ggml-org/embeddinggemma-300m-gguf overview

{
  "metadata": {},
  "card_data": {
    "base_model": [
      "google/embeddinggemma-300M"
    ],
    "frontmatter": {
      "base_model": [
        "google/embeddinggemma-300M"
      ]
    },
    "hero_image_url": "",
    "summary": "# embeddinggemma-300M GGUF Recommended way to run this model: ``sh llama-server -hf ggml-org/embeddinggemma-300M-GGUF --embeddings ` Then the endpoint can be accessed at http://localhost:8080/embedding, for example using curl: `console curl --request POST \\ --url http://localhost:8080/embedding \\ --header \"Content-Type: application/json\" \\ --data '{\"input\": \"Hello embeddings\"}' \\ --silent ` Alternatively, the llama-embedding command line tool can be used: `sh llama-embedding -hf ggml-org/embeddinggemma-300M-GGUF --verbose-prompt -p \"Hello embeddings\" ` #### embd_normalize When a model uses pooling, or the pooling method is specified using --pooling, the normalization can be controlled by the embd_normalize parameter. The default value is 2 which means that the embeddings are normalized using the Euclidean norm (L2). Other options are: * -1 No normalization *  0 Max absolute *  1 Taxicab *  2 Euclidean/L2 * \\>2 P-Norm This can be passed in the request body to llama-server, for example: `sh --data '{\"input\": \"Hello embeddings\", \"embd_normalize\": -1}' \\ ` And for llama-embedding, by passing --embd-normalize , for example: `sh llama-embedding -hf ggml-org/embeddinggemma-300M-GGUF  --embd-normalize -1 -p \"Hello embeddings\" ``",
    "quick_links": [],
    "benchmark_table_html": "",
    "readme_markdown": "---\nbase_model:\n- google/embeddinggemma-300M\n---\n# embeddinggemma-300M GGUF\n\nRecommended way to run this model:\n\n```sh\nllama-server -hf ggml-org/embeddinggemma-300M-GGUF --embeddings\n```\n\nThen the endpoint can be accessed at http://localhost:8080/embedding, for\nexample using `curl`:\n```console\ncurl --request POST \\\n    --url http://localhost:8080/embedding \\\n    --header \"Content-Type: application/json\" \\\n    --data '{\"input\": \"Hello embeddings\"}' \\\n    --silent\n```\n\nAlternatively, the `llama-embedding` command line tool can be used:\n```sh\nllama-embedding -hf ggml-org/embeddinggemma-300M-GGUF --verbose-prompt -p \"Hello embeddings\"\n```\n\n#### embd_normalize\nWhen a model uses pooling, or the pooling method is specified using `--pooling`,\nthe normalization can be controlled by the `embd_normalize` parameter.\n\nThe default value is `2` which means that the embeddings are normalized using\nthe Euclidean norm (L2). Other options are:\n* -1 No normalization\n*  0 Max absolute\n*  1 Taxicab\n*  2 Euclidean/L2\n* \\>2 P-Norm\n\nThis can be passed in the request body to `llama-server`, for example:\n```sh\n    --data '{\"input\": \"Hello embeddings\", \"embd_normalize\": -1}' \\\n```\n\nAnd for `llama-embedding`, by passing `--embd-normalize <value>`, for example:\n```sh\nllama-embedding -hf ggml-org/embeddinggemma-300M-GGUF  --embd-normalize -1 -p \"Hello embeddings\"\n```\n",
    "related_quantizations": []
  },
  "tags": [
    "gguf",
    "base_model:google/embeddinggemma-300m",
    "base_model:quantized:google/embeddinggemma-300m",
    "endpoints_compatible",
    "region:us"
  ],
  "likes": 26,
  "downloads": 349290,
  "gated": false,
  "private": false,
  "last_modified": "2025-09-04T16:19:45.000Z",
  "created_at": "2025-09-04T16:14:00.000Z",
  "pipeline_tag": "",
  "library_name": ""
}

{
  "_id": "68b9bac89c760f6782d78a57",
  "id": "ggml-org/embeddinggemma-300M-GGUF",
  "modelId": "ggml-org/embeddinggemma-300M-GGUF",
  "sha": "a0f9bb7a85ab3b2b2ff2e540b91930b784518576",
  "createdAt": "2025-09-04T16:14:00.000Z",
  "lastModified": "2025-09-04T16:19:45.000Z",
  "author": "ggml-org",
  "downloads": 349290,
  "likes": 26,
  "gated": false,
  "private": false,
  "pipeline_tag": "",
  "library_name": "",
  "siblings_count": 3
}