{
  "architectures": [
    "ViTMAEWithRegistersForPreTraining"
  ],
  "attention_probs_dropout_prob": 0.0,
  "attn_implementation": "sdpa",
  "decoder_hidden_size": 512,
  "decoder_intermediate_size": 2048,
  "decoder_num_attention_heads": 16,
  "decoder_num_hidden_layers": 8,
  "drop_path_rate": 0.1,
  "hidden_act": "gelu",
  "hidden_dropout_prob": 0.0,
  "hidden_size": 768,
  "image_size": 224,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "layer_norm_eps": 1e-12,
  "mask_ratio": 0.0,
  "model_type": "vit_mae",
  "n_registers": 4,
  "norm_pix_loss": false,
  "num_attention_heads": 12,
  "num_channels": 3,
  "num_hidden_layers": 12,
  "patch_size": 16,
  "qkv_bias": true,
  "torch_dtype": "float32",
  "transformers_version": "4.46.3",
  "use_dino_encoder": false
}