--- tags: - Multilingual license: mit language: - af - am - ar - hy - as - ast - az - be - bn - bs - bg - my - ca - ceb - zho - hr - cs - da - nl - en - et - tl - fi - fr - ff - gl - lg - ka - de - el - gu - ha - he - hi - hu - is - ig - id - ga - it - ja - jv - kea - kam - kn - kk - km - ko - ky - lo - lv - ln - lt - luo - lb - mk - ms - ml - mt - mi - mr - mn - ne - ns - no - ny - oc - or - om - ps - fa - pl - pt - pa - ro - ru - sr - sn - sd - sk - sl - so - ku - es - sw - sv - tg - ta - te - th - tr - uk - umb - ur - uz - vi - cy - wo - xh - yo - zu --- ### Model Sources - **Paper**: LLaMAX: Scaling Linguistic Horizons of LLM by Enhancing Translation Capabilities Beyond 100 Languages - **Link**: https://arxiv.org/pdf/2407.05975 - **Repository**: https://github.com/CONE-MT/LLaMAX/ - **Demo**: https://huggingface.co/spaces/vilarin/LLaMAX3-Translator Thanks for the efforts from @AnnioDance. ### Model Description LLaMAX is a language model with powerful multilingual capabilities without loss instruction-following capabilities. We collected extensive training sets in 102 languages for continued pre-training of Llama2 and leveraged the English instruction fine-tuning dataset, Alpaca, to fine-tune its instruction-following capabilities. ### 🔥 Effortless Multilingual Translation with a Simple Prompt LLaMAX supports translation between more than 100 languages, surpassing the performance of similarly scaled LLMs. ```angular2html def Prompt_template(query, src_language, trg_language): instruction = f'Translate the following sentences from {src_language} to {trg_language}.' prompt = ( 'Below is an instruction that describes a task, paired with an input that provides further context. ' 'Write a response that appropriately completes the request.\n' f'### Instruction:\n{instruction}\n' f'### Input:\n{query}\n### Response:' ) return prompt ``` And then run the following codes to execute translation: ```angular2html from transformers import AutoTokenizer, LlamaForCausalLM model = LlamaForCausalLM.from_pretrained(PATH_TO_CONVERTED_WEIGHTS) tokenizer = AutoTokenizer.from_pretrained(PATH_TO_CONVERTED_TOKENIZER) query = "你好，今天是个好日子" prompt = Prompt_template(query, 'Chinese', 'English') inputs = tokenizer(prompt, return_tensors="pt") generate_ids = model.generate(inputs.input_ids, max_length=30) tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0] # => "Hello, today is a good day" ``` ### 🔥 Excellent Translation Performance LLaMAX3-8B-Alpaca achieves an average spBLEU score improvement of over **5 points** compared to the LLaMA3-8B-Alpaca model on the Flores-101 dataset. | System | Size | en-X (COMET) | en-X (BLEU) | zh-X (COMET)| zh-X (BLEU) | de-X (COMET) | de-X (BLEU) | ne-X (COMET) | ne-X (BLEU) |ar-X (COMET) | ar-X (BLEU) | az-X (COMET) | az-X (BLEU) | ceb-X (COMET) | ceb-X (BLEU)| |--------------------|------|--------------------|-------------| ----| ---- | ---- | ---- | ---- | ---- | ---- | ---- | ---- | ---- | ---- | ---- | | LLaMA3-8B-Alpaca | 8B |67.97|17.23|64.65|10.14|64.67|13.62|62.95|7.96|63.45|11.27|60.61|6.98|55.26|8.52| | LLaMAX3-8B-Alpaca | 8B |75.52|22.77|73.16|14.43|73.47|18.95|75.13|15.32|72.29|16.42|72.06|12.41|68.88|15.85| | System | Size | X-en (COMET) | X-en (BLEU) | X-zh (COMET)| X-zh (BLEU) | X-de (COMET) | X-de (BLEU) | X-ne (COMET) | X-ne (BLEU) |X-ar (COMET) | X-ar (BLEU) | X-az (COMET) | X-az (BLEU) | X-ceb (COMET) | X-ceb (BLEU) | |--------------------|------|----------------|-------------| ----| ---- | ---- | ---- | ---- | ---- | ---- | ---- | ---- | ---- | ---- |--------------| | LLaMA3-8B-Alpaca | 8B |77.43|26.55|73.56|13.17|71.59|16.82|46.56|3.83|66.49|10.20|58.30|4.81|52.68|4.18| | LLaMAX3-8B-Alpaca | 8B |81.28|31.85|78.34|16.46|76.23|20.64|65.83|14.16|75.84|15.45|70.61|9.32|63.35|12.66| ### Supported Languages Akrikaans (af), Amharic (am), Arabic (ar), Armenian (hy), Assamese (as), Asturian (ast), Azerbaijani (az), Belarusian (be), Bengali (bn), Bosnian (bs), Bulgarian (bg), Burmese (my), Catalan (ca), Cebuano (ceb), Chinese Simpl (zho), Chinese Trad (zho), Croatian (hr), Czech (cs), Danish (da), Dutch (nl), English (en), Estonian (et), Filipino (tl), Finnish (fi), French (fr), Fulah (ff), Galician (gl), Ganda (lg), Georgian (ka), German (de), Greek (el), Gujarati (gu), Hausa (ha), Hebrew (he), Hindi (hi), Hungarian (hu), Icelandic (is), Igbo (ig), Indonesian (id), Irish (ga), Italian (it), Japanese (ja), Javanese (jv), Kabuverdianu (kea), Kamba (kam), Kannada (kn), Kazakh (kk), Khmer (km), Korean (ko), Kyrgyz (ky), Lao (lo), Latvian (lv), Lingala (ln), Lithuanian (lt), Luo (luo), Luxembourgish (lb), Macedonian (mk), Malay (ms), Malayalam (ml), Maltese (mt), Maori (mi), Marathi (mr), Mongolian (mn), Nepali (ne), Northern Sotho (ns), Norwegian (no), Nyanja (ny), Occitan (oc), Oriya (or), Oromo (om), Pashto (ps), Persian (fa), Polish (pl), Portuguese (pt), Punjabi (pa), Romanian (ro), Russian (ru), Serbian (sr), Shona (sn), Sindhi (sd), Slovak (sk), Slovenian (sl), Somali (so), Sorani Kurdish (ku), Spanish (es), Swahili (sw), Swedish (sv), Tajik (tg), Tamil (ta), Telugu (te), Thai (th), Turkish (tr), Ukrainian (uk), Umbundu (umb), Urdu (ur), Uzbek (uz), Vietnamese (vi), Welsh (cy), Wolof (wo), Xhosa (xh), Yoruba (yo), Zulu (zu) ### Model Index We implement multiple versions of the LLaMAX model, the model links are as follows: | Model | LLaMAX | LLaMAX-Alpaca | |---------|----------------------------------------------------------|-----------------------------------------------------------------| | Llama-2 | [Link](https://huggingface.co/LLaMAX/LLaMAX2-7B) | [Link](https://huggingface.co/LLaMAX/LLaMAX2-7B-Alpaca) | | Llama-3 | [Link](https://huggingface.co/LLaMAX/LLaMAX3-8B-8B) | [Link](https://huggingface.co/LLaMAX/LLaMAX3-8B-8B-Alpaca) | ### Citation If our model helps your work, please cite this paper: ``` @inproceedings{lu-etal-2024-llamax, title = "{LL}a{MAX}: Scaling Linguistic Horizons of {LLM} by Enhancing Translation Capabilities Beyond 100 Languages", author = "Lu, Yinquan and Zhu, Wenhao and Li, Lei and Qiao, Yu and Yuan, Fei", editor = "Al-Onaizan, Yaser and Bansal, Mohit and Chen, Yun-Nung", booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2024", month = nov, year = "2024", address = "Miami, Florida, USA", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2024.findings-emnlp.631", doi = "10.18653/v1/2024.findings-emnlp.631", pages = "10748--10772", abstract = "Large Language Models (LLMs) demonstrate remarkable translation capabilities in high-resource language tasks, yet their performance in low-resource languages is hindered by insufficient multilingual data during pre-training. To address this, we conduct extensive multilingual continual pre-training on the LLaMA series models, enabling translation support across more than 100 languages. Through a comprehensive analysis of training strategies, such as vocabulary expansion and data augmentation, we develop LLaMAX. Remarkably, without sacrificing its generalization ability, LLaMAX achieves significantly higher translation performance compared to existing open-source LLMs (by more than 10 spBLEU points) and performs on-par with specialized translation model (M2M-100-12B) on the Flores-101 benchmark. Extensive experiments indicate that LLaMAX can serve as a robust multilingual foundation model. The code and the models are publicly available.", } ```