{
  "id": "internvideo2-stage2-1b-224p-f4",
  "identity": {
    "family": "internvideo2",
    "display_name": "InternVideo2 Stage2 1B (224p, 4 frames)",
    "hub_id": "OpenGVLab/InternVideo2-Stage2_1B-224p-f4",
    "api_name": null,
    "revision": null,
    "variants": [],
    "license": "Apache-2.0",
    "commercial_use": true,
    "commercial_use_note": "Apache-2.0 per the hub tag; access gated by an automatic agreement on the hub.",
    "release_date": "2024-03",
    "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": "Video-text retrieval and zero-shot action recognition on short clips; motion-aware where frame CLIP models are not.",
    "worst_at": "Only 4 frames at 224p per clip: long or slow scenes need splitting; research code base, not a packaged loader."
  },
  "io": {
    "input": {
      "video": {
        "frames": 4,
        "resolution": "224x224",
        "sampling": "uniform over the clip"
      },
      "text": {
        "max_tokens": 40
      }
    },
    "output": {
      "vector": {
        "dims": "projection dim per the repo config (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": "Weights are a single .pt pickle and the model is defined only in the InternVideo GitHub repo (multi_modality), so it needs vendored model code plus a safetensors conversion before it can be a plugin; GPU in practice for 1B.",
    "serve_apps": [],
    "runtime": "InternVideo repo (multi_modality InternVideo2 Stage2)",
    "weights_format": "pickle",
    "trust_remote_code": false,
    "size_on_disk_mb": 2821,
    "params_millions": 1000,
    "preprocessing": "Sample 4 frames uniformly, resize to 224, normalize ImageNet mean/std.",
    "postprocessing": "L2-normalize video and text 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": 8,
      "num_cpus": 2,
      "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_1B-224p-f4@4362e1f88a992e7edbfd7696f7f78b7f79426dfd"
  ],
  "formats": [
    {
      "ref": "hf://OpenGVLab/InternVideo2-Stage2_1B-224p-f4@4362e1f88a992e7edbfd7696f7f78b7f79426dfd",
      "format": "pytorch_pt",
      "file": "InternVideo2-stage2_1b-224p-f4.pt",
      "loader": "torch.load (pickle); Mixpeek refuses pickle weights (shared/security/torch_pickle.py)",
      "supported_today": false
    }
  ],
  "use_cases": [
    "video-understanding",
    "action-recognition"
  ],
  "recommended": {
    "value": false,
    "for": [],
    "reason": null
  },
  "demos": [],
  "seo": {
    "title": "InternVideo2 Stage2 1B (224p, 4 frames): embedding model",
    "description": "InternVideo2 Stage2 1B (224p, 4 frames) is a video and text embedding model (Apache-2.0). Not runnable on Mixpeek today.",
    "keywords": [
      "InternVideo2 Stage2 1B (224p, 4 frames)",
      "internvideo2",
      "embedding",
      "OpenGVLab/InternVideo2-Stage2_1B-224p-f4",
      "video-understanding",
      "action-recognition"
    ],
    "faqs": [
      {
        "q": "What is InternVideo2 Stage2 1B (224p, 4 frames) best at?",
        "a": "Video-text retrieval and zero-shot action recognition on short clips; motion-aware where frame CLIP models are not."
      },
      {
        "q": "What are the limits of InternVideo2 Stage2 1B (224p, 4 frames)?",
        "a": "Only 4 frames at 224p per clip: long or slow scenes need splitting; research code base, not a packaged loader."
      },
      {
        "q": "Can InternVideo2 Stage2 1B (224p, 4 frames) be used commercially?",
        "a": "Yes. Apache-2.0 per the hub tag; access gated by an automatic agreement on the hub."
      },
      {
        "q": "How does InternVideo2 Stage2 1B (224p, 4 frames) run on Mixpeek?",
        "a": "Weights are a single .pt pickle and the model is defined only in the InternVideo GitHub repo (multi_modality), so it needs vendored model code plus a safetensors conversion before it can be a plugin; GPU in practice for 1B."
      }
    ]
  },
  "last_verified": "2026-11-05",
  "sources": [
    "https://huggingface.co/OpenGVLab/InternVideo2-Stage2_1B-224p-f4",
    "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": "Apache-2.0",
    "commercial_use": true,
    "modalities_in": [
      "video",
      "text"
    ],
    "modalities_out": [
      "vector"
    ],
    "use_cases": [
      "video-understanding",
      "action-recognition"
    ],
    "recommended": false,
    "release_date": "2024-03",
    "params_millions": 1000,
    "size_on_disk_mb": 2821,
    "vector_dims": null,
    "vector_distance": "cosine",
    "weights_format": "pickle",
    "formats": [
      "pytorch_pt"
    ],
    "loadable_today": false,
    "trust_remote_code": false,
    "cpu_fallback": false,
    "num_gpus": 1,
    "memory_gb": 8,
    "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
}