david-ar/20q-gguf overview

{
  "metadata": {},
  "card_data": {
    "license": "mit",
    "base_model": "david-ar/20q",
    "tags": [
      "gguf",
      "2-bit",
      "quantized",
      "twenty-questions"
    ],
    "frontmatter": {
      "license": "mit",
      "base_model": "david-ar/20q",
      "tags": [
        "gguf",
        "2-bit",
        "quantized",
        "twenty-questions"
      ]
    },
    "hero_image_url": "",
    "summary": "GGUF quantized versions of david-ar/20q, the world's smallest chat model. This model was **natively trained at 2-bit precision**. All quantization levels above Q2_K are technically upscaled. Q2_K is the model's native precision.",
    "quick_links": [],
    "benchmark_table_html": "",
    "readme_markdown": "---\nlicense: mit\nbase_model: david-ar/20q\ntags:\n  - gguf\n  - 2-bit\n  - quantized\n  - twenty-questions\n---\n\n# TwentyQ — GGUF\n\nGGUF quantized versions of [david-ar/20q](https://huggingface.co/david-ar/20q), the world's smallest chat model.\n\nThis model was **natively trained at 2-bit precision**. All quantization levels above Q2_K are technically upscaled. Q2_K is the model's native precision.\n\n## Available Quantizations\n\n| File | Quant | Size | Quality Loss |\n|---|---|---|---|\n| twentyq-f32.gguf | F32 | 762 KB | 0% |\n| twentyq-f16.gguf | F16 | 397 KB | 0% |\n| twentyq-q8_0.gguf | Q8_0 | 228 KB | 0% |\n| twentyq-q4_0.gguf | Q4_0 | 135 KB | 0% |\n| twentyq-q2_k.gguf | Q2_K | 95 KB | 0% |\n\nAll quantizations are **lossless** because the original weights are 2-bit integers (values 0-3). Q2_K is the only quantization level that doesn't waste bits.\n\n## Architecture\n\n```\ngeneral.architecture: twentyq\ntwentyq.block_count: 0\ntwentyq.embedding_length: 156\ntwentyq.attention.head_count: 156\ntwentyq.context_length: 20\ntwentyq.vocab_size: 1200\n```\n\nZero transformer blocks. 156 attention heads. 20-token context window. The output projection layer (`output.weight`) contains the entire model.\n\n## Compatibility\n\nThese files require a runtime with `twentyq` architecture support, which does not currently exist in llama.cpp, ollama, or any other GGUF runtime. For inference, use the [original model](https://huggingface.co/david-ar/20q) via the `transformers` library, or the [live demo](https://huggingface.co/spaces/david-ar/20q).\n",
    "related_quantizations": []
  },
  "tags": [
    "gguf",
    "2-bit",
    "quantized",
    "twenty-questions",
    "base_model:david-ar/20q",
    "base_model:quantized:david-ar/20q",
    "license:mit",
    "endpoints_compatible",
    "region:us"
  ],
  "likes": 0,
  "downloads": 84,
  "gated": false,
  "private": false,
  "last_modified": "2026-03-20T11:06:46.000Z",
  "created_at": "2026-03-20T11:06:37.000Z",
  "pipeline_tag": "",
  "library_name": ""
}

{
  "_id": "69bd2a3d7ce15b06bbec244f",
  "id": "david-ar/20q-GGUF",
  "modelId": "david-ar/20q-GGUF",
  "sha": "ad861a2d75bf98b6b47a5d98043b985d78eb09e8",
  "createdAt": "2026-03-20T11:06:37.000Z",
  "lastModified": "2026-03-20T11:06:46.000Z",
  "author": "david-ar",
  "downloads": 84,
  "likes": 0,
  "gated": false,
  "private": false,
  "pipeline_tag": "",
  "library_name": "",
  "siblings_count": 7
}