config.json

{
  "architectures": [
    "SamHQModel"
  ],
  "initializer_range": 0.02,
  "mask_decoder_config": {
    "attention_downsample_rate": 2,
    "hidden_act": "relu",
    "hidden_size": 256,
    "iou_head_depth": 3,
    "iou_head_hidden_dim": 256,
    "layer_norm_eps": 1e-06,
    "mlp_dim": 2048,
    "model_type": "",
    "num_attention_heads": 8,
    "num_hidden_layers": 2,
    "num_multimask_outputs": 3,
    "vit_dim": 1280
  },
  "model_type": "sam_hq",
  "prompt_encoder_config": {
    "hidden_act": "gelu",
    "hidden_size": 256,
    "image_embedding_size": 64,
    "image_size": 1024,
    "layer_norm_eps": 1e-06,
    "mask_input_channels": 16,
    "model_type": "",
    "num_point_embeddings": 4,
    "patch_size": 16
  },
  "torch_dtype": "float32",
  "transformers_version": "4.52.0.dev0",
  "vision_config": {
    "attention_dropout": 0.0,
    "global_attn_indexes": [
      7,
      15,
      23,
      31
    ],
    "hidden_act": "gelu",
    "hidden_size": 1280,
    "image_size": 1024,
    "initializer_range": 1e-10,
    "layer_norm_eps": 1e-06,
    "mlp_dim": 5120,
    "mlp_ratio": 4.0,
    "model_type": "sam_hq_vision_model",
    "num_attention_heads": 16,
    "num_channels": 3,
    "num_hidden_layers": 32,
    "num_pos_feats": 128,
    "output_channels": 256,
    "patch_size": 16,
    "qkv_bias": true,
    "use_abs_pos": true,
    "use_rel_pos": true,
    "window_size": 14
  }
}