advised to consider discontinuing Bible studies with students who show no interest in becoming members. Witnesses are taught they are under a biblical command to engage in public preaching. They are instructed to devote as much time as possible to their ministry and are required to submit an individual monthly "Field Service Report". Baptized members who fail to report a month of preaching are termed "irregular" and may be counseled by elders; those who do not submit reports for six consecutive months are termed "inactive".", ...</code> | <code>reformed</code> | <code>The Verge</code> | <code>adversarial_qa_dbert_answer_the_following_q</code> | <code>fs_opt</code> |
751
+ * Loss: [<code>CachedMultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cachedmultiplenegativesrankingloss) with these parameters:
752
+ ```json
753
+ {
754
+ "scale": 20.0,
755
+ "similarity_fct": "cos_sim",
756
+ "mini_batch_size": 32,
757
+ "gather_across_devices": false
758
+ }
759
+ ```
760
+ </details>
761
 
762
  ### Training Hyperparameters
763
  #### Non-Default Hyperparameters
 
896
  ### Training Logs
897
  | Epoch | Step | Training Loss |
898
  |:------:|:-----:|:-------------:|
899
+ | 0.0347 | 500 | 5.015 |
900
+ | 0.0694 | 1000 | 3.6072 |
901
+ | 0.1041 | 1500 | 3.6206 |
902
+ | 0.1387 | 2000 | 3.5205 |
903
+ | 0.1734 | 2500 | 3.8717 |
904
+ | 0.2081 | 3000 | 3.2363 |
905
+ | 0.2428 | 3500 | 3.0776 |
906
+ | 0.2775 | 4000 | 3.1094 |
907
+ | 0.3122 | 4500 | 3.3586 |
908
+ | 0.3468 | 5000 | 3.2504 |
909
+ | 0.3815 | 5500 | 2.9393 |
910
+ | 0.4162 | 6000 | 2.8626 |
911
+ | 0.4509 | 6500 | 3.1186 |
912
+ | 0.4856 | 7000 | 2.9852 |
913
+ | 0.5203 | 7500 | 2.8228 |
914
+ | 0.5549 | 8000 | 2.9656 |
915
+ | 0.5896 | 8500 | 2.6737 |
916
+ | 0.6243 | 9000 | 2.6191 |
917
+ | 0.6590 | 9500 | 2.7254 |
918
+ | 0.6937 | 10000 | 2.6937 |
919
+ | 0.7284 | 10500 | 2.8813 |
920
+ | 0.7630 | 11000 | 2.8221 |
921
+ | 0.7977 | 11500 | 2.9132 |
922
+ | 0.8324 | 12000 | 2.4675 |
923
+ | 0.8671 | 12500 | 2.9528 |
924
+ | 0.9018 | 13000 | 2.4298 |
925
+ | 0.9365 | 13500 | 2.3748 |
926
+ | 0.9711 | 14000 | 2.6557 |
927
 
928
 
929
  ### Framework Versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:834345ed278e11adc629a56fb2e40f7ce67d922f9bce2e7da5185c5ac560e0bc
3
  size 67193928
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:42965efac128bdb06678f2ffbda2054bb31f286f68c10a5eea5796a170fd68b4
3
  size 67193928