2601.06843v1.pdf

{
  "dorsal/arxiv": [
    {
      "date_modified": "2026-02-17T05:53:08.860000Z",
      "hash": "53bdd9f7537ab8db3b9b266a4e0b9a1d52d7dc9293d6846f04e217b9bbf8e5cc",
      "id": "fbaccc69-b080-45fb-be88-b6e198d549a0",
      "source": {
        "execution_id": "ca6a70aa-197f-4032-a2ef-8d9365df3397",
        "id": "arXiv Dataset",
        "type": "Model",
        "variant": "snapshot-2026-01-17",
        "version": "0.1.0"
      },
      "user_id": 1000002
    }
  ],
  "file/base": {
    "record": {
      "extension": ".pdf",
      "hash": "53bdd9f7537ab8db3b9b266a4e0b9a1d52d7dc9293d6846f04e217b9bbf8e5cc",
      "media_type": "application/pdf",
      "media_type_prefix": "application",
      "name": "2601.06843v1.pdf",
      "size": 2064835
    },
    "source": {
      "execution_id": "becf6817-9793-47e2-a9a7-f873fed0492b",
      "id": "file/base",
      "type": "Model",
      "version": "1.0.0"
    }
  },
  "file/pdf": {
    "private": false,
    "record": {
      "keywords": [],
      "page_count": 15,
      "producer": "pikepdf 8.15.1",
      "title": "Speak While Watching: Unleashing TRUE Real-Time Video Understanding Capability of Multimodal Large Language Models",
      "version": "1.7"
    },
    "source": {
      "execution_id": "9fde8863-0bce-4b0b-9100-7900c0256001",
      "id": "dorsal/pdf",
      "type": "Model",
      "version": "1.1.0"
    }
  }
}