qwp4w3hyb/meta-llama-3.1-70b-instruct-imat-gguf overview

{
  "metadata": {},
  "card_data": {
    "language": [
      "en",
      "de",
      "fr",
      "it",
      "pt",
      "hi",
      "es",
      "th"
    ],
    "license": "llama3.1",
    "pipeline_tag": "text-generation",
    "tags": [
      "facebook",
      "meta",
      "pytorch",
      "llama",
      "llama-3",
      "gguf",
      "imatrix"
    ],
    "base_model": "meta-llama/Meta-Llama-3.1-70B-Instruct",
    "frontmatter": {
      "language": [
        "en",
        "de",
        "fr",
        "it",
        "pt",
        "hi",
        "es",
        "th"
      ],
      "license": "llama3.1",
      "pipeline_tag": "text-generation",
      "tags": [
        "facebook",
        "meta",
        "pytorch",
        "llama",
        "llama-3",
        "gguf",
        "imatrix"
      ],
      "base_model": "meta-llama/Meta-Llama-3.1-70B-Instruct"
    },
    "hero_image_url": "",
    "summary": "`` ./imatrix -m $model_name-bf16.gguf -f calibration_datav3.txt -o $model_name.imatrix `` # Original Model Card: TODO",
    "quick_links": [],
    "benchmark_table_html": "",
    "readme_markdown": "---\nlanguage:\n  - en\n  - de\n  - fr\n  - it\n  - pt\n  - hi\n  - es\n  - th\nlicense: llama3.1\npipeline_tag: text-generation\ntags:\n  - facebook\n  - meta\n  - pytorch\n  - llama\n  - llama-3\n  - gguf\n  - imatrix\nbase_model: meta-llama/Meta-Llama-3.1-70B-Instruct\n---\n\n# Quant Infos\n\n- ~Requires latest master + [Rope Scaling PR](https://github.com/ggerganov/llama.cpp/pull/8676).~ Rope scaling is merged, so just a a recent master is required now.\n- quants done with an importance matrix for improved quantization loss\n- Quantized ggufs & imatrix from hf bf16, through bf16. `safetensors bf16 -> gguf bf16 -> quant` for *optimal* quant loss.\n- Wide coverage of different gguf quant types from Q\\_8\\_0 down to IQ1\\_S\n  - experimental custom quant types\n    - `_L` with `--output-tensor-type f16 --token-embedding-type f16`, which supposedly leads to better accuracy.\n- Imatrix generated with [this](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8) multi-purpose dataset by [bartowski](https://huggingface.co/bartowski).\n  ```\n  ./imatrix -m $model_name-bf16.gguf -f calibration_datav3.txt -o $model_name.imatrix\n  ```\n\n# Original Model Card:\n\nTODO",
    "related_quantizations": []
  },
  "tags": [
    "gguf",
    "facebook",
    "meta",
    "pytorch",
    "llama",
    "llama-3",
    "imatrix",
    "text-generation",
    "en",
    "de",
    "fr",
    "it",
    "pt",
    "hi",
    "es",
    "th",
    "base_model:meta-llama/Llama-3.1-70B-Instruct",
    "base_model:quantized:meta-llama/Llama-3.1-70B-Instruct",
    "license:llama3.1",
    "endpoints_compatible",
    "region:us",
    "conversational"
  ],
  "likes": 8,
  "downloads": 247,
  "gated": false,
  "private": false,
  "last_modified": "2024-08-06T09:10:09.000Z",
  "created_at": "2024-07-25T12:18:38.000Z",
  "pipeline_tag": "text-generation",
  "library_name": ""
}

{
  "_id": "66a2429e614dd62973aa9f39",
  "id": "qwp4w3hyb/Meta-Llama-3.1-70B-Instruct-iMat-GGUF",
  "modelId": "qwp4w3hyb/Meta-Llama-3.1-70B-Instruct-iMat-GGUF",
  "sha": "6a77687765d18011e4778a544cd94a18381955b7",
  "createdAt": "2024-07-25T12:18:38.000Z",
  "lastModified": "2024-08-06T09:10:09.000Z",
  "author": "qwp4w3hyb",
  "downloads": 247,
  "likes": 8,
  "gated": false,
  "private": false,
  "pipeline_tag": "text-generation",
  "library_name": "",
  "siblings_count": 27
}