Update README.md
Browse files
README.md
CHANGED
@@ -6,42 +6,16 @@ tags:
|
|
6 |
- sentence-transformers
|
7 |
- sentence-similarity
|
8 |
- feature-extraction
|
9 |
-
base_model: tohoku-nlp/bert-
|
10 |
widget: []
|
11 |
pipeline_tag: sentence-similarity
|
12 |
license: apache-2.0
|
|
|
|
|
13 |
---
|
14 |
|
15 |
-
#
|
16 |
|
17 |
-
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [tohoku-nlp/bert-large-japanese-v2](https://huggingface.co/tohoku-nlp/bert-large-japanese-v2). 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.
|
18 |
-
|
19 |
-
## Model Details
|
20 |
-
|
21 |
-
### Model Description
|
22 |
-
- **Model Type:** Sentence Transformer
|
23 |
-
- **Base model:** [tohoku-nlp/bert-large-japanese-v2](https://huggingface.co/tohoku-nlp/bert-large-japanese-v2) <!-- at revision 75b828083735e953e3ed13e2ad6ea945c1fdb390 -->
|
24 |
-
- **Maximum Sequence Length:** 512 tokens
|
25 |
-
- **Output Dimensionality:** 1024 tokens
|
26 |
-
- **Similarity Function:** Cosine Similarity
|
27 |
-
<!-- - **Training Dataset:** Unknown -->
|
28 |
-
<!-- - **Language:** Unknown -->
|
29 |
-
<!-- - **License:** Unknown -->
|
30 |
-
|
31 |
-
### Model Sources
|
32 |
-
|
33 |
-
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
|
34 |
-
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
|
35 |
-
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
|
36 |
-
|
37 |
-
### Full Model Architecture
|
38 |
-
|
39 |
-
```
|
40 |
-
MySentenceTransformer(
|
41 |
-
(0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel
|
42 |
-
(1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
|
43 |
-
)
|
44 |
-
```
|
45 |
|
46 |
## Usage
|
47 |
|
@@ -55,64 +29,82 @@ pip install -U sentence-transformers
|
|
55 |
|
56 |
Then you can load this model and run inference.
|
57 |
```python
|
|
|
58 |
from sentence_transformers import SentenceTransformer
|
59 |
|
60 |
# Download from the 🤗 Hub
|
61 |
-
model = SentenceTransformer("
|
62 |
-
|
|
|
63 |
sentences = [
|
64 |
-
|
65 |
-
"
|
66 |
-
|
|
|
67 |
]
|
68 |
-
embeddings = model.encode(sentences)
|
69 |
-
print(embeddings.shape)
|
70 |
-
# [3, 1024]
|
71 |
-
|
72 |
-
# Get the similarity scores for the embeddings
|
73 |
-
similarities = model.similarity(embeddings, embeddings)
|
74 |
-
print(similarities.shape)
|
75 |
-
# [3, 3]
|
76 |
-
```
|
77 |
-
|
78 |
-
<!--
|
79 |
-
### Direct Usage (Transformers)
|
80 |
|
81 |
-
|
|
|
|
|
82 |
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
<!--
|
87 |
-
### Downstream Usage (Sentence Transformers)
|
88 |
-
|
89 |
-
You can finetune this model on your own dataset.
|
90 |
|
91 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
92 |
|
93 |
-
</details>
|
94 |
-
-->
|
95 |
|
96 |
-
<!--
|
97 |
-
### Out-of-Scope Use
|
98 |
|
99 |
-
|
100 |
-
-->
|
101 |
|
102 |
-
|
103 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
104 |
|
105 |
-
|
106 |
-
-->
|
107 |
|
108 |
-
|
109 |
-
|
|
|
|
|
|
|
|
|
110 |
|
111 |
-
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
|
112 |
-
-->
|
113 |
|
114 |
## Training Details
|
115 |
|
|
|
116 |
### Framework Versions
|
117 |
- Python: 3.10.13
|
118 |
- Sentence Transformers: 3.0.0
|
@@ -122,24 +114,10 @@ You can finetune this model on your own dataset.
|
|
122 |
- Datasets: 2.19.1
|
123 |
- Tokenizers: 0.19.1
|
124 |
|
125 |
-
## Citation
|
126 |
|
127 |
### BibTeX
|
|
|
128 |
|
129 |
-
|
130 |
-
|
131 |
-
|
132 |
-
*Clearly define terms in order to be accessible across audiences.*
|
133 |
-
-->
|
134 |
-
|
135 |
-
<!--
|
136 |
-
## Model Card Authors
|
137 |
-
|
138 |
-
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
|
139 |
-
-->
|
140 |
-
|
141 |
-
<!--
|
142 |
-
## Model Card Contact
|
143 |
-
|
144 |
-
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
|
145 |
-
-->
|
|
|
6 |
- sentence-transformers
|
7 |
- sentence-similarity
|
8 |
- feature-extraction
|
9 |
+
base_model: tohoku-nlp/bert-base-japanese-v3
|
10 |
widget: []
|
11 |
pipeline_tag: sentence-similarity
|
12 |
license: apache-2.0
|
13 |
+
datasets:
|
14 |
+
- cl-nagoya/ruri-dataset-ft
|
15 |
---
|
16 |
|
17 |
+
# Ruri: Japanese General Text Embeddings
|
18 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
19 |
|
20 |
## Usage
|
21 |
|
|
|
29 |
|
30 |
Then you can load this model and run inference.
|
31 |
```python
|
32 |
+
import torch.nn.functional as F
|
33 |
from sentence_transformers import SentenceTransformer
|
34 |
|
35 |
# Download from the 🤗 Hub
|
36 |
+
model = SentenceTransformer("cl-nagoya/ruri-pt-base")
|
37 |
+
|
38 |
+
# Don't forget to add the prefix "クエリ: " for query-side or "文章: " for passage-side texts.
|
39 |
sentences = [
|
40 |
+
"クエリ: 瑠璃色はどんな色?",
|
41 |
+
"文章: 瑠璃色(るりいろ)は、紫みを帯びた濃い青。名は、半貴石の瑠璃(ラピスラズリ、英: lapis lazuli)による。JIS慣用色名では「こい紫みの青」(略号 dp-pB)と定義している[1][2]。",
|
42 |
+
"クエリ: ワシやタカのように、鋭いくちばしと爪を持った大型の鳥類を総称して「何類」というでしょう?",
|
43 |
+
"文章: ワシ、タカ、ハゲワシ、ハヤブサ、コンドル、フクロウが代表的である。これらの猛禽類はリンネ前後の時代(17~18世紀)には鷲類・鷹類・隼類及び梟類に分類された。ちなみにリンネは狩りをする鳥を単一の目(もく)にまとめ、vultur(コンドル、ハゲワシ)、falco(ワシ、タカ、ハヤブサなど)、strix(フクロウ)、lanius(モズ)の4属を含めている。",
|
44 |
]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
45 |
|
46 |
+
embeddings = model.encode(sentences, convert_to_tensor=True)
|
47 |
+
print(embeddings.size())
|
48 |
+
# [4, 1024]
|
49 |
|
50 |
+
similarities = F.cosine_similarity(embeddings.unsqueeze(0), embeddings.unsqueeze(1), dim=2)
|
51 |
+
print(similarities)
|
52 |
+
```
|
|
|
|
|
|
|
|
|
53 |
|
54 |
+
## Benchmarks
|
55 |
+
|
56 |
+
### JMTEB
|
57 |
+
Evaluated with [JMTEB](https://github.com/sbintuitions/JMTEB).
|
58 |
+
|
59 |
+
|Model|#Param.|Avg.|Retrieval|STS|Classfification|Reranking|Clustering|PairClassification|
|
60 |
+
|:-:|:-:|:-:|:-:|:-:|:-:|:-:|:-:|:-:|
|
61 |
+
|[cl-nagoya/sup-simcse-ja-base](https://huggingface.co/cl-nagoya/sup-simcse-ja-base)|111M|68.56|49.64|82.05|73.47|91.83|51.79|62.57|
|
62 |
+
|[cl-nagoya/sup-simcse-ja-large](https://huggingface.co/cl-nagoya/sup-simcse-ja-large)|337M|66.51|37.62|83.18|73.73|91.48|50.56|62.51|
|
63 |
+
|[cl-nagoya/unsup-simcse-ja-base](https://huggingface.co/cl-nagoya/unsup-simcse-ja-base)|111M|65.07|40.23|78.72|73.07|91.16|44.77|62.44|
|
64 |
+
|[cl-nagoya/unsup-simcse-ja-large](https://huggingface.co/cl-nagoya/unsup-simcse-ja-large)|337M|66.27|40.53|80.56|74.66|90.95|48.41|62.49|
|
65 |
+
|[pkshatech/GLuCoSE-base-ja](https://huggingface.co/pkshatech/GLuCoSE-base-ja)|133M|70.44|59.02|78.71|76.82|91.90|49.78|66.39|
|
66 |
+
||||||||||
|
67 |
+
|[sentence-transformers/LaBSE](https://huggingface.co/sentence-transformers/LaBSE)|472M|64.70|40.12|76.56|72.66|91.63|44.88|62.33|
|
68 |
+
|[intfloat/multilingual-e5-small](https://huggingface.co/intfloat/multilingual-e5-small)|118M|69.52|67.27|80.07|67.62|93.03|46.91|62.19|
|
69 |
+
|[intfloat/multilingual-e5-base](https://huggingface.co/intfloat/multilingual-e5-base)|278M|70.12|68.21|79.84|69.30|92.85|48.26|62.26|
|
70 |
+
|[intfloat/multilingual-e5-large](https://huggingface.co/intfloat/multilingual-e5-large)|560M|71.65|70.98|79.70|72.89|92.96|51.24|62.15|
|
71 |
+
||||||||||
|
72 |
+
|OpenAI/text-embedding-ada-002|-|69.48|64.38|79.02|69.75|93.04|48.30|62.40|
|
73 |
+
|OpenAI/text-embedding-3-small|-|70.86|66.39|79.46|73.06|92.92|51.06|62.27|
|
74 |
+
|OpenAI/text-embedding-3-large|-|73.97|74.48|82.52|77.58|93.58|53.32|62.35|
|
75 |
+
||||||||||
|
76 |
+
|[Ruri-Small](https://huggingface.co/cl-nagoya/ruri-small)|68M|71.53|69.41|82.79|76.22|93.00|51.19|62.11|
|
77 |
+
|[Ruri-Base](https://huggingface.co/cl-nagoya/ruri-base)|111M|71.91|69.82|82.87|75.58|92.91|54.16|62.38|
|
78 |
+
|[Ruri-Large](https://huggingface.co/cl-nagoya/ruri-large)|337M|73.31|73.02|83.13|77.43|92.99|51.82|62.29|
|
79 |
|
|
|
|
|
80 |
|
|
|
|
|
81 |
|
82 |
+
## Model Details
|
|
|
83 |
|
84 |
+
### Model Description
|
85 |
+
- **Model Type:** Sentence Transformer
|
86 |
+
- **Base model:** [tohoku-nlp/bert-large-japanese-v2](https://huggingface.co/tohoku-nlp/bert-large-japanese-v2)
|
87 |
+
- **Maximum Sequence Length:** 512 tokens
|
88 |
+
- **Output Dimensionality:** 1024
|
89 |
+
- **Similarity Function:** Cosine Similarity
|
90 |
+
- **Language:** Japanese
|
91 |
+
- **License:** Apache 2.0
|
92 |
+
- **Paper:** https://arxiv.org/abs/2409.07737
|
93 |
+
<!-- - **Training Dataset:** Unknown -->
|
94 |
|
95 |
+
### Full Model Architecture
|
|
|
96 |
|
97 |
+
```
|
98 |
+
SentenceTransformer(
|
99 |
+
(0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel
|
100 |
+
(1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
|
101 |
+
)
|
102 |
+
```
|
103 |
|
|
|
|
|
104 |
|
105 |
## Training Details
|
106 |
|
107 |
+
|
108 |
### Framework Versions
|
109 |
- Python: 3.10.13
|
110 |
- Sentence Transformers: 3.0.0
|
|
|
114 |
- Datasets: 2.19.1
|
115 |
- Tokenizers: 0.19.1
|
116 |
|
117 |
+
<!-- ## Citation
|
118 |
|
119 |
### BibTeX
|
120 |
+
-->
|
121 |
|
122 |
+
## License
|
123 |
+
This model is published under the [Apache License, Version 2.0](https://www.apache.org/licenses/LICENSE-2.0).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|