File size: 4,986 Bytes
ae4fce8
 
 
2209e43
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
---
license: cc-by-nc-4.0
---

# SONAR
[[Paper]]()
[[Demo]](#usage)

We introduce SONAR, a new multilingual and multimodal fixed-size sentence embedding space. Our single **text encoder, covering 200 languages**, substantially outperforms existing sentence embeddings such as LASER3 and LabSE on the xsim and xsim++ multilingual similarity search tasks. 

Speech segments can be embedded in the same \sonar embedding space using language-specific speech encoders trained in a teacher-student setting on speech transcription data. Our encoders outperform existing speech encoders on similarity search tasks.
We also provide a **text decoder for 200 languages**, which allows us to perform text-to-text and speech-to-text machine translation, including for zero-shot language and modality combinations. 

Our text-to-text results are competitive compared to the state-of-the-art NLLB~1B model, despite the fixed-size bottleneck representation. Our zero-shot speech-to-text translation results compare favorably with strong supervised baselines such as Whisper.


Model inference support thanks [Fairseq2](https://github.com/facebookresearch/fairseq2)


## Installing

See our github [repo](https://reimagined-broccoli-941276ee.pages.github.io/nightly/installation/from_source_conda)

## Usage
Compute text sentence embeddings:
```python
from sonar.inference_pipelines.text import TextToEmbeddingModelPipeline
t2vec_model = TextToEmbeddingModelPipeline(encoder="text_sonar_basic_encoder",
                                           tokenizer="text_sonar_basic_encoder")
sentences = ['My name is SONAR.', 'I can embed the sentences into vectorial space.']
t2vec_model.predict(sentences, source_lang="eng_Latn").shape
# torch.Size([2, 1024])
```

Translate with SONAR
```python
from sonar.inference_pipelines.text import TextToTextModelPipeline
t2t_model = TextToTextModelPipeline(encoder="text_sonar_basic_encoder",
                                    decoder="text_sonar_basic_decoder",
                                    tokenizer="text_sonar_basic_encoder")  # tokenizer is attached to both encoder and decoder cards

sentences = ['My name is SONAR.', 'I can embed the sentences into vectorial space.']
t2t_model.predict(sentences, source_lang="eng_Latn", target_lang="fra_Latn")
# ['Mon nom est SONAR.', "Je peux intégrer les phrases dans l'espace vectoriel."]
```

Compute speech sentence embeddings:
```python
import torch
from sonar.inference_pipelines.speech import SpeechToEmbeddingPipeline, SpeechInferenceParams

speech_embedding_dp_builder = SpeechToEmbeddingPipeline.load_from_name("sonar_speech_encoder_eng")

speech_ctx = SpeechInferenceParams(
    data_file="..../test_fleurs_fra-eng.tsv",
    audio_root_dir=".../audio_zips",
    audio_path_index=2,
    batch_size=4,
)

speech_embedding_dp = speech_embedding_dp_builder.build_pipeline(speech_ctx)
with torch.inference_mode():
    speech_emb = next(iter(speech_embedding_dp))
speech_emb["audio"]["data"].sentence_embeddings
```


Speech-to-text with SONAR
```python
import torch
from sonar.inference_pipelines import SpeechToTextPipeline, SpeechInferenceParams

speech_to_text_dp_builder = SpeechToTextPipeline.load_from_name(encoder_name="sonar_speech_encoder_eng", 
                                                                decoder_name="text_sonar_basic_decoder")

speech_ctx = SpeechInferenceParams(
    data_file=".../test_fleurs_fra-eng.tsv",
    audio_root_dir=".../audio_zips",
    audio_path_index=2,
    target_lang='fra_Latn',
    batch_size=4,
)
speech_to_text_dp = speech_to_text_dp_builder.build_pipeline(speech_ctx)
with torch.inference_mode():
    speech_text_translation = next(iter(speech_to_text_dp))
speech_text_translation
```

Predicting [cross-lingual semantic similarity](https://github.com/facebookresearch/fairseq/tree/nllb/examples/nllb/human_XSTS_eval) 
with BLASER-2 models
```Python
import torch
from sonar.models.blaser.loader import load_blaser_model

blaser_ref = load_blaser_model("blaser_st2st_ref_v2_0").eval()
blaser_qe = load_blaser_model("blaser_st2st_qe_v2_0").eval()
# BLASER-2 is supposed to work with SONAR speech and text embeddings,
# but we didn't include their extraction in this snippet, to keep it simple.
emb = torch.ones([1, 1024])
print(blaser_ref(src=emb, ref=emb, mt=emb).item())  # 5.2552
print(blaser_qe(src=emb, mt=emb).item())  # 4.9819
```

See more complete demo notebooks :
* [sonar text2text similarity and translation](examples/sonar_text_demo.ipynb)
* [sonar speech2text and other data pipeline examples](examples/inference_pipelines.ipynb)


## Model details

- **Developed by:** Paul-Ambroise Duquenne et al.
- **License:** CC-BY-NC 4.0 license
- **Cite as:**

  @article{Duquenne:2023:sonar_arxiv,
    author = {Paul-Ambroise Duquenne and Holger Schwenk and Benoit Sagot},
    title = {{SONAR:} Sentence-Level Multimodal and Language-Agnostic Representations},
    publisher = {arXiv},
    year = {2023},
    url = {https://arxiv.org/abs/unk},
  }