Instructions to use OptimaMT/en-vi with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers.js
How to use OptimaMT/en-vi with Transformers.js:
// npm i @huggingface/transformers import { pipeline } from '@huggingface/transformers'; // Allocate pipeline const pipe = await pipeline('translation', 'OptimaMT/en-vi');
File size: 2,536 Bytes
add3e13 7492c08 77a9f2e 7492c08 add3e13 7492c08 add3e13 | 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 | ---
license: cc-by-4.0
language:
- en
- vi
pipeline_tag: translation
library_name: transformers.js
base_model: Xenova/opus-mt-en-vi
tags:
- translation
- marian
- opus-mt
- onnx
---
# OptimaMT en → vi (English to Vietnamese)
Drop-in copy of [Xenova/opus-mt-en-vi](https://huggingface.co/Xenova/opus-mt-en-vi) at revision `30bcd461c0d7` with a lighter merged decoder: ONNX Runtime's fast mode (`disable_quant_qdq = 1`, 2.3–2.5× faster) at half the memory. Same weights, same tokens.
Part of [OptimaMT](https://github.com/AmerAlreyahi/OptimaMT): [all 51 pairs](https://huggingface.co/collections/OptimaMT/optimamt-lighter-opus-mt-onnx-models-6af228dc9cd91f015e56e869), named `OptimaMT/<src>-<tgt>`.
**Used in production:**
- [OpenCircle](https://github.com/AmerAlreyahi/OpenCircle), a desktop screen translator for Windows and Linux, ships these models since v0.1.4.
| Memory after load | Linux | Windows | Identical tokens |
|---|---:|---:|---|
| `disable_quant_qdq = 1` | 464 → 257 MiB (-45 %) | 441 → 232 MiB (-47 %) | yes |
| ONNX Runtime defaults | 251 → 198 MiB (-21 %) | 179 → 155 MiB (-14 %) | yes |
FLORES-200 devtest (1,012 sentences): the same translations as the original, BLEU 24.5, chrF 43.68 (greedy).
ONNX Runtime 1.31.1, CPU, 4 threads (Linux; Windows in a VM). Interactive results: [OptimaMT/results](https://huggingface.co/spaces/OptimaMT/results) ·
method: [REPORT.md](https://github.com/AmerAlreyahi/OptimaMT/blob/main/REPORT.md).
## Use
```js
import { pipeline } from "@huggingface/transformers";
const translator = await pipeline("translation", "OptimaMT/en-vi", {
dtype: "q8",
session_options: { extra: { session: { disable_quant_qdq: "1" } } },
});
const [out] = await translator("The lighthouse stands on the coast.");
console.log(out.translation_text);
```
Python ([translate.py](https://github.com/AmerAlreyahi/OptimaMT/blob/main/scripts/translate.py): onnxruntime, sentencepiece, numpy):
```bash
python translate.py OptimaMT/en-vi "The lighthouse stands on the coast."
```
## Files
Only `onnx/decoder_model_merged_quantized.onnx` is rebuilt; every other file is the original, byte for byte.
Vocabulary: 53,685 tokens.
## Licence and credits
- Weights: [OPUS-MT](https://huggingface.co/Helsinki-NLP/opus-mt-en-vi) by the University of Helsinki, CC BY 4.0.
- ONNX export: [Xenova/opus-mt-en-vi](https://huggingface.co/Xenova/opus-mt-en-vi) at revision `30bcd461c0d7`.
- Modified: the decoder's output projection was moved out of its two branches (graph change only). CC BY 4.0.
|