open_clip_config.json

{
  "model_cfg": {
    "embed_dim": 512,
    "vision_cfg": {
      "timm_model_name": "swinv2_base_window12_192.ms_in22k",
      "timm_model_pretrained": true,
      "patch_dropout": 0.6,
      "timm_pool": "avg",
      "timm_proj": "linear",
      "image_size": 192
    },
    "text_cfg": {
      "hf_model_name": "roberta-base",
      "hf_tokenizer_name": "roberta-base",
      "hf_pooler_type": "mean_pooler"
    }
  },
  "preprocess_cfg": {
    "mean": [
      0.48145466,
      0.4578275,
      0.40821073
    ],
    "std": [
      0.26862954,
      0.26130258,
      0.27577711
    ],
    "interpolation": "bicubic",
    "resize_mode": "shortest"
  }
}