cstr/octen-embedding-0.6b-gguf overview

{
  "metadata": {},
  "card_data": {
    "license": "apache-2.0",
    "tags": [
      "sentence-transformers",
      "feature-extraction",
      "gguf",
      "crispembed",
      "qwen3"
    ],
    "language": [
      "multilingual"
    ],
    "base_model": "Octen/Octen-Embedding-0.6B",
    "pipeline_tag": "feature-extraction",
    "frontmatter": {
      "license": "apache-2.0",
      "tags": [
        "sentence-transformers",
        "feature-extraction",
        "gguf",
        "crispembed",
        "qwen3"
      ],
      "language": [
        "multilingual"
      ],
      "base_model": "Octen/Octen-Embedding-0.6B",
      "pipeline_tag": "feature-extraction"
    },
    "hero_image_url": "",
    "summary": "GGUF conversion of Octen/Octen-Embedding-0.6B for use with CrispEmbed.",
    "quick_links": [],
    "benchmark_table_html": "",
    "readme_markdown": "---\nlicense: apache-2.0\ntags:\n  - sentence-transformers\n  - feature-extraction\n  - gguf\n  - crispembed\n  - qwen3\nlanguage:\n  - multilingual\nbase_model: Octen/Octen-Embedding-0.6B\npipeline_tag: feature-extraction\n---\n\n# Octen-Embedding-0.6B — GGUF\n\nGGUF conversion of [Octen/Octen-Embedding-0.6B](https://huggingface.co/Octen/Octen-Embedding-0.6B) for use with [CrispEmbed](https://github.com/CrispStrobe/CrispEmbed).\n\n## Model Details\n\n- **Architecture:** Qwen3 decoder with GQA (16 Q heads, 8 KV heads, head_dim=128)\n- **Parameters:** 0.6B (28 layers, 1024 hidden, 3072 intermediate)\n- **Embedding dim:** 1024\n- **Pooling:** Last-token\n- **Tokenizer:** GPT-2 BPE (151K vocab)\n- **RoPE:** theta=1,000,000\n- **License:** Apache-2.0\n\n## Files\n\n| File | Type | Size | CosSim vs HF |\n|------|------|------|-------------|\n| `octen-0.6b-f32.gguf` | F32 | 2.3 GB | 0.9999 |\n| `octen-0.6b-q8_0.gguf` | Q8_0 | 609 MB | 0.9993 |\n| `octen-0.6b-q4_k.gguf` | Q4_K | 325 MB | 0.9570 |\n\n## Usage with CrispEmbed\n\n```bash\n./crispembed -m octen-0.6b-q8_0.gguf \"Hello world\"\n# prints 1024-dim L2-normalized embedding\n\n# Server mode\n./crispembed-server -m octen-0.6b-q8_0.gguf --port 8080\ncurl -X POST http://localhost:8080/embed -d '{\"texts\": [\"Hello world\"]}'\n```\n\n## Conversion\n\nConverted from the original PyTorch model using `models/convert-decoder-embed-to-gguf.py` from the CrispEmbed repo. Verified bit-identical (cos≥0.999) to HuggingFace sentence-transformers output.\n",
    "related_quantizations": []
  },
  "tags": [
    "sentence-transformers",
    "gguf",
    "feature-extraction",
    "crispembed",
    "qwen3",
    "multilingual",
    "base_model:Octen/Octen-Embedding-0.6B",
    "base_model:quantized:Octen/Octen-Embedding-0.6B",
    "license:apache-2.0",
    "endpoints_compatible",
    "region:us"
  ],
  "likes": 0,
  "downloads": 103,
  "gated": false,
  "private": false,
  "last_modified": "2026-04-14T19:10:30.000Z",
  "created_at": "2026-04-14T19:09:11.000Z",
  "pipeline_tag": "feature-extraction",
  "library_name": "sentence-transformers"
}

{
  "_id": "69de90d715af821c90d21425",
  "id": "cstr/Octen-Embedding-0.6B-GGUF",
  "modelId": "cstr/Octen-Embedding-0.6B-GGUF",
  "sha": "a8cbc36adf7614b099eab7300fefde4ce7eba7f5",
  "createdAt": "2026-04-14T19:09:11.000Z",
  "lastModified": "2026-04-14T19:10:30.000Z",
  "author": "cstr",
  "downloads": 103,
  "likes": 0,
  "gated": false,
  "private": false,
  "pipeline_tag": "feature-extraction",
  "library_name": "sentence-transformers",
  "siblings_count": 5
}