--- license: cc-by-4.0 base_model: HPLT/translate-ko-en-v2.0-hplt_opus library_name: ctranslate2 pipeline_tag: translation tags: [translation, marian, ctranslate2, int8, ko, en] --- # translate-windy-hplt-ko-en — ko to en Bilingual machine translation in **CTranslate2 INT8** for fast CPU inference. Fine-tuned by **Windstorm Labs** from [`HPLT/translate-ko-en-v2.0-hplt_opus`](https://huggingface.co/HPLT/translate-ko-en-v2.0-hplt_opus). ## Attribution Derived from [`HPLT/translate-ko-en-v2.0-hplt_opus`](https://huggingface.co/HPLT/translate-ko-en-v2.0-hplt_opus), from the **HPLT** project, licensed under **CC-BY-4.0**. **Modified by Windstorm Labs.** Attribution is retained as the licence requires. ## What we did Converted from the upstream Marian release, LoRA fine-tuned on parallel data for this language pair, merged into the base weights, then quantized to INT8. | | | |---|---| | Rank / alpha | 8 / 16 | | Learning rate / steps | 2.5e-06 / 50 | | Target modules | `q_proj`, `v_proj` | | Seed | 42 | | Training data | opus100_ko-en (clinic cache), 2,000 sentence pairs | | Tensors modified | 36 | ## Provenance ``` base model.bin sha256 3b32451b3a44d148ae309300bf269f15d17aef0765dc19bd81d499825ba51a6d this model.bin sha256 66515ad86368a0fce3f148f45b0c7f9e6b436099813a57ead5e6b40e764b3b34 ``` Verified after INT8 quantization, on the published artifact. ## Evaluation FLORES-200 devtest, 200 sentences, beam 4, CTranslate2 INT8 on CPU. | | spBLEU | chrF | |---|---:|---:| | this model | **28.25** | 54.19 | ## Notes - Single language pair: ko to en. - FLORES-200 is news and encyclopedic prose. - No human evaluation was performed.