2512.04810v3.pdf

{
  "dorsal/arxiv": [
    {
      "date_modified": "2026-02-17T05:42:17.767000Z",
      "hash": "cf1c458832449c590d1601edc88ad8dd6a64d543fe1dcdcc507945fcda7571db",
      "id": "9f8ffbcf-c872-44db-8f5b-db95fce747f1",
      "source": {
        "id": "arXiv Dataset",
        "type": "Model",
        "variant": "snapshot-2026-01-17",
        "version": "0.1.0"
      },
      "user_id": 1000002
    }
  ],
  "file/base": {
    "record": {
      "extension": ".pdf",
      "hash": "cf1c458832449c590d1601edc88ad8dd6a64d543fe1dcdcc507945fcda7571db",
      "media_type": "application/pdf",
      "media_type_prefix": "application",
      "name": "2512.04810v3.pdf",
      "size": 9871217
    },
    "source": {
      "id": "file/base",
      "type": "Model",
      "version": "1.0.0"
    }
  },
  "file/pdf": {
    "private": false,
    "record": {
      "keywords": [],
      "page_count": 18,
      "producer": "pikepdf 8.15.1",
      "title": "EMMA: Efficient Multimodal Understanding, Generation, and Editing with a Unified Architecture",
      "version": "1.7"
    },
    "source": {
      "id": "dorsal/pdf",
      "type": "Model",
      "version": "1.1.0"
    }
  }
}