{
  "architectures": [
    "PIXELForPreTraining"
  ],
  "attention_probs_dropout_prob": 0.1,
  "decoder_hidden_size": 264,
  "decoder_intermediate_size": 1056,
  "decoder_num_attention_heads": 6,
  "decoder_num_hidden_layers": 4,
  "hidden_act": "gelu",
  "hidden_dropout_prob": 0.1,
  "hidden_size": 384,
  "image_size": [
    32,
    18432
  ],
  "initializer_range": 0.02,
  "intermediate_size": 1536,
  "layer_norm_eps": 1e-12,
  "mask_ratio": 0.25,
  "model_type": "pixel",
  "noise_scheduler": "cosine",
  "norm_pix_loss": true,
  "norm_pix_pred": false,
  "num_attention_heads": 6,
  "num_channels": 3,
  "num_hidden_layers": 12,
  "patch_size": 32,
  "qkv_bias": true,
  "torch_dtype": "float32",
  "transformers_version": "4.17.0"
}