{
  "model_cfg": {
    "embed_dim": 640,
    "vision_cfg": {
      "image_size": 288,
      "layers": [
        4,
        6,
        10,
        6
      ],
      "width": 80,
      "patch_size": null
    },
    "text_cfg": {
      "context_length": 77,
      "vocab_size": 49408,
      "width": 640,
      "heads": 10,
      "layers": 12
    },
    "quick_gelu": true
  },
  "preprocess_cfg": {
    "mean": [
      0.48145466,
      0.4578275,
      0.40821073
    ],
    "std": [
      0.26862954,
      0.26130258,
      0.27577711
    ],
    "interpolation": "bicubic",
    "resize_mode": "shortest"
  }
}