Instructions to use OptimaMT/de-en with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers.js
How to use OptimaMT/de-en with Transformers.js:
// npm i @huggingface/transformers import { pipeline } from '@huggingface/transformers'; // Allocate pipeline const pipe = await pipeline('translation', 'OptimaMT/de-en');
|
Download README.md from OptimaMT/de-en: direct link, hf CLI and curl.
- Browser
- Download file 2.53 kB
-
https://huggingface.co/OptimaMT/de-en/resolve/main/README.md
- Command line
-
hf download hf://OptimaMT/de-en/README.md
-
curl -L -o README.md https://huggingface.co/OptimaMT/de-en/resolve/main/README.md
2.53 kB
| license: cc-by-4.0 | |
| language: | |
| - de | |
| - en | |
| pipeline_tag: translation | |
| library_name: transformers.js | |
| base_model: Xenova/opus-mt-de-en | |
| tags: | |
| - translation | |
| - marian | |
| - opus-mt | |
| - onnx | |
| # OptimaMT de → en (German to English) | |
| Drop-in copy of [Xenova/opus-mt-de-en](https://huggingface.co/Xenova/opus-mt-de-en) at revision `399dfd687067` 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` | 495 → 268 MiB (-46 %) | 474 → 245 MiB (-48 %) | yes | | |
| | ONNX Runtime defaults | 265 → 208 MiB (-21 %) | 190 → 158 MiB (-17 %) | yes | | |
| FLORES-200 devtest (1,012 sentences): the same translations as the original, BLEU 40.04, chrF 66.18 (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/de-en", { | |
| dtype: "q8", | |
| session_options: { extra: { session: { disable_quant_qdq: "1" } } }, | |
| }); | |
| const [out] = await translator("Der Leuchtturm steht an der Küste."); | |
| 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/de-en "Der Leuchtturm steht an der Küste." | |
| ``` | |
| ## Files | |
| Only `onnx/decoder_model_merged_quantized.onnx` is rebuilt; every other file is the original, byte for byte. | |
| Vocabulary: 58,101 tokens. | |
| ## Licence and credits | |
| - Weights: [OPUS-MT](https://huggingface.co/Helsinki-NLP/opus-mt-de-en) by the University of Helsinki, CC BY 4.0. | |
| - ONNX export: [Xenova/opus-mt-de-en](https://huggingface.co/Xenova/opus-mt-de-en) at revision `399dfd687067`. | |
| - Modified: the decoder's output projection was moved out of its two branches (graph change only). CC BY 4.0. | |