Hugging Face's logo

michaelfeil
/
ct2fast-opus-mt-fr-en

Translation
Transformers
ctranslate2
Inference Endpoints
Model card Files Community
ct2fast-opus-mt-fr-en
File size: 2,862 Bytes 
cb3b2d6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
 
---
tags:
- ctranslate2
- translation
license: apache-2.0
---
# # Fast-Inference with Ctranslate2
Speedup inference by 2x-8x using int8 inference in C++

quantized version of [Helsinki-NLP/opus-mt-fr-en](https://huggingface.co/Helsinki-NLP/opus-mt-fr-en)
```bash
pip install hf-hub-ctranslate2>=1.0.0 ctranslate2>=3.13.0
```
Converted using
```
ct2-transformers-converter --model Helsinki-NLP/opus-mt-fr-en --output_dir /home/michael/tmp-ct2fast-opus-mt-fr-en --force --copy_files README.md generation_config.json tokenizer_config.json vocab.json source.spm .gitattributes target.spm --quantization float16
```

Checkpoint compatible to [ctranslate2](https://github.com/OpenNMT/CTranslate2) and [hf-hub-ctranslate2](https://github.com/michaelfeil/hf-hub-ctranslate2)
- `compute_type=int8_float16` for `device="cuda"` 
- `compute_type=int8`  for `device="cpu"`

```python
from hf_hub_ctranslate2 import TranslatorCT2fromHfHub, GeneratorCT2fromHfHub
from transformers import AutoTokenizer

model_name = "michaelfeil/ct2fast-opus-mt-fr-en"
# use either TranslatorCT2fromHfHub or GeneratorCT2fromHfHub here, depending on model.
model = TranslatorCT2fromHfHub(
        # load in int8 on CUDA
        model_name_or_path=model_name, 
        device="cuda",
        compute_type="int8_float16",
        tokenizer=AutoTokenizer.from_pretrained("Helsinki-NLP/opus-mt-fr-en")
)
outputs = model.generate(
    text=["How do you call a fast Flan-ingo?", "User: How are you doing?"],
)
print(outputs)
```

# Licence and other remarks:
This is just a quantized version. Licence conditions are intended to be idential to original huggingface repo.

# Original description
    

### opus-mt-fr-en

* source languages: fr
* target languages: en
*  OPUS readme: [fr-en](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/fr-en/README.md)

*  dataset: opus
* model: transformer-align
* pre-processing: normalization + SentencePiece
* download original weights: [opus-2020-02-26.zip](https://object.pouta.csc.fi/OPUS-MT-models/fr-en/opus-2020-02-26.zip)
* test set translations: [opus-2020-02-26.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/fr-en/opus-2020-02-26.test.txt)
* test set scores: [opus-2020-02-26.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/fr-en/opus-2020-02-26.eval.txt)

## Benchmarks

| testset               | BLEU  | chr-F |
|-----------------------|-------|-------|
| newsdiscussdev2015-enfr.fr.en 	| 33.1 	| 0.580 |
| newsdiscusstest2015-enfr.fr.en 	| 38.7 	| 0.614 |
| newssyscomb2009.fr.en 	| 30.3 	| 0.569 |
| news-test2008.fr.en 	| 26.2 	| 0.542 |
| newstest2009.fr.en 	| 30.2 	| 0.570 |
| newstest2010.fr.en 	| 32.2 	| 0.590 |
| newstest2011.fr.en 	| 33.0 	| 0.597 |
| newstest2012.fr.en 	| 32.8 	| 0.591 |
| newstest2013.fr.en 	| 33.9 	| 0.591 |
| newstest2014-fren.fr.en 	| 37.8 	| 0.633 |
| Tatoeba.fr.en 	| 57.5 	| 0.720 |