am-azadi commited on
Commit
4084353
·
verified ·
1 Parent(s): 9a0c83b

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. README.md +151 -90
  2. config.json +1 -1
  3. model.safetensors +1 -1
README.md CHANGED
@@ -4,72 +4,138 @@ tags:
4
  - sentence-similarity
5
  - feature-extraction
6
  - generated_from_trainer
7
- - dataset_size:25743
8
  - loss:MultipleNegativesRankingLoss
9
- base_model: am-azadi/bilingual-embedding-large_Fine_Tuned
10
  widget:
11
- - source_sentence: Anabel Hernandez This morning I received this in an envelope to
12
- my name that said "We're coming for you." Both me and my family we feel safe.
13
- This happens when the government Instead of defending the press is the first that
14
- we attack. 00| After exhibiting negotiations between the Morenoite government
15
- and organized crime, drug trafficking expert Anabel Hernández denounces being
16
- harassed and receiving death threats. We demand to activate the protection protocol
17
- for journalists
18
  sentences:
19
- - Covid-19 vaccines contain tracking devices
20
- - The Mexican journalist Anabel Hernández denounced on Twitter a bullet threat received
21
- in an envelope
22
- - Video shows Indian PM Modi giving a speech in January 2022 in Goa
23
- - source_sentence: Richard German 19 min. According to a publication of portal Plain
24
- Words, the mother of the Chapo Guzman, Maria Consuelo Loera Pérez contributed
25
- to Morena a total of 900 millions. brunette pr FOUNDER BONG $900,000,000 fometre.
26
- Marma of mine ... September 12 48149 To Yo K R 89 art Rib Parts Hodon yesWITHOUT
27
- WORDS
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
28
  sentences:
29
- - Members of the Vox political party in Spain hold a banner against a sex toy
30
- - Image shows Islamic missionaries killed in Pakistan
31
- - The mother of "Chapo" Guzmán donated 900 million pesos to Morena
32
- - source_sentence: EXA playground for children and children Stay from 7am to 8pm permitted.
33
- head of state Hanover subject area Un and city green Languatastro17 2016 ver Telepo,
34
- 160-439 11It really is time to stop giving these idiots any more leeway!
 
 
 
35
  sentences:
36
- - The city of Hanover changed the word "children" to "children" on a playground
37
- sign
38
- - Russia Did Target Africa With Missiles For Talking About Its War With Ukraine
39
- - Periodic table assembled to say “F**k Boris” in front of the UK PM while he attends
40
- a press conference
41
- - source_sentence: on the encounter Why is Faith sad?When Telangana Police killed
42
- the victims of Hyderabad rape and murder case in an encounter, Kavi Kumar Vishwas
43
- gave a big statement! , , ,
 
 
 
 
44
  sentences:
45
- - Photo of suspects killed by police in Hyderabad rape-murder case
46
- - A video in which the water washes away several buildings, supposedly in Turkey,
47
- has been shared more than 220,000 times on Facebook. In reality, it is an altered
48
- version of a video recorded during the tsunami that hit Japan in 2011. In 9 minutes
49
- an entire neighborhood in Turkey disappears
50
- - “vaccinated individuals carry 251 times the load of COVID-19 viruses in their
51
- nostrils compared to the unvaccinated”
52
- - source_sentence: Well it is Australian Dr's point of view that Queen Elizabeth may
53
- benefit from Ivermectin, that's right common old horse paste.
 
54
  sentences:
55
- - Australian doctor told news programme that Queen Elizabeth II should use ivermectin
56
- to treat Covid-19
57
- - These 12 rights are no longer in the proposed Chilean Constitution
58
- - “If you want to know who controls you, just look at whom you cannot criticize,”
59
- Voltaire said.
60
  pipeline_tag: sentence-similarity
61
  library_name: sentence-transformers
62
  ---
63
 
64
- # SentenceTransformer based on am-azadi/bilingual-embedding-large_Fine_Tuned
65
 
66
- This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [am-azadi/bilingual-embedding-large_Fine_Tuned](https://huggingface.co/am-azadi/bilingual-embedding-large_Fine_Tuned). It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
67
 
68
  ## Model Details
69
 
70
  ### Model Description
71
  - **Model Type:** Sentence Transformer
72
- - **Base model:** [am-azadi/bilingual-embedding-large_Fine_Tuned](https://huggingface.co/am-azadi/bilingual-embedding-large_Fine_Tuned) <!-- at revision 0b9b4a8c5e6212588461d6df7c9428f571b1c823 -->
73
  - **Maximum Sequence Length:** 512 tokens
74
  - **Output Dimensionality:** 1024 dimensions
75
  - **Similarity Function:** Cosine Similarity
@@ -111,9 +177,9 @@ from sentence_transformers import SentenceTransformer
111
  model = SentenceTransformer("sentence_transformers_model_id")
112
  # Run inference
113
  sentences = [
114
- "Well it is Australian Dr's point of view that Queen Elizabeth may benefit from Ivermectin, that's right common old horse paste.",
115
- 'Australian doctor told news programme that Queen Elizabeth II should use ivermectin to treat Covid-19',
116
- 'These 12 rights are no longer in the proposed Chilean Constitution',
117
  ]
118
  embeddings = model.encode(sentences)
119
  print(embeddings.shape)
@@ -167,19 +233,19 @@ You can finetune this model on your own dataset.
167
 
168
  #### Unnamed Dataset
169
 
170
- * Size: 25,743 training samples
171
- * Columns: <code>sentence_0</code>, <code>sentence_1</code>, and <code>label</code>
172
  * Approximate statistics based on the first 1000 samples:
173
- | | sentence_0 | sentence_1 | label |
174
- |:--------|:------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:--------------------------------------------------------------|
175
- | type | string | string | float |
176
- | details | <ul><li>min: 2 tokens</li><li>mean: 116.09 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 19.96 tokens</li><li>max: 130 tokens</li></ul> | <ul><li>min: 1.0</li><li>mean: 1.0</li><li>max: 1.0</li></ul> |
177
  * Samples:
178
- | sentence_0 | sentence_1 | label |
179
- |:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-----------------|
180
- | <code>10 66% 8:06 PM M<R RIVALS Extinction Rebellion Member charged with starting Green Wattle Creek bushfi n Hussey AAP/7NEWS Sunday, 5 January 2020 12:44 pm FILE IMAGE: A 26-year-old Man from Sydney has been charged with intentionally lighting four fire Credit: KELLY BARNES/AAP A 26-year-old Man from Sydney has been charged with intentionally lighting four grass and scrub fires in the state's southeast in recent days Police will allege the man lit fires on December 15 and January 1, and th two on Saturday, all around the Anderson Fire Trail, Wentworth Falls In the video above: Bushfires continue to rage across Australia It's all about the dessert.This taken from another page.. I don't know if it's true or not but this is what it said.. "This photo was taken of a story CH7 did and posted to their FB page. Now, that post has been removed and no trace of it can be found on google. Share it far and wide "</code> | <code>Extinction Rebellion member charged with bushfire arson</code> | <code>1.0</code> |
181
- | <code>Even with the coronavirus, no one wants to eat Vegan</code> | <code>A photo shows that, even with the coronavirus, no one wants to eat vegan</code> | <code>1.0</code> |
182
- | <code>CONFUSION IN BENI: For the Population of this Corner, the Mask has just Fallen from the MONUSCO-ADF Marriage. Attached, the Adf surprised in the Monusco facilities with FARDC and PNC outfits. Let's figure it out.</code> | <code>Members of the armed group of Ugandan origin of the Allied Democratic Forces (ADF), surprised in the UN installations with the uniforms of the Congolese army</code> | <code>1.0</code> |
183
  * Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
184
  ```json
185
  {
@@ -194,7 +260,6 @@ You can finetune this model on your own dataset.
194
  - `per_device_train_batch_size`: 2
195
  - `per_device_eval_batch_size`: 2
196
  - `num_train_epochs`: 1
197
- - `fp16`: True
198
  - `multi_dataset_batch_sampler`: round_robin
199
 
200
  #### All Hyperparameters
@@ -239,7 +304,7 @@ You can finetune this model on your own dataset.
239
  - `jit_mode_eval`: False
240
  - `use_ipex`: False
241
  - `bf16`: False
242
- - `fp16`: True
243
  - `fp16_opt_level`: O1
244
  - `half_precision_backend`: auto
245
  - `bf16_full_eval`: False
@@ -321,31 +386,27 @@ You can finetune this model on your own dataset.
321
  ### Training Logs
322
  | Epoch | Step | Training Loss |
323
  |:------:|:-----:|:-------------:|
324
- | 0.0388 | 500 | 0.0125 |
325
- | 0.0777 | 1000 | 0.0072 |
326
- | 0.1165 | 1500 | 0.0209 |
327
- | 0.1554 | 2000 | 0.0112 |
328
- | 0.1942 | 2500 | 0.0268 |
329
- | 0.2331 | 3000 | 0.017 |
330
- | 0.2719 | 3500 | 0.0211 |
331
- | 0.3108 | 4000 | 0.039 |
332
- | 0.3496 | 4500 | 0.013 |
333
- | 0.3884 | 5000 | 0.0225 |
334
- | 0.4273 | 5500 | 0.0182 |
335
- | 0.4661 | 6000 | 0.0208 |
336
- | 0.5050 | 6500 | 0.0071 |
337
- | 0.5438 | 7000 | 0.0071 |
338
- | 0.5827 | 7500 | 0.0132 |
339
- | 0.6215 | 8000 | 0.0101 |
340
- | 0.6603 | 8500 | 0.015 |
341
- | 0.6992 | 9000 | 0.0062 |
342
- | 0.7380 | 9500 | 0.0037 |
343
- | 0.7769 | 10000 | 0.0061 |
344
- | 0.8157 | 10500 | 0.0056 |
345
- | 0.8546 | 11000 | 0.0084 |
346
- | 0.8934 | 11500 | 0.0208 |
347
- | 0.9323 | 12000 | 0.0052 |
348
- | 0.9711 | 12500 | 0.0081 |
349
 
350
 
351
  ### Framework Versions
 
4
  - sentence-similarity
5
  - feature-extraction
6
  - generated_from_trainer
7
+ - dataset_size:21988
8
  - loss:MultipleNegativesRankingLoss
9
+ base_model: Lajavaness/bilingual-embedding-large
10
  widget:
11
+ - source_sentence: BEAS INTESTINES 2901 718935 wwwIsrael under heavy attack from Gaza
12
+ There were more than 600 rockets launched against Israel. There are some civilians
13
+ wounded and dead
 
 
 
 
14
  sentences:
15
+ - Photo shows cloud of smoke after attack in Israel
16
+ - Claudia López with a book thanking the FARC
17
+ - Wife of Chinese official shot in US
18
+ - source_sentence: 'People''s Network people.cn People''s Daily: Scientifically grasp
19
+ the law of population development Balanced Population Development in the New Era
20
+ - January 2022 From the 1st, the one-child policy will be completely abolished.
21
+ Newlyweds must have at least two children Wang Peian April 1, 2021 06:18 Source:
22
+ People''s Daily Online, People''s Daily Executive summary: ■After the founding
23
+ of New China, the implementation of family planning was based on the basic national
24
+ conditions of my country''s large population and relatively insufficient resources
25
+ A major strategic decision, which makes the population''s pressure on resources
26
+ and the environment get a preliminary understanding: it creates a longer demographic
27
+ dividend period, It has effectively promoted economic development, social progress
28
+ and the improvement of people''s living standards, and the country''s capacity
29
+ for sustainable development has been greatly enhanced. ■Since the beginning of
30
+ the new century, my country''s population situation has undergone major changes.
31
+ Strive to achieve the level of active fertility, vigorously improve the quality
32
+ and skills of workers, and implement the comprehensive two-child policy, which
33
+ is the key to population development. Three issues that must be addressed in the
34
+ field. ■ Attention should be paid to the research on population development strategies,
35
+ comprehensively and profoundly understand and grasp the laws of population, and
36
+ promote the coordination between population and economy and society. development,
37
+ and promote the long-term balanced development of the population. choice of history
38
+ my country has been a country with the largest population in the world since ancient
39
+ times. In traditional society, if there is an entrance, there will be a license
40
+ and tax, and the country will be strengthened. If there is a population, there
41
+ will be soldiers. The rulers of successive dynasties have vigorously encouraged
42
+ population reproduction. Once the society is stable and production develops, the
43
+ total population will decrease. The threshold will increase greatly; when the
44
+ dynasty is changed, the army will be in chaos, famine and flag epidemics will
45
+ be intertwined, and the population will be sharp or small. Look, before the 17th
46
+ century, my country''s population grew slowly in a cyclical ups and downs. The
47
+ introduction of high-yielding food crops such as corn, sweet potato and potato
48
+ in the late Ming Dynasty, especially the century-long Kanggan in the early Qing
49
+ Dynasty. The prosperous age made my country''s population grow rapidly, breaking
50
+ through the 200 million, 300 million mark successively, and the 400 million mark
51
+ in the Daoguang years, which led to Legal Migrant Workers People''s Network people.cn
52
+ People''s Daily: Scientifically grasp the law of population development Balanced
53
+ Population Development in the New Era - January 2022 From the 1st, the one-child
54
+ policy will be completely abolished. Newlyweds must have at least two children
55
+ Wang Peian April 1, 2021 06:18 Source: People''s Daily Online, People''s Daily
56
+ Executive summary: ■After the founding of New China, the implementation of family
57
+ planning was based on the basic national conditions of my country''s large population
58
+ and relatively insufficient resources A major strategic decision, which makes
59
+ the population''s pressure on resources and the environment get a preliminary
60
+ understanding: it creates a longer demographic dividend period, It has effectively
61
+ promoted economic development, social progress and the improvement of people''s
62
+ living standards, and the country''s capacity for sustainable development has
63
+ been greatly enhanced. ■Since the beginning of the new century, my country''s
64
+ population situation has undergone major changes. Strive to achieve the level
65
+ of active fertility, vigorously improve the quality and skills of workers, and
66
+ implement the comprehensive two-child policy, which is the key to population development.
67
+ Three issues that must be addressed in the field. ■ Attention should be paid to
68
+ the research on population development strategies, comprehensively and profoundly
69
+ understand and grasp the laws of population, and promote the coordination between
70
+ population and economy and society. development, and promote the long-term balanced
71
+ development of the population. choice of history my country has been a country
72
+ with the largest population in the world since ancient times. In traditional society,
73
+ if there is an entrance, there will be a license and tax, and the country will
74
+ be strengthened. If there is a population, there will be soldiers. The rulers
75
+ of successive dynasties have vigorously encouraged population reproduction. Once
76
+ the society is stable and production develops, the total population will decrease.
77
+ The threshold will increase greatly; when the dynasty is changed, the army will
78
+ be in chaos, famine and flag epidemics will be intertwined, and the population
79
+ will be sharp or small. Look, before the 17th century, my country''s population
80
+ grew slowly in a cyclical ups and downs. The introduction of high-yielding food
81
+ crops such as corn, sweet potato and potato in the late Ming Dynasty, especially
82
+ the century-long Kanggan in the early Qing Dynasty. The prosperous age made my
83
+ country''s population grow rapidly, breaking through the 200 million, 300 million
84
+ mark successively, and the 400 million mark in the Daoguang years, which led to Legal
85
+ Migrant WorkersA warning to those prosperous forces who often talk about human
86
+ rights: China has human rights, and we have approved that Chinese people must
87
+ get married, and they must have two children after they get married!'
88
  sentences:
89
+ - Hamad bin Jassim told the BBC In a new interview, we paid the defected Syrian
90
+ officer $30,000 and the regular soldier $15,000.
91
+ - State-run newspaper announces Chinese couples ‘must have two children’ starting
92
+ January 2022
93
+ - This is the draw for judges for the case of former Ecuadorian President Rafael
94
+ Correa
95
+ - source_sentence: Part 1 Resignation sir jokowi JOKOWI REGISTERED COMPASS DKI DPRD
96
+ HOLDS Plenary MEETING CARIS JAKARTA KOMPASTV Tik TokIs it true that the President
97
+ of Indonesia, Joko Widodo, has resigned from his position?
98
  sentences:
99
+ - BBC reports on release of 'Unabomber' Ted Kaczynski
100
+ - Thai children flash three fingered salute to Thai PM Prayut
101
+ - President Joko Widodo, alias Jokowi, resigns from his post
102
+ - source_sentence: The organization 'Vegan Society' calls for a ban on animal-shaped
103
+ children's cookies. They consider that these cookies "incite children to see animals
104
+ as something inferior and at our disposal." This is the , which is dangerous even
105
+ for anti-bullfighting. It's not that they don't want bullfighting. It is that
106
+ they want to impose even the shape of the cookies that your children eat. And
107
+ it's not the first time. Barnum cookies have already "freed" the animals in their
108
+ boxes to have a better brand image. They may seem like funny news. But they are
109
+ not. They hide a prohibitionist ideology full of censorship. 𝗘𝗹 𝗮𝗻𝗶𝗺����𝗹𝗶𝘀𝗺𝗼 𝗲𝘀
110
+ 𝗽𝗲𝗹𝗶𝗴𝗿𝗼 𝗽𝗮𝗿𝗮 𝗻𝘂𝗲𝘀𝘁𝗿𝗮 𝘀𝗼𝗰𝗶𝗲𝗱𝗮𝗱
111
  sentences:
112
+ - Vegan NGO Vegan Society wants to ban the sale of animal-shaped cookies in France
113
+ - Cans of food containing pork with a "halal" stamp
114
+ - Pfizer announces Covid-19 vaccine update with Microsoft chip for symptom reduction
115
+ - source_sentence: a . . . . . (177. FO Accident st THE LEADER IN ACCIDENT REPORTING
116
+ Reckless driving by a minor Kuliapitiya Kanadulla after a defender collided with
117
+ a motorcycle An accident occurred in front of Maha Vidyalaya today (01) afternoon
118
+ A young man on a motorcycle and about 4 years old A young child (father and son)
119
+ unfortunately Lost his life. Behaved provocatively with the accident Villagers
120
+ set fire to the defender car that caused the accident had May that innocent father
121
+ and little son rest in peace! 94 site
122
  sentences:
123
+ - The image of a Syrian child who sleeps next to the graves of his parents
124
+ - Accident kills four-year-old in northwestern Sri Lanka
125
+ - Masks are ineffective because some packaging says they don't protect
 
 
126
  pipeline_tag: sentence-similarity
127
  library_name: sentence-transformers
128
  ---
129
 
130
+ # SentenceTransformer based on Lajavaness/bilingual-embedding-large
131
 
132
+ This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Lajavaness/bilingual-embedding-large](https://huggingface.co/Lajavaness/bilingual-embedding-large). It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
133
 
134
  ## Model Details
135
 
136
  ### Model Description
137
  - **Model Type:** Sentence Transformer
138
+ - **Base model:** [Lajavaness/bilingual-embedding-large](https://huggingface.co/Lajavaness/bilingual-embedding-large) <!-- at revision e83179d7a66e8aed1b3015e98bb5ae234ed89598 -->
139
  - **Maximum Sequence Length:** 512 tokens
140
  - **Output Dimensionality:** 1024 dimensions
141
  - **Similarity Function:** Cosine Similarity
 
177
  model = SentenceTransformer("sentence_transformers_model_id")
178
  # Run inference
179
  sentences = [
180
+ 'a . . . . . (177. FO Accident st THE LEADER IN ACCIDENT REPORTING Reckless driving by a minor Kuliapitiya Kanadulla after a defender collided with a motorcycle An accident occurred in front of Maha Vidyalaya today (01) afternoon A young man on a motorcycle and about 4 years old A young child (father and son) unfortunately Lost his life. Behaved provocatively with the accident Villagers set fire to the defender car that caused the accident had May that innocent father and little son rest in peace! 94 site',
181
+ 'Accident kills four-year-old in northwestern Sri Lanka',
182
+ 'The image of a Syrian child who sleeps next to the graves of his parents',
183
  ]
184
  embeddings = model.encode(sentences)
185
  print(embeddings.shape)
 
233
 
234
  #### Unnamed Dataset
235
 
236
+ * Size: 21,988 training samples
237
+ * Columns: <code>sentence_0</code> and <code>sentence_1</code>
238
  * Approximate statistics based on the first 1000 samples:
239
+ | | sentence_0 | sentence_1 |
240
+ |:--------|:-----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|
241
+ | type | string | string |
242
+ | details | <ul><li>min: 2 tokens</li><li>mean: 119.9 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 7 tokens</li><li>mean: 19.25 tokens</li><li>max: 128 tokens</li></ul> |
243
  * Samples:
244
+ | sentence_0 | sentence_1 |
245
+ |:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------|
246
+ | <code>ANK DBS DBS IT department at ChangiThis is actually happening as confirmed by my brother who does contract work with DBS at Changi Business Park. Wonder if PAP knows this or turning a blind eye and pretending not to know.</code> | <code>Photo shows foreign staff of the IT department at DBS Bank in Singapore</code> |
247
+ | <code>29th 30th 31st 32nd 33rd 34th 35th 36th 37th 38th 39th 40th 41st 42nd 43rd 44th 45th 46th 47th 48th 49th 50th 51st 52nd 53rd 54th 55th Urban Planning Foreign Languages Animal Science Law Economics Political Science Education Advertising Journalism Finance Hospitality Criminology Accounting Anthropology Psychology History Geography Information Technology Sociology Sports Science Social Sciences Real Estate Liberal Arts Communications and Mass Media Business Marketing Public Relations 1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th 13th 14th 15th 16th 17th 18th 19th 20th 21st 22nd 23rd 24th 25th 26th 27th 28th Architecture Chemical Engineering Chemistry Electrical Engineering Physics Mechanical Engineering Civil Engineering Biochemistry Medicine Pharmacy Engineering Nursing Math Biology Philosophy Mathematics Statistics Music Microbiology Psychology Accounting Finance Environmental Science Creative Writing Hospitality International Relations Art History Ecology55 most difficult course...</code> | <code>Harvard list of its 50 most difficult courses</code> |
248
+ | <code>The 30,000 sheep donated by Mongolia to China entered through the Erenhot port, which is very spectacular. [Qiang] Yesterday there were people who were worried about how to transport so many sheep. It turned out that they came by themselves, and they didn't even need transport tools.</code> | <code>These videos show 30,000 sheep donated to China by Mongolia during the novel coronavirus epidemic</code> |
249
  * Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
250
  ```json
251
  {
 
260
  - `per_device_train_batch_size`: 2
261
  - `per_device_eval_batch_size`: 2
262
  - `num_train_epochs`: 1
 
263
  - `multi_dataset_batch_sampler`: round_robin
264
 
265
  #### All Hyperparameters
 
304
  - `jit_mode_eval`: False
305
  - `use_ipex`: False
306
  - `bf16`: False
307
+ - `fp16`: False
308
  - `fp16_opt_level`: O1
309
  - `half_precision_backend`: auto
310
  - `bf16_full_eval`: False
 
386
  ### Training Logs
387
  | Epoch | Step | Training Loss |
388
  |:------:|:-----:|:-------------:|
389
+ | 0.0455 | 500 | 0.0505 |
390
+ | 0.0910 | 1000 | 0.0637 |
391
+ | 0.1364 | 1500 | 0.039 |
392
+ | 0.1819 | 2000 | 0.0269 |
393
+ | 0.2274 | 2500 | 0.0527 |
394
+ | 0.2729 | 3000 | 0.0576 |
395
+ | 0.3184 | 3500 | 0.0278 |
396
+ | 0.3638 | 4000 | 0.0471 |
397
+ | 0.4093 | 4500 | 0.0486 |
398
+ | 0.4548 | 5000 | 0.025 |
399
+ | 0.5003 | 5500 | 0.0324 |
400
+ | 0.5458 | 6000 | 0.0169 |
401
+ | 0.5912 | 6500 | 0.0218 |
402
+ | 0.6367 | 7000 | 0.0476 |
403
+ | 0.6822 | 7500 | 0.0124 |
404
+ | 0.7277 | 8000 | 0.0247 |
405
+ | 0.7731 | 8500 | 0.0231 |
406
+ | 0.8186 | 9000 | 0.01 |
407
+ | 0.8641 | 9500 | 0.0145 |
408
+ | 0.9096 | 10000 | 0.0267 |
409
+ | 0.9551 | 10500 | 0.0111 |
 
 
 
 
410
 
411
 
412
  ### Framework Versions
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "am-azadi/bilingual-embedding-large_Fine_Tuned",
3
  "architectures": [
4
  "BilingualModel"
5
  ],
 
1
  {
2
+ "_name_or_path": "Lajavaness/bilingual-embedding-large",
3
  "architectures": [
4
  "BilingualModel"
5
  ],
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d5027ccc9b2e2301ff74039dbd9841c268da1b7877d167ab214a86fa5392173f
3
  size 2239607176
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:544b5dea808c43e84c38838160a1d0090df6b1e8d839cebfb873915bcd19a15e
3
  size 2239607176