config.json

{
    "_commit_hash": null,
    "architectures": [
      "ViTaminCLIP"
    ],
    "auto_map": {
      "AutoConfig": "configuration_vitamin.ViTaminConfig",
      "AutoModel": "model.ViTaminCLIP"
    },
    "embed_dim": 1152,
    "vision_cfg": {
      "timm_model_name": "vitamin_xlarge_336",
      "timm_model_pretrained": false,
      "timm_pool": "",
      "timm_proj": "linear",
      "timm_drop": 0.0,
      "timm_drop_path": 0.1,
      "image_size": 336
    },
    "text_cfg": {
      "context_length": 77,
      "vocab_size": 49408,
      "width": 1152,
      "heads": 16,
      "layers": 27
  }
}