{
  "id": "internvideo2-stage2-6b",
  "identity": {
    "family": "internvideo2",
    "display_name": "InternVideo2 Stage2 6B",
    "hub_id": "OpenGVLab/InternVideo2-Stage2_6B",
    "api_name": null,
    "revision": null,
    "variants": [
      "OpenGVLab/InternVideo2-Stage2_6B-224p-f4 (Apache-2.0, .pt with audio encoder)"
    ],
    "license": "MIT",
    "commercial_use": true,
    "commercial_use_note": "MIT per the hub tag.",
    "release_date": "2025-02",
    "paper_url": "https://arxiv.org/abs/2403.15377",
    "repo_url": "https://github.com/OpenGVLab/InternVideo"
  },
  "task": {
    "category": "embedding",
    "modalities_in": [
      "video",
      "text"
    ],
    "modalities_out": [
      "vector"
    ],
    "best_at": "State-of-the-art video-text retrieval and action recognition among open video encoders (per the paper).",
    "worst_at": "6.4B parameters: multi-GPU-class memory; far too large for scale-to-zero CPU."
  },
  "io": {
    "input": {
      "video": {
        "frames": "4 to 8 per clip",
        "resolution": "224x224"
      },
      "text": {}
    },
    "output": {
      "vector": {
        "dims": "not verified here",
        "distance": "cosine",
        "type": "dense"
      }
    }
  },
  "mixpeek_wiring": {
    "extractors": [],
    "feature_uris": [],
    "vector_indexes": [],
    "retriever_stages": [],
    "composes_with": [
      "scene splits (transcoding/multimodal extractor) -> clip-level vectors -> feature_search with a text query"
    ]
  },
  "run": {
    "status": "not_runnable",
    "status_reason": "Requires trust_remote_code (custom_code on the hub) and a large GPU. Would run as a GPU batch plugin only after vendoring the model code.",
    "serve_apps": [],
    "runtime": "transformers AutoModel with trust_remote_code=True",
    "weights_format": "safetensors",
    "trust_remote_code": true,
    "size_on_disk_mb": 25000,
    "params_millions": 6367,
    "preprocessing": "Sample frames, resize to 224, normalize.",
    "postprocessing": "L2-normalize projections.",
    "serving_shape": "batch",
    "batch_size": null,
    "serve_batch": null,
    "scale_to_zero": {
      "min_replicas": 0,
      "max_replicas": 1,
      "idle_timeout_s": 600,
      "memory_gb": 40,
      "num_cpus": 8,
      "num_gpus": 1,
      "cold_start_s": {
        "warm_node_s": null,
        "new_node_s": null,
        "source": null
      },
      "cpu_fallback": false
    }
  },
  "cost": {
    "unmeasured": true,
    "measured": null,
    "note": "No committed measurement; plan row 3 measures it in the standalone image."
  },
  "evidence": {
    "unmeasured": true,
    "items": []
  },
  "refs": [
    "hf://OpenGVLab/InternVideo2-Stage2_6B@fd8ef758daf132d4738325d39546dd76ce26911d"
  ],
  "formats": [
    {
      "ref": "hf://OpenGVLab/InternVideo2-Stage2_6B@fd8ef758daf132d4738325d39546dd76ce26911d",
      "format": "safetensors",
      "file": "13 files, e.g. model-00001-of-00013.safetensors",
      "loader": "transformers / open_clip from_pretrained; Mixpeek Hub import (api/models/huggingface.py)",
      "supported_today": true
    }
  ],
  "use_cases": [
    "video-understanding",
    "action-recognition"
  ],
  "recommended": {
    "value": false,
    "for": [],
    "reason": null
  },
  "demos": [],
  "seo": {
    "title": "InternVideo2 Stage2 6B: embedding model",
    "description": "InternVideo2 Stage2 6B is a video and text embedding model (MIT). Not runnable on Mixpeek today.",
    "keywords": [
      "InternVideo2 Stage2 6B",
      "internvideo2",
      "embedding",
      "OpenGVLab/InternVideo2-Stage2_6B",
      "video-understanding",
      "action-recognition"
    ],
    "faqs": [
      {
        "q": "What is InternVideo2 Stage2 6B best at?",
        "a": "State-of-the-art video-text retrieval and action recognition among open video encoders (per the paper)."
      },
      {
        "q": "What are the limits of InternVideo2 Stage2 6B?",
        "a": "6.4B parameters: multi-GPU-class memory; far too large for scale-to-zero CPU."
      },
      {
        "q": "Can InternVideo2 Stage2 6B be used commercially?",
        "a": "Yes. MIT per the hub tag."
      },
      {
        "q": "How does InternVideo2 Stage2 6B run on Mixpeek?",
        "a": "Requires trust_remote_code (custom_code on the hub) and a large GPU. Would run as a GPU batch plugin only after vendoring the model code."
      }
    ]
  },
  "last_verified": "2026-11-05",
  "sources": [
    "https://huggingface.co/OpenGVLab/InternVideo2-Stage2_6B",
    "https://arxiv.org/abs/2403.15377",
    "https://github.com/OpenGVLab/InternVideo"
  ],
  "registry": null,
  "templates": [],
  "usage": {
    "namespaces": null,
    "collections": null,
    "uncounted": true,
    "templates": 0,
    "counted_at": null,
    "source": null
  },
  "facets": {
    "status": "not_runnable",
    "category": "embedding",
    "family": "internvideo2",
    "license": "MIT",
    "commercial_use": true,
    "modalities_in": [
      "video",
      "text"
    ],
    "modalities_out": [
      "vector"
    ],
    "use_cases": [
      "video-understanding",
      "action-recognition"
    ],
    "recommended": false,
    "release_date": "2025-02",
    "params_millions": 6367,
    "size_on_disk_mb": 25000,
    "vector_dims": null,
    "vector_distance": "cosine",
    "weights_format": "safetensors",
    "formats": [
      "safetensors"
    ],
    "loadable_today": true,
    "trust_remote_code": true,
    "cpu_fallback": false,
    "num_gpus": 1,
    "memory_gb": 40,
    "min_replicas": 0,
    "serving_shape": "batch",
    "served": false,
    "ref_schemes": [
      "hf"
    ],
    "has_evidence": false,
    "cost_measured": false,
    "template_count": 0,
    "usage_namespaces": null,
    "usage_collections": null,
    "usage_uncounted": true,
    "has_demo": false,
    "rank_score": 0
  },
  "rank_score": 0
}