aisingapore/llama-sea-lion-v3.5-70b-r-gguf overview

{
  "metadata": {},
  "card_data": {
    "library_name": "transformers",
    "pipeline_tag": "text-generation",
    "base_model": [
      "aisingapore/Llama-SEA-LION-v3-70B-IT"
    ],
    "language": [
      "en",
      "zh",
      "vi",
      "id",
      "th",
      "fil",
      "ta",
      "ms",
      "km",
      "lo",
      "my",
      "jv",
      "su"
    ],
    "license": "llama3.1",
    "frontmatter": {
      "library_name": "transformers",
      "pipeline_tag": "text-generation",
      "base_model": [
        "aisingapore/Llama-SEA-LION-v3-70B-IT"
      ],
      "language": [
        "en",
        "zh",
        "vi",
        "id",
        "th",
        "fil",
        "ta",
        "ms",
        "km",
        "lo",
        "my",
        "jv",
        "su"
      ],
      "license": "llama3.1"
    },
    "hero_image_url": "llama_sea_lion_3.5_70b_r_banner.png",
    "summary": "**SEA-LION** is a collection of Large Language Models (LLMs) which have been pretrained and instruct-tuned for the Southeast Asia (SEA) region. ### Model Description  SEA-LION stands for *Southeast Asian Languages In One Network*. Quantization was performed on Llama-SEA-LION-v3.5-70B-R to produce optimized variants that reduce memory requirements while maintaining model quality. These quantized models support inference on a range of consumer-grade GPUs and are compatible with various inference engines. For tokenization, the model employs the default tokenizer used in Llama 3.1-70B-Instruct. This repo contains GGUF format models files for aisingapore/Llama-SEA-LION-v3.5-70B-R Model Weights included in this repository: > [!NOTE] > Take note that some GGUFs are split into parts. Most tools such as llama.cpp and those built on it do support split GGUFs, > pointing the platform to the first split will be sufficient for it to function. In the event where a merge is necessary, > it can be done using llama.cpp's gguf-split: ./gguf-split --merge ./path/to/first-split ./path/to/output-gguf More details: > gguf-split guide & README",
    "quick_links": [],
    "benchmark_table_html": "",
    "readme_markdown": "---\nlibrary_name: transformers\npipeline_tag: text-generation\nbase_model:\n- aisingapore/Llama-SEA-LION-v3-70B-IT\nlanguage:\n- en\n- zh\n- vi\n- id\n- th\n- fil\n- ta\n- ms\n- km\n- lo\n- my\n- jv\n- su\nlicense: llama3.1\n---\n\n<div>\n  <img src=\"llama_sea_lion_3.5_70b_r_banner.png\"/>\n</div>\n\nLast updated: 2025-14-04\n\n# Llama-SEA-LION-v3.5-70B-R-GGUF\n\n[**SEA-LION**](https://arxiv.org/abs/2504.05747) is a collection of Large Language Models (LLMs) which have been pretrained and instruct-tuned \nfor the Southeast Asia (SEA) region.\n\n\n\n### Model Description\n\n<!-- Provide a longer summary of what this model is. -->\n\nSEA-LION stands for *Southeast Asian Languages In One Network*. \n\nQuantization was performed on Llama-SEA-LION-v3.5-70B-R to produce optimized variants that reduce memory requirements \nwhile maintaining model quality. These quantized models support inference on a range of consumer-grade GPUs \nand are compatible with various inference engines.\n\n\nFor tokenization, the model employs the default tokenizer used in Llama 3.1-70B-Instruct. \n\n\n- **Developed by:** Products Pillar, AI Singapore\n- **Funded by:** Singapore NRF\n- **Model type:** Decoder\n- **Context length:** 128k tokens\n- **Language(s):** Burmese, Chinese, English, Filipino, Indonesian, Javanese, Khmer, Lao, Malay, Sundanese, Tamil, Thai, Vietnamese\n- **License:** [Llama 3.1 Community License](https://huggingface.co/meta-llama/Llama-3.1-70B-Instruct/blob/main/LICENSE)\n- **Quantized from model:** Llama-SEA-LION-v3.5-70B-R\n\nThis repo contains `GGUF` format models files for [aisingapore/Llama-SEA-LION-v3.5-70B-R](https://huggingface.co/aisingapore/Llama-SEA-LION-v3.5-70B-R)\n\nModel Weights included in this repository:\n- [Llama-SEA-LION-v3.5-70B-R-F16](https://huggingface.co/aisingapore/Llama-SEA-LION-v3.5-70B-R-GGUF/blob/main/Llama-SEA-LION-v3.5-70B-R-F16-00001-of-00008.gguf)\n- [Llama-SEA-LION-v3.5-70B-R-Q2_K](https://huggingface.co/aisingapore/Llama-SEA-LION-v3.5-70B-R-GGUF/blob/main/Llama-SEA-LION-v3.5-70B-R-Q2_K.gguf)\n- [Llama-SEA-LION-v3.5-70B-R-Q3_K_M](https://huggingface.co/aisingapore/Llama-SEA-LION-v3.5-70B-R-GGUF/blob/main/Llama-SEA-LION-v3.5-70B-R-Q3_K_M.gguf)\n- [Llama-SEA-LION-v3.5-70B-R-Q4_0](https://huggingface.co/aisingapore/Llama-SEA-LION-v3.5-70B-R-GGUF/blob/main/Llama-SEA-LION-v3.5-70B-R-Q4_0.gguf)\n- [Llama-SEA-LION-v3.5-70B-R-Q4_K_M](https://huggingface.co/aisingapore/Llama-SEA-LION-v3.5-70B-R-GGUF/blob/main/Llama-SEA-LION-v3.5-70B-R-Q4_K_M.gguf)\n- [Llama-SEA-LION-v3.5-70B-R-Q5_0](https://huggingface.co/aisingapore/Llama-SEA-LION-v3.5-70B-R-GGUF/blob/main/Llama-SEA-LION-v3.5-70B-R-Q5_0.gguf)\n- [Llama-SEA-LION-v3.5-70B-R-Q5_K_M](https://huggingface.co/aisingapore/Llama-SEA-LION-v3.5-70B-R-GGUF/blob/main/Llama-SEA-LION-v3.5-70B-R-Q5_K_M.gguf)\n- [Llama-SEA-LION-v3.5-70B-R-Q6_K](https://huggingface.co/aisingapore/Llama-SEA-LION-v3.5-70B-R-GGUF/blob/main/Llama-SEA-LION-v3.5-70B-R-Q6_K-00001-of-00003.gguf)\n- [Llama-SEA-LION-v3.5-70B-R-Q8_0](https://huggingface.co/aisingapore/Llama-SEA-LION-v3.5-70B-R-GGUF/blob/main/Llama-SEA-LION-v3.5-70B-R-Q8_0-00001-of-00004.gguf)\n\n> [!NOTE]\n> Take note that some GGUFs are split into parts. Most tools such as llama.cpp and those built on it do support split GGUFs, \n> pointing the platform to the first split will be sufficient for it to function. In the event where a merge is necessary, \n> it can be done using llama.cpp's gguf-split: ./gguf-split --merge ./path/to/first-split ./path/to/output-gguf More details: \n> gguf-split guide & [README](https://github.com/ggerganov/llama.cpp/tree/master/examples/gguf-split) \n \n\n## Evaluation\n\n<!-- This section describes the evaluation protocols and provides the results. -->\n\n### Test Results\n\nFor details on Llama-SEA-LION-v3.5-70B-R performance, please refer to the SEA-HELM leaderboard, [Leaderboard results on SEA-HELM](https://leaderboard.sea-lion.ai/).\n\n\n### Out-of-Scope Use\n\n<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->\n\nThe model has not been aligned for safety. Developers and users should perform their own safety \nfine-tuning and related security measures. In no event shall the authors be held liable for any claims, damages, or other liabilities arising from the use of the released weights and codes.\n\n\n## Bias, Risks, and Limitations\n\n<!-- This section is meant to convey both technical and sociotechnical limitations. -->\n\n*The model was not tested for robustness against adversarial prompting.* It is important for users to be aware that our model exhibits certain limitations that warrant consideration. \nLike many LLMs, the model can hallucinate and occasionally generates irrelevant content, \nintroducing fictional elements that are not grounded in the provided context. \nUsers should also exercise caution in interpreting and validating the model's responses \ndue to the potential inconsistencies.\n\n\n\n## More Information\n\nThis is the repository for the commercial instruction-tuned model. \nThe model has not been aligned for safety. Developers and users should perform their own safety \nfine-tuning and related security measures. In no event shall the authors be held liable \nfor any claims, damages, or other liabilities arising from the use of the released weights and codes.\n\n[AI Singapore](https://aisingapore.org/) is a national programme supported by the National Research Foundation, Singapore and hosted by the National University of Singapore. \nAny opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and \ndo not reflect the views of the National Research Foundation or the National University of Singapore.\n\n[Link to SEA-LION's GitHub repository](https://github.com/aisingapore/sealion)\n\nFor more info, please contact us at sealion@aisingapore.org\n\n\n## Team\n\nAntonyrex Sajeban, Chan Adwin, Cheng Nicholas, Choa Esther, Huang Yuli, Hulagadri Adithya Venkatadri, Lau Wayne, Lee Chwan Ren, Leong Wai Yi, Leong Wei Qi, Liew Rachel, Limkonchotiwat Peerat, Liu Bing Jie Darius, \nMontalan Jann Railey, Ng Boon Cheong Raymond, Ngui Jian Gang, Nguyen Thanh Ngan, Ong Brandon, Ong Tat-Wee David, \nOng Zhi Hao, Rengarajan Hamsawardhini, Siow Bryan, Susanto Yosephine, Tai Ngee Chia, Tan Choon Meng, Teng Walter, \nTeo Eng Sipp Leslie, Teo Wei Yi, Tjhi William, Yeo Yeow Tong, Yong Xianbin\n\n\n## Contact\n\nsealion@aisingapore.org",
    "related_quantizations": []
  },
  "tags": [
    "transformers",
    "gguf",
    "text-generation",
    "en",
    "zh",
    "vi",
    "id",
    "th",
    "fil",
    "ta",
    "ms",
    "km",
    "lo",
    "my",
    "jv",
    "su",
    "arxiv:2504.05747",
    "base_model:aisingapore/Llama-SEA-LION-v3-70B-IT",
    "base_model:quantized:aisingapore/Llama-SEA-LION-v3-70B-IT",
    "license:llama3.1",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ],
  "likes": 0,
  "downloads": 217,
  "gated": false,
  "private": false,
  "last_modified": "2025-09-01T10:34:50.000Z",
  "created_at": "2025-04-25T05:55:03.000Z",
  "pipeline_tag": "text-generation",
  "library_name": "transformers"
}

{
  "_id": "680b23b7c131c3be24f1948f",
  "id": "aisingapore/Llama-SEA-LION-v3.5-70B-R-GGUF",
  "modelId": "aisingapore/Llama-SEA-LION-v3.5-70B-R-GGUF",
  "sha": "c8b0318affa79c6e6527ac870b5ed9b5e8835c41",
  "createdAt": "2025-04-25T05:55:03.000Z",
  "lastModified": "2025-09-01T10:34:50.000Z",
  "author": "aisingapore",
  "downloads": 217,
  "likes": 0,
  "gated": false,
  "private": false,
  "pipeline_tag": "text-generation",
  "library_name": "transformers",
  "siblings_count": 24
}