config.json

{
  "architectures": [
    "OneFormerModel"
  ],
  "backbone": null,
  "backbone_config": {
    "attention_probs_dropout_prob": 0.0,
    "depths": [
      1,
      1,
      1,
      1
    ],
    "drop_path_rate": 0.3,
    "embed_dim": 16,
    "encoder_stride": 32,
    "hidden_act": "gelu",
    "hidden_dropout_prob": 0.0,
    "hidden_size": 16,
    "image_size": 224,
    "initializer_range": 0.02,
    "layer_norm_eps": 1e-05,
    "mlp_ratio": 4.0,
    "model_type": "swin",
    "num_channels": 3,
    "num_heads": [
      1,
      1,
      2,
      2
    ],
    "num_layers": 4,
    "out_features": [
      "stage1",
      "stage2",
      "stage3",
      "stage4"
    ],
    "out_indices": [
      1,
      2,
      3,
      4
    ],
    "patch_size": 4,
    "qkv_bias": true,
    "stage_names": [
      "stem",
      "stage1",
      "stage2",
      "stage3",
      "stage4"
    ],
    "use_absolute_embeddings": false,
    "window_size": 7
  },
  "class_weight": 2.0,
  "common_stride": 4,
  "contrastive_temperature": 0.07,
  "contrastive_weight": 0.5,
  "conv_dim": 64,
  "decoder_layers": 2,
  "dice_weight": 5.0,
  "dim_feedforward": 64,
  "dropout": 0.1,
  "dtype": "float32",
  "encoder_feedforward_dim": 32,
  "encoder_layers": 2,
  "enforce_input_proj": false,
  "hidden_dim": 64,
  "id2label": {
    "0": "LABEL_0",
    "1": "LABEL_1",
    "2": "LABEL_2",
    "3": "LABEL_3"
  },
  "ignore_value": 255,
  "importance_sample_ratio": 0.75,
  "init_std": 0.02,
  "init_xavier_std": 1.0,
  "is_training": false,
  "label2id": {
    "LABEL_0": 0,
    "LABEL_1": 1,
    "LABEL_2": 2,
    "LABEL_3": 3
  },
  "layer_norm_eps": 1e-05,
  "mask_dim": 64,
  "mask_weight": 5.0,
  "max_seq_len": 77,
  "model_type": "oneformer",
  "no_object_weight": 0.1,
  "norm": "GN",
  "num_attention_heads": 8,
  "num_queries": 10,
  "output_auxiliary_logits": true,
  "oversample_ratio": 3.0,
  "pre_norm": false,
  "query_dec_layers": 2,
  "strides": [
    4,
    8,
    16,
    32
  ],
  "task_seq_len": 77,
  "text_encoder_context_length": 77,
  "text_encoder_n_ctx": 4,
  "text_encoder_num_layers": 6,
  "text_encoder_proj_layers": 2,
  "text_encoder_vocab_size": 99,
  "text_encoder_width": 64,
  "train_num_points": 12544,
  "transformers_version": "5.16.0.dev0",
  "use_auxiliary_loss": true,
  "use_task_norm": true
}