Hugging Face's logo

Ehsanl
/
cncpt_e5

Sentence Similarity
sentence-transformers
Safetensors
xlm-roberta
feature-extraction
dense
Generated from Trainer
dataset_size:295188
loss:WeightedMultiPositiveCachedLoss
loss:MultipleNegativesRankingLoss
Eval Results (legacy)
text-embeddings-inference
Model card Files
xet
 Community
Ehsanl commited on
Commit
1997bc7
·
verified ·
1 Parent(s): 3921d13

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +0 -167
README.md CHANGED
@@ -9,82 +9,6 @@ tags:
9
  - loss:WeightedMultiPositiveCachedLoss
10
  - loss:MultipleNegativesRankingLoss
11
  base_model: intfloat/multilingual-e5-large-instruct
12
- widget:
13
- - source_sentence: 'Instruct: Given a Dutch legal question or case description, retrieve
14
- the legal concepts it involves.
15
-
16
- Query: mag je werkkleding van de zaak zelf thuis wassen'
17
- sentences:
18
- - 'parameter: gewastype'
19
- - 'parameter: gewassoort'
20
- - 'parameter: na reiniging'
21
- - 'parameter: schoonmaken'
22
- - 'regulated issue: hygiëne-eisen'
23
- - 'parameter: gewas'
24
- - 'parameter: reiniging'
25
- - 'parameter: gewasnaam'
26
- - 'regulated issue: hygiëne'
27
- - 'regulated issue: beroepskleding'
28
- - source_sentence: 'Instruct: Given a Dutch legal question or case description, retrieve
29
- the legal concepts it involves.
30
-
31
- Query: -'
32
- sentences:
33
- - 'parameter: gegoedheid'
34
- - 'parameter: onvoltooide manoeuvre'
35
- - 'parameter: oploopmanoeuvre'
36
- - 'parameter: voorrangsweg'
37
- - 'legal mechanism: verkeersregel'
38
- - 'regulated issue: aansprakelijkheid voor verkeersongevallen'
39
- - 'remedy: vergoeding van schaden'
40
- - 'parameter: rijbaan voorrangsweg'
41
- - 'parameter: keren'
42
- - 'parameter: voorrangsregeling'
43
- - source_sentence: 'Instruct: Given a Dutch legal question or case description, retrieve
44
- the legal concepts it involves.
45
-
46
- Query: verbod op bodemverharding noordzee zonder watervergunning'
47
- sentences:
48
- - 'parameter: ner'
49
- - 'parameter: bodemberoering'
50
- - 'legal mechanism: verbodsrecht'
51
- - 'parameter: norea'
52
- - 'regulated issue: gebruik noordzee'
53
- - 'parameter: installeren'
54
- - 'parameter: noordzee'
55
- - 'parameter: bodemhelling'
56
- - 'parameter: bodemophoging'
57
- - 'regulated issue: noordzeeactiviteiten'
58
- - source_sentence: 'Instruct: Given a Dutch legal question or case description, retrieve
59
- the legal concepts it involves.
60
-
61
- Query: veiligheidseisen landing luchtballon voor dieren op de grond'
62
- sentences:
63
- - 'parameter: dieren'
64
- - 'parameter: begrip zaak'
65
- - 'regulated issue: handhaving veiligheid'
66
- - 'legal mechanism: gevaarzetting'
67
- - 'parameter: zoogdieren'
68
- - 'regulated issue: landing'
69
- - 'parameter: zaken'
70
- - 'parameter: bepaalde zaak'
71
- - 'parameter: per dier'
72
- - 'legal mechanism: gevaarschepping'
73
- - source_sentence: 'Instruct: Given a Dutch legal question or case description, retrieve
74
- the legal concepts it involves.
75
-
76
- Query: uitzonderingsgronden voor het verbod op vluchten tegen vergoeding'
77
- sentences:
78
- - 'parameter: luchtrecht'
79
- - 'parameter: luchtwerk'
80
- - 'parameter: ballonnen'
81
- - 'parameter: luchtschip'
82
- - 'legal mechanism: vergunningsvereiste'
83
- - 'parameter: luchtvaartuigen'
84
- - 'parameter: lucht'
85
- - 'parameter: vrije ballon'
86
- - 'legal mechanism: vergunningvereiste'
87
- - 'parameter: luchtvering'
88
  pipeline_tag: sentence-similarity
89
  library_name: sentence-transformers
90
  metrics:
@@ -181,65 +105,6 @@ SentenceTransformer(
181
  (2): Normalize({'module_input_name': 'sentence_embedding', 'module_output_name': 'sentence_embedding'})
182
  )
183
  ```
184
-
185
- ## Usage
186
-
187
- ### Direct Usage (Sentence Transformers)
188
-
189
- First install the Sentence Transformers library:
190
-
191
- ```bash
192
- pip install -U sentence-transformers
193
- ```
194
- Then you can load this model and run inference.
195
- ```python
196
- from sentence_transformers import SentenceTransformer
197
-
198
- # Download from the 🤗 Hub
199
- model = SentenceTransformer("sentence_transformers_model_id")
200
- # Run inference
201
- queries = [
202
- 'Instruct: Given a Dutch legal question or case description, retrieve the legal concepts it involves.\nQuery: uitzonderingsgronden voor het verbod op vluchten tegen vergoeding',
203
- ]
204
- documents = [
205
- 'parameter: vrije ballon',
206
- 'parameter: luchtschip',
207
- 'parameter: luchtwerk',
208
- ]
209
- query_embeddings = model.encode_query(queries)
210
- document_embeddings = model.encode_document(documents)
211
- print(query_embeddings.shape, document_embeddings.shape)
212
- # [1, 1024] [3, 1024]
213
-
214
- # Get the similarity scores for the embeddings
215
- similarities = model.similarity(query_embeddings, document_embeddings)
216
- print(similarities)
217
- # tensor([[0.3279, 0.3127, 0.2823]])
218
- ```
219
- <!--
220
- ### Direct Usage (Transformers)
221
-
222
- <details><summary>Click to see the direct usage in Transformers</summary>
223
-
224
- </details>
225
- -->
226
-
227
- <!--
228
- ### Downstream Usage (Sentence Transformers)
229
-
230
- You can finetune this model on your own dataset.
231
-
232
- <details><summary>Click to expand</summary>
233
-
234
- </details>
235
- -->
236
-
237
- <!--
238
- ### Out-of-Scope Use
239
-
240
- *List how the model may foreseeably be misused and address what users ought not to do with the model.*
241
- -->
242
-
243
  ## Evaluation
244
 
245
  ### Metrics
@@ -282,38 +147,6 @@ You can finetune this model on your own dataset.
282
 
283
  #### json
284
 
285
- * Dataset: json
286
- * Size: 295,188 training samples
287
- * Columns: <code>anchor</code>, <code>label</code>, <code>candidate_0</code>, <code>candidate_1</code>, <code>candidate_2</code>, <code>candidate_3</code>, <code>candidate_4</code>, <code>candidate_5</code>, <code>candidate_6</code>, <code>candidate_7</code>, <code>candidate_8</code>, and <code>candidate_9</code>
288
- * Approximate statistics based on the first 100 samples:
289
- | | anchor | label | candidate_0 | candidate_1 | candidate_2 | candidate_3 | candidate_4 | candidate_5 | candidate_6 | candidate_7 | candidate_8 | candidate_9 |
290
- |:---------|:------------------------------------------------------------------------------------|:------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|
291
- | type | string | list | string | string | string | string | string | string | string | string | string | string |
292
- | modality | text | | text | text | text | text | text | text | text | text | text | text |
293
- | details | <ul><li>min: 30 tokens</li><li>mean: 86.84 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>size: 45 elements</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 9.15 tokens</li><li>max: 15 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 9.58 tokens</li><li>max: 15 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 8.68 tokens</li><li>max: 12 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 9.16 tokens</li><li>max: 13 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 9.69 tokens</li><li>max: 15 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 9.49 tokens</li><li>max: 14 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 9.49 tokens</li><li>max: 16 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 9.33 tokens</li><li>max: 14 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 9.13 tokens</li><li>max: 13 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 9.27 tokens</li><li>max: 15 tokens</li></ul> |
294
- * Samples:
295
- | anchor | label | candidate_0 | candidate_1 | candidate_2 | candidate_3 | candidate_4 | candidate_5 | candidate_6 | candidate_7 | candidate_8 | candidate_9 |
296
- |:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:--------------------------------------------|:--------------------------------------------------------|:----------------------------------------------------|:--------------------------------------------------------|:-----------------------------------------------|:--------------------------------------------------|:----------------------------------------------------------------|:----------------------------------------------|:----------------------------------------------------------|:--------------------------------------------------|:------------------------------------------------------|
297
- | <code>Instruct: Given a Dutch legal question or case description, retrieve the legal concepts it involves.<br>Query: Een burger krijgt een Wahv-sanctie omdat zijn bromfiets op een rollentestbank een te hoge constructiesnelheid behaalde. De burger dient een bezwaarschrift in waarin hij stelt dat hij de boete van € 120,- niet kan betalen door schulden en verzoekt om matiging. De officier van justitie wijst dit schriftelijk af met de enkele reden: "De wet kent geen regeling voor matiging van verkeersboetes op basis van inkomen of vermogen." De burger voert daarnaast aan dat de meting ongeldig is omdat het ijkrapport van de rollentestbank niet bij de stukken zit en de opsporingsambtenaar zelf niet op de bromfiets zat tijdens de test, wat volgens de burger de weerstand beïnvloedt. Is het ontbreken van een keuringscertificaat in het dossier fataal voor de bewijskracht van de meting?</code> | <code>[1.0, 1.0, 1.0, 1.0, 0.0, ...]</code> | <code>legal mechanism: passeerregeling</code> | <code>parameter: benadelingscriterium</code> | <code>parameter: draagkracht</code> | <code>remedy: vernietiging van zaken</code> | <code>parameter: draagvermogen</code> | <code>legal mechanism: motiveringseis</code> | <code>remedy: gehele vernietiging</code> | <code>remedy: vergoeding van betaalde proceskosten</code> | <code>legal mechanism: motiveringsvereiste</code> | <code>remedy: proceskostencompensatie</code> |
298
- | <code>Instruct: Given a Dutch legal question or case description, retrieve the legal concepts it involves.<br>Query: kentekenaansprakelijkheid eigenaar niet-bestuurder verweer</code> | <code>[1.0, 1.0, 1.0, 1.0, 0.0, ...]</code> | <code>legal mechanism: kentekenaansprakelijkheid</code> | <code>regulated issue: snelheidsovertreding</code> | <code>parameter: ernstig vermoeden</code> | <code>legal mechanism: novum-vereiste</code> | <code>remedy: herziening bekostiging</code> | <code>legal mechanism: kentekenopgave</code> | <code>legal mechanism: kentekenplicht</code> | <code>regulated issue: snelheidscontrole</code> | <code>remedy: herzieningsprocedure</code> | <code>regulated issue: snelheidsoverschrijding</code> |
299
- | <code>Instruct: Given a Dutch legal question or case description, retrieve the legal concepts it involves.<br>Query: vermogensvermindering WUV invloed uitkeringsgerechtigde</code> | <code>[1.0, 1.0, 1.0, 1.0, 0.0, ...]</code> | <code>regulated issue: vervolgingsslachtoffers</code> | <code>regulated issue: vermogensvaststelling</code> | <code>parameter: nieuwe feiten of omstandigheden</code> | <code>legal mechanism: marginaltoetsing</code> | <code>regulated issue: vermogensopstelling</code> | <code>regulated issue: uitkering vervolgingsslachtoffers</code> | <code>legal mechanism: marginale toets</code> | <code>legal mechanism: discretionaire bevoegheid</code> | <code>legal mechanism: marginaletoetsing</code> | <code>remedy: herzieningsprocedure</code> |
300
- * Loss: <code>multipositive_loss.WeightedMultiPositiveCachedLoss</code> with these parameters:
301
- ```json
302
- {
303
- "scale": 20.0,
304
- "similarity_fct": "cos_sim",
305
- "mini_batch_size": 64,
306
- "mini_batch_num_tokens": null,
307
- "gather_across_devices": false,
308
- "directions": [
309
- "query_to_doc"
310
- ],
311
- "partition_mode": "joint",
312
- "hardness_mode": null,
313
- "hardness_strength": 0.0,
314
- "n_answers": 24
315
- }
316
- ```
317
 
318
  ### Training Hyperparameters
319
  #### Non-Default Hyperparameters
 
9
  - loss:WeightedMultiPositiveCachedLoss
10
  - loss:MultipleNegativesRankingLoss
11
  base_model: intfloat/multilingual-e5-large-instruct
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
12
  pipeline_tag: sentence-similarity
13
  library_name: sentence-transformers
14
  metrics:
 
105
  (2): Normalize({'module_input_name': 'sentence_embedding', 'module_output_name': 'sentence_embedding'})
106
  )
107
  ```
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
108
  ## Evaluation
109
 
110
  ### Metrics
 
147
 
148
  #### json
149
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
150
 
151
  ### Training Hyperparameters
152
  #### Non-Default Hyperparameters