{
  "architectures": [
    "PoolFormerForImageClassification"
  ],
  "depths": [
    2,
    2,
    2,
    2
  ],
  "drop_path_rate": 0.0,
  "hidden_act": "gelu",
  "hidden_dropout_prob": 0.1,
  "hidden_sizes": [
    16,
    32,
    64,
    128
  ],
  "image_size": 64,
  "initializer_range": 0.02,
  "layer_scale_init_value": 1e-05,
  "mlp_ratio": 4.0,
  "model_type": "poolformer",
  "num_channels": 3,
  "num_encoder_blocks": 4,
  "padding": [
    2,
    1,
    1,
    1
  ],
  "patch_size": 16,
  "patch_sizes": [
    7,
    3,
    3,
    3
  ],
  "pool_size": 3,
  "stride": 16,
  "strides": [
    4,
    2,
    2,
    2
  ],
  "torch_dtype": "float32",
  "transformers_version": "4.28.0.dev0",
  "use_layer_scale": true
}