README.md
3.8 KB · 156 lines · markdown Raw
1 ---
2 language:
3 - multilingual
4 - ar
5 - bg
6 - ca
7 - cs
8 - da
9 - de
10 - el
11 - en
12 - es
13 - et
14 - fa
15 - fi
16 - fr
17 - gl
18 - gu
19 - he
20 - hi
21 - hr
22 - hu
23 - hy
24 - id
25 - it
26 - ja
27 - ka
28 - ko
29 - ku
30 - lt
31 - lv
32 - mk
33 - mn
34 - mr
35 - ms
36 - my
37 - nb
38 - nl
39 - pl
40 - pt
41 - ro
42 - ru
43 - sk
44 - sl
45 - sq
46 - sr
47 - sv
48 - th
49 - tr
50 - uk
51 - ur
52 - vi
53 license: apache-2.0
54 library_name: sentence-transformers
55 tags:
56 - sentence-transformers
57 - feature-extraction
58 - sentence-similarity
59 - transformers
60 language_bcp47:
61 - fr-ca
62 - pt-br
63 - zh-cn
64 - zh-tw
65 pipeline_tag: sentence-similarity
66 ---
67
68 # sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2
69
70 This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search.
71
72
73
74 ## Usage (Sentence-Transformers)
75
76 Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
77
78 ```
79 pip install -U sentence-transformers
80 ```
81
82 Then you can use the model like this:
83
84 ```python
85 from sentence_transformers import SentenceTransformer
86 sentences = ["This is an example sentence", "Each sentence is converted"]
87
88 model = SentenceTransformer('sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2')
89 embeddings = model.encode(sentences)
90 print(embeddings)
91 ```
92
93
94
95 ## Usage (HuggingFace Transformers)
96 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.
97
98 ```python
99 from transformers import AutoTokenizer, AutoModel
100 import torch
101
102
103 # Mean Pooling - Take attention mask into account for correct averaging
104 def mean_pooling(model_output, attention_mask):
105 token_embeddings = model_output[0] #First element of model_output contains all token embeddings
106 input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
107 return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
108
109
110 # Sentences we want sentence embeddings for
111 sentences = ['This is an example sentence', 'Each sentence is converted']
112
113 # Load model from HuggingFace Hub
114 tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2')
115 model = AutoModel.from_pretrained('sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2')
116
117 # Tokenize sentences
118 encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
119
120 # Compute token embeddings
121 with torch.no_grad():
122 model_output = model(**encoded_input)
123
124 # Perform pooling. In this case, max pooling.
125 sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
126
127 print("Sentence embeddings:")
128 print(sentence_embeddings)
129 ```
130
131
132
133 ## Full Model Architecture
134 ```
135 SentenceTransformer(
136 (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel
137 (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
138 )
139 ```
140
141 ## Citing & Authors
142
143 This model was trained by [sentence-transformers](https://www.sbert.net/).
144
145 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):
146 ```bibtex
147 @inproceedings{reimers-2019-sentence-bert,
148 title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
149 author = "Reimers, Nils and Gurevych, Iryna",
150 booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
151 month = "11",
152 year = "2019",
153 publisher = "Association for Computational Linguistics",
154 url = "http://arxiv.org/abs/1908.10084",
155 }
156 ```