config.json

{
  "architectures": [
    "OneFormerForUniversalSegmentation"
  ],
  "backbone_config": {
    "attention_probs_dropout_prob": 0.0,
    "depths": [
      3,
      4,
      18,
      5
    ],
    "dilations": [
      [
        1,
        18,
        1
      ],
      [
        1,
        5,
        1,
        9
      ],
      [
        1,
        2,
        1,
        3,
        1,
        4,
        1,
        2,
        1,
        3,
        1,
        4,
        1,
        2,
        1,
        3,
        1,
        4
      ],
      [
        1,
        2,
        1,
        2,
        1
      ]
    ],
    "drop_path_rate": 0.3,
    "embed_dim": 192,
    "encoder_stride": 32,
    "feature_channels": [
      192,
      384,
      768,
      1536
    ],
    "hidden_act": "gelu",
    "hidden_dropout_prob": 0,
    "kernel_size": 7,
    "layer_scale_init_value": 0.0,
    "mlp_ratio": 2.0,
    "num_channels": 3,
    "num_heads": [
      6,
      12,
      24,
      48
    ],
    "patch_size": 4,
    "qkv_bias": true,
    "strides": [
      4,
      8,
      16,
      32
    ]
  },
  "decoder_config": {
    "common_stride": 4,
    "conv_dim": 256,
    "decoder_layers": 10,
    "dim_feedforward": 2048,
    "dropout": 0.1,
    "encoder_feedforward_dim": 1024,
    "encoder_layers": 6,
    "enforce_input_proj": false,
    "hidden_dim": 256,
    "mask_dim": 256,
    "norm": "GN",
    "num_heads": 8,
    "pre_norm": false,
    "query_dec_layers": 2,
    "use_task_norm": true
  },
  "general_config": {
    "backbone_type": "dinat",
    "class_weight": 2.0,
    "contrastive_temperature": 0.07,
    "contrastive_weight": 0.5,
    "deep_supervision": true,
    "dice_weight": 5.0,
    "ignore_value": 255,
    "importance_sample_ratio": 0.75,
    "init_std": 0.02,
    "init_xavier_std": 1.0,
    "is_train": false,
    "layer_norm_eps": 1e-05,
    "mask_weight": 5.0,
    "no_object_weight": 0.1,
    "num_classes": 19,
    "num_queries": 250,
    "output_auxiliary_logits": true,
    "oversample_ratio": 3.0,
    "train_num_points": 12544,
    "use_auxiliary_loss": true
  },
  "hidden_size": 256,
  "id2label": {
    "0": "road",
    "1": "sidewalk",
    "2": "building",
    "3": "wall",
    "4": "fence",
    "5": "pole",
    "6": "traffic light",
    "7": "traffic sign",
    "8": "vegetation",
    "9": "terrain",
    "10": "sky",
    "11": "person",
    "12": "rider",
    "13": "car",
    "14": "truck",
    "15": "bus",
    "16": "train",
    "17": "motorcycle",
    "18": "bicycle"
  },
  "init_std": 0.02,
  "init_xavier_std": 1.0,
  "label2id": {
    "bicycle": 18,
    "building": 2,
    "bus": 15,
    "car": 13,
    "fence": 4,
    "motorcycle": 17,
    "person": 11,
    "pole": 5,
    "rider": 12,
    "road": 0,
    "sidewalk": 1,
    "sky": 10,
    "terrain": 9,
    "traffic light": 6,
    "traffic sign": 7,
    "train": 16,
    "truck": 14,
    "vegetation": 8,
    "wall": 3
  },
  "model_type": "oneformer",
  "num_attention_heads": 8,
  "num_hidden_layers": 10,
  "output_attentions": true,
  "output_hidden_states": true,
  "text_encoder_config": {
    "max_seq_len": 77,
    "task_seq_len": 77,
    "text_encoder_context_length": 77,
    "text_encoder_n_ctx": 16,
    "text_encoder_num_layers": 6,
    "text_encoder_proj_layers": 2,
    "text_encoder_vocab_size": 49408,
    "text_encoder_width": 256
  },
  "torch_dtype": "float32",
  "transformers_version": "4.25.0.dev0"
}