nreimers commited on
Commit
5d079c4
1 Parent(s): e1f20cf

Add new SentenceTransformer model.

Browse files

1_Pooling/config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
1
+ {
2
+ "word_embedding_dimension": 768,
3
+ "pooling_mode_cls_token": true,
4
+ "pooling_mode_mean_tokens": false,
5
+ "pooling_mode_max_tokens": false,
6
+ "pooling_mode_mean_sqrt_len_tokens": false
7
+ }
README.md CHANGED
@@ -1,71 +1,97 @@
1
  ---
2
- language: en
3
  tags:
4
- - exbert
5
  - sentence-transformers
6
  - feature-extraction
7
- license: apache-2.0
8
- datasets:
9
- - snli
10
- - multi_nli
11
  ---
12
 
13
- # BERT base model (uncased) for Sentence Embeddings
14
- This is the `bert-base-nli-cls-token` model from the [sentence-transformers](https://github.com/UKPLab/sentence-transformers)-repository. The sentence-transformers repository allows to train and use Transformer models for generating sentence and text embeddings.
15
- The model is described in the paper [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084)
16
 
17
- ## Usage (HuggingFace Models Repository)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
18
 
19
- You can use the model directly from the model repository to compute sentence embeddings. The CLS token of each input represents the sentence embedding:
20
  ```python
21
  from transformers import AutoTokenizer, AutoModel
22
  import torch
23
 
24
 
25
- #Sentences we want sentence embeddings for
26
- sentences = ['This framework generates embeddings for each input sentence',
27
- 'Sentences are passed as a list of string.',
28
- 'The quick brown fox jumps over the lazy dog.']
 
 
29
 
30
- #Load AutoModel from huggingface model repository
31
- tokenizer = AutoTokenizer.from_pretrained("sentence-transformers/bert-base-nli-cls-token")
32
- model = AutoModel.from_pretrained("sentence-transformers/bert-base-nli-cls-token")
33
 
34
- #Tokenize sentences
35
- encoded_input = tokenizer(sentences, padding=True, truncation=True, max_length=128, return_tensors='pt')
36
 
37
- #Compute token embeddings
38
  with torch.no_grad():
39
  model_output = model(**encoded_input)
40
- sentence_embeddings = model_output[0][:,0] #Take the first token ([CLS]) from each sentence
 
 
41
 
42
  print("Sentence embeddings:")
43
  print(sentence_embeddings)
44
  ```
45
 
46
- ## Usage (Sentence-Transformers)
47
- Using this model becomes more convenient when you have [sentence-transformers](https://github.com/UKPLab/sentence-transformers) installed:
48
- ```
49
- pip install -U sentence-transformers
50
- ```
51
 
52
- Then you can use the model like this:
53
- ```python
54
- from sentence_transformers import SentenceTransformer
55
- model = SentenceTransformer('bert-base-nli-cls-token')
56
- sentences = ['This framework generates embeddings for each input sentence',
57
- 'Sentences are passed as a list of string.',
58
- 'The quick brown fox jumps over the lazy dog.']
59
- sentence_embeddings = model.encode(sentences)
60
 
61
- print("Sentence embeddings:")
62
- print(sentence_embeddings)
63
- ```
 
 
 
64
 
65
 
 
 
 
 
 
 
 
 
66
  ## Citing & Authors
 
 
 
67
  If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084):
68
- ```
69
  @inproceedings{reimers-2019-sentence-bert,
70
  title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
71
  author = "Reimers, Nils and Gurevych, Iryna",
@@ -75,4 +101,4 @@ If you find this model helpful, feel free to cite our publication [Sentence-BERT
75
  publisher = "Association for Computational Linguistics",
76
  url = "http://arxiv.org/abs/1908.10084",
77
  }
78
- ```
1
  ---
2
+ pipeline_tag: sentence-similarity
3
  tags:
 
4
  - sentence-transformers
5
  - feature-extraction
6
+ - sentence-similarity
7
+ - transformers
8
+ - transformers
 
9
  ---
10
 
11
+ # sentence-transformers/bert-base-nli-cls-token
 
 
12
 
13
+ This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
14
+
15
+
16
+
17
+ ## Usage (Sentence-Transformers)
18
+
19
+ Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
20
+
21
+ ```
22
+ pip install -U sentence-transformers
23
+ ```
24
+
25
+ Then you can use the model like this:
26
+
27
+ ```python
28
+ from sentence_transformers import SentenceTransformer
29
+ sentences = ["This is an example sentence", "Each sentence is converted"]
30
+
31
+ model = SentenceTransformer('sentence-transformers/bert-base-nli-cls-token')
32
+ embeddings = model.encode(sentences)
33
+ print(embeddings)
34
+ ```
35
+
36
+
37
+
38
+ ## Usage (HuggingFace Transformers)
39
+ Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.
40
 
 
41
  ```python
42
  from transformers import AutoTokenizer, AutoModel
43
  import torch
44
 
45
 
46
+ def cls_pooling(model_output, attention_mask):
47
+ return model_output[0][:,0]
48
+
49
+
50
+ # Sentences we want sentence embeddings for
51
+ sentences = ['This is an example sentence', 'Each sentence is converted']
52
 
53
+ # Load model from HuggingFace Hub
54
+ tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/bert-base-nli-cls-token')
55
+ model = AutoModel.from_pretrained('sentence-transformers/bert-base-nli-cls-token')
56
 
57
+ # Tokenize sentences
58
+ encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
59
 
60
+ # Compute token embeddings
61
  with torch.no_grad():
62
  model_output = model(**encoded_input)
63
+
64
+ # Perform pooling. In this case, max pooling.
65
+ sentence_embeddings = cls_pooling(model_output, encoded_input['attention_mask'])
66
 
67
  print("Sentence embeddings:")
68
  print(sentence_embeddings)
69
  ```
70
 
 
 
 
 
 
71
 
 
 
 
 
 
 
 
 
72
 
73
+ ## Evaluation Results
74
+
75
+
76
+
77
+ For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/bert-base-nli-cls-token)
78
+
79
 
80
 
81
+ ## Full Model Architecture
82
+ ```
83
+ SentenceTransformer(
84
+ (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel
85
+ (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
86
+ )
87
+ ```
88
+
89
  ## Citing & Authors
90
+
91
+ This model was trained by [sentence-transformers](https://www.sbert.net/).
92
+
93
  If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084):
94
+ ```bibtex
95
  @inproceedings{reimers-2019-sentence-bert,
96
  title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
97
  author = "Reimers, Nils and Gurevych, Iryna",
101
  publisher = "Association for Computational Linguistics",
102
  url = "http://arxiv.org/abs/1908.10084",
103
  }
104
+ ```
config.json CHANGED
@@ -1,4 +1,5 @@
1
  {
 
2
  "architectures": [
3
  "BertModel"
4
  ],
@@ -15,6 +16,9 @@
15
  "num_attention_heads": 12,
16
  "num_hidden_layers": 12,
17
  "pad_token_id": 0,
 
 
18
  "type_vocab_size": 2,
 
19
  "vocab_size": 30522
20
- }
1
  {
2
+ "_name_or_path": "old_models/bert-base-nli-cls-token/0_BERT",
3
  "architectures": [
4
  "BertModel"
5
  ],
16
  "num_attention_heads": 12,
17
  "num_hidden_layers": 12,
18
  "pad_token_id": 0,
19
+ "position_embedding_type": "absolute",
20
+ "transformers_version": "4.7.0",
21
  "type_vocab_size": 2,
22
+ "use_cache": true,
23
  "vocab_size": 30522
24
+ }
config_sentence_transformers.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
1
+ {
2
+ "__version__": {
3
+ "sentence_transformers": "2.0.0",
4
+ "transformers": "4.7.0",
5
+ "pytorch": "1.9.0+cu102"
6
+ }
7
+ }
modules.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "idx": 0,
4
+ "name": "0",
5
+ "path": "",
6
+ "type": "sentence_transformers.models.Transformer"
7
+ },
8
+ {
9
+ "idx": 1,
10
+ "name": "1",
11
+ "path": "1_Pooling",
12
+ "type": "sentence_transformers.models.Pooling"
13
+ }
14
+ ]
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d413a6c8c71809c51608a6e267bf80002bde4e9192516ce16764a7bcc091fc85
3
- size 438006864
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ba777e44e81c45b2515e744582f383de507e2699d75ffd1e2d313bc0982fca8e
3
+ size 438007537
sentence_bert_config.json CHANGED
@@ -1,3 +1,4 @@
1
  {
2
- "max_seq_length": 128
 
3
  }
1
  {
2
+ "max_seq_length": 128,
3
+ "do_lower_case": false
4
  }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
1
+ {"do_lower_case": true, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "special_tokens_map_file": "old_models/bert-base-nli-cls-token/0_BERT/special_tokens_map.json", "name_or_path": "old_models/bert-base-nli-cls-token/0_BERT", "do_basic_tokenize": true, "never_split": null}