config.json

{
  "_attn_implementation_autoset": true,
  "_name_or_path": "hf-internal-testing/tiny-random-MaskFormerModel",
  "architectures": [
    "MaskFormerModel"
  ],
  "backbone": null,
  "backbone_config": {
    "depths": [
      1,
      1,
      1,
      1
    ],
    "drop_path_rate": 0.1,
    "embed_dim": 96,
    "encoder_stride": 32,
    "hidden_size": 768,
    "image_size": 224,
    "model_type": "maskformer-swin",
    "num_heads": [
      3,
      6,
      12,
      24
    ],
    "out_features": [
      "stage1",
      "stage2",
      "stage3",
      "stage4"
    ],
    "out_indices": [
      1,
      2,
      3,
      4
    ],
    "path_norm": true,
    "window_size": 7
  },
  "backbone_kwargs": null,
  "cross_entropy_weight": 1.0,
  "decoder_config": {
    "d_model": 32,
    "decoder_attention_heads": 2,
    "decoder_ffn_dim": 128,
    "max_position_embeddings": 1024,
    "model_type": "detr",
    "num_queries": 10,
    "scale_embedding": false
  },
  "dice_weight": 1.0,
  "fpn_feature_size": 32,
  "id2label": {
    "0": "LABEL_0",
    "1": "LABEL_1",
    "2": "LABEL_2",
    "3": "LABEL_3"
  },
  "init_std": 0.02,
  "init_xavier_std": 1.0,
  "label2id": {
    "LABEL_0": 0,
    "LABEL_1": 1,
    "LABEL_2": 2,
    "LABEL_3": 3
  },
  "mask_feature_size": 32,
  "mask_weight": 20.0,
  "model_type": "maskformer",
  "no_object_weight": 0.1,
  "num_attention_heads": 8,
  "num_channels": 3,
  "num_hidden_layers": 6,
  "output_auxiliary_logits": null,
  "transformers_version": "4.48.2",
  "use_auxiliary_loss": false,
  "use_pretrained_backbone": false,
  "use_timm_backbone": false
}