{ "architectures": [ "MaskFormerForInstanceSegmentation" ], "backbone_config": { "add_cross_attention": false, "attention_probs_dropout_prob": 0.0, "bos_token_id": null, "cross_attention_hidden_size": null, "decoder_start_token_id": null, "depths": [ 2, 2, 18, 2 ], "drop_path_rate": 0.3, "dtype": "float32", "embed_dim": 128, "encoder_stride": 32, "eos_token_id": null, "finetuning_task": null, "hidden_act": "gelu", "hidden_dropout_prob": 0.0, "hidden_size": 1024, "image_size": 384, "in_channels": 3, "initializer_range": 0.02, "is_decoder": false, "layer_norm_eps": 1e-05, "mlp_ratio": 4.0, "model_type": "maskformer-swin", "num_channels": 3, "num_heads": [ 4, 8, 16, 32 ], "num_layers": 4, "out_features": [ "stage1", "stage2", "stage3", "stage4" ], "out_indices": [ 1, 2, 3, 4 ], "pad_token_id": null, "patch_size": 4, "path_norm": true, "prefix": null, "pretrain_img_size": 384, "pruned_heads": {}, "qkv_bias": true, "sep_token_id": null, "stage_names": [ "stem", "stage1", "stage2", "stage3", "stage4" ], "task_specific_params": null, "tie_encoder_decoder": false, "tie_word_embeddings": true, "tokenizer_class": null, "torchscript": false, "use_absolute_embeddings": false, "use_bfloat16": false, "window_size": 12 }, "ce_weight": 1.0, "cross_entropy_weight": 1.0, "decoder_config": { "activation_dropout": 0.0, "activation_function": "relu", "add_cross_attention": false, "attention_dropout": 0.0, "auxiliary_loss": false, "backbone_config": { "_name_or_path": "", "architectures": null, "backbone": "resnet50", "chunk_size_feed_forward": 0, "dtype": null, "features_only": true, "freeze_batch_norm_2d": false, "id2label": { "0": "LABEL_0", "1": "LABEL_1" }, "is_encoder_decoder": false, "label2id": { "LABEL_0": 0, "LABEL_1": 1 }, "model_type": "timm_backbone", "num_channels": 3, "out_features": null, "out_indices": [ 1, 2, 3, 4 ], "output_attentions": false, "output_hidden_states": false, "output_stride": null, "problem_type": null, "return_dict": true, "use_pretrained_backbone": false }, "bbox_cost": 5, "bbox_loss_coefficient": 5, "bos_token_id": null, "class_cost": 1, "cross_attention_hidden_size": null, "d_model": 256, "decoder_attention_heads": 8, "decoder_ffn_dim": 2048, "decoder_layerdrop": 0.0, "decoder_layers": 6, "decoder_start_token_id": null, "dice_loss_coefficient": 1, "dilation": false, "dropout": 0.1, "dtype": "float32", "encoder_attention_heads": 8, "encoder_ffn_dim": 2048, "encoder_layerdrop": 0.0, "encoder_layers": 6, "eos_coefficient": 0.1, "eos_token_id": null, "finetuning_task": null, "giou_cost": 2, "giou_loss_coefficient": 2, "init_std": 0.02, "init_xavier_std": 1.0, "is_decoder": false, "mask_loss_coefficient": 1, "max_position_embeddings": 1024, "model_type": "detr", "num_channels": 3, "num_queries": 100, "pad_token_id": null, "position_embedding_type": "sine", "prefix": null, "pruned_heads": {}, "scale_embedding": false, "sep_token_id": null, "task_specific_params": null, "tie_encoder_decoder": false, "tie_word_embeddings": true, "tokenizer_class": null, "torchscript": false, "use_bfloat16": false }, "dice_weight": 1.0, "dtype": "float32", "fpn_feature_size": 256, "id2label": { "0": "bg", "1": "tumor" }, "init_std": 0.02, "init_xavier_std": 1.0, "label2id": { "animal": 126, "arcade machine": 78, "armchair": 30, "awning, sunshade, sunblind": 86, "bag": 115, "ball": 119, "bannister, banister, balustrade, balusters, handrail": 95, "bar": 77, "barrel, cask": 111, "base, pedestal, stand": 40, "basket, handbasket": 112, "bed": 7, "bench": 69, "bicycle": 127, "blanket, cover": 131, "blind, screen": 63, "boat": 76, "book": 67, "bookcase": 62, "booth": 88, "bottle": 98, "box": 41, "bridge, span": 61, "buffet, counter, sideboard": 99, "building": 1, "bulletin board": 144, "bus": 80, "cabinet": 10, "canopy": 106, "car": 20, "case, display case, showcase, vitrine": 55, "ceiling": 5, "chair": 19, "chandelier": 85, "chest of drawers, chest, bureau, dresser": 44, "clock": 148, "clothes": 92, "coffee table": 64, "column, pillar": 42, "computer": 74, "conveyer belt, conveyor belt, conveyer, conveyor, transporter": 105, "counter": 45, "countertop": 70, "cradle": 117, "crt screen": 141, "curtain": 18, "cushion": 39, "desk": 33, "dirt track": 91, "dishwasher": 129, "door": 14, "earth, ground": 13, "escalator, moving staircase, moving stairway": 96, "falls": 113, "fan": 139, "fence": 32, "field": 29, "fireplace": 49, "flag": 149, "floor": 3, "flower": 66, "food, solid food": 120, "fountain": 104, "glass, drinking glass": 147, "grandstand, covered stand": 51, "grass": 9, "hill": 68, "hood, exhaust hood": 133, "house": 25, "hovel, hut, hutch, shack, shanty": 79, "kitchen island": 73, "lake": 128, "lamp": 36, "land, ground, soil": 94, "light": 82, "microwave": 124, "minibike, motorbike": 116, "mirror": 27, "monitor": 143, "mountain, mount": 16, "ottoman, pouf, pouffe, puff, hassock": 97, "oven": 118, "painting, picture": 22, "palm, palm tree": 72, "path": 52, "person": 12, "pier": 140, "pillow": 57, "plane": 90, "plant": 17, "plate": 142, "plaything, toy": 108, "pole": 93, "pool": 109, "pool table, billiard table, snooker table": 56, "poster, posting, placard, notice, bill, card": 100, "pot": 125, "radiator": 146, "rail": 38, "refrigerator, icebox": 50, "river": 60, "road, route": 6, "rock, stone": 34, "rug": 28, "runway": 54, "sand": 46, "sconce": 134, "screen": 130, "screen door, screen": 58, "sculpture": 132, "sea": 26, "seat": 31, "shelf": 24, "ship": 103, "shower": 145, "sidewalk, pavement": 11, "signboard, sign": 43, "sink": 47, "sky": 2, "skyscraper": 48, "sofa": 23, "stage": 101, "stairs": 53, "stairway, staircase": 59, "step, stair": 121, "stool": 110, "stove": 71, "street lamp": 87, "swivel chair": 75, "table": 15, "tank, storage tank": 122, "tent": 114, "toilet, can, commode, crapper, pot, potty, stool, throne": 65, "towel": 81, "tower": 84, "trade name": 123, "traffic light": 136, "trash can": 138, "tray": 137, "tree": 4, "truck": 83, "tub": 37, "tv": 89, "van": 102, "vase": 135, "wall": 0, "wardrobe, closet, press": 35, "washer, automatic washer, washing machine": 107, "water": 21, "window ": 8 }, "mask_feature_size": 256, "mask_weight": 20.0, "model_type": "maskformer", "no_object_weight": 0.1, "num_attention_heads": 8, "num_hidden_layers": 6, "num_queries": 100, "output_auxiliary_logits": null, "transformers_version": "5.6.2", "use_auxiliary_loss": false }