File size: 1,259 Bytes
d0c9db0 |
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 |
---
license: apache-2.0
tags:
- translation
- opus-mt
- ctranslate2
- quantized
language:
- multilingual
pipeline_tag: translation
---
# opus-mt-zls-zls-ctranslate2-android
This is a quantized INT8 version of `Helsinki-NLP/opus-mt-zls-zls` converted to CTranslate2 format for efficient inference.
## Model Details
- **Original Model**: Helsinki-NLP/opus-mt-zls-zls
- **Format**: CTranslate2
- **Quantization**: INT8
- **Framework**: OPUS-MT
- **Converted by**: Automated conversion pipeline
## Usage
### With CTranslate2
```python
import ctranslate2
import sentencepiece as spm
# Load the model
translator = ctranslate2.Translator("path/to/model")
# Load tokenizers
sp_source = spm.SentencePieceProcessor(model_file="source.spm")
sp_target = spm.SentencePieceProcessor(model_file="target.spm")
# Translate
source_tokens = sp_source.encode("Your text here", out_type=str)
results = translator.translate_batch([source_tokens])
translation = sp_target.decode(results[0].hypotheses[0])
```
## Performance
This INT8 quantized version provides:
- ~75% reduction in model size
- Faster inference speed
- Maintained translation quality
- Mobile-friendly deployment
## Original Model
Based on the OPUS-MT project: https://github.com/Helsinki-NLP/Opus-MT
|