Add config

config.json

@@ -1,11 +1,17 @@
 {
   "act_decoder": "relu",
   "act_encoder": "silu",
-  "architectures": [
-    "RTDetrModel"
-  ],
   "aux_loss": true,
-  "backbone": "resnet50d",
+  "backbone_config": {
+    "backbone": "resnet50d",
+    "freeze_batch_norm_2d": true,
+    "model_type": "timm_backbone",
+    "out_indices": [
+      2,
+      3,
+      4
+    ]
+  },
   "box_noise_scale": 1.0,
   "depth_mult": 1.0,
   "dim_feedforward": 1024,
@@ -32,19 +38,10 @@
   ],
   "focal_loss_alpha": 0.75,
   "focal_loss_gamma": 2.0,
-  "freeze_batch_norm_2d": true,
   "hidden_dim": 256,
   "id2label": {
     "0": "person",
     "1": "bicycle",
-    "2": "car",
-    "3": "motorcycle",
-    "4": "airplane",
-    "5": "bus",
-    "6": "train",
-    "7": "truck",
-    "8": "boat",
-    "9": "traffic light",
     "10": "fire hydrant",
     "11": "stop sign",
     "12": "parking meter",
@@ -55,6 +52,7 @@
     "17": "horse",
     "18": "sheep",
     "19": "cow",
+    "2": "car",
     "20": "elephant",
     "21": "bear",
     "22": "zebra",
@@ -65,6 +63,7 @@
     "27": "tie",
     "28": "suitcase",
     "29": "frisbee",
+    "3": "motorcycle",
     "30": "skis",
     "31": "snowboard",
     "32": "sports ball",
@@ -75,6 +74,7 @@
     "37": "surfboard",
     "38": "tennis racket",
     "39": "bottle",
+    "4": "airplane",
     "40": "wine glass",
     "41": "cup",
     "42": "fork",
@@ -85,6 +85,7 @@
     "47": "apple",
     "48": "sandwich",
     "49": "orange",
+    "5": "bus",
     "50": "broccoli",
     "51": "carrot",
     "52": "hot dog",
@@ -95,6 +96,7 @@
     "57": "couch",
     "58": "potted plant",
     "59": "bed",
+    "6": "train",
     "60": "dining table",
     "61": "toilet",
     "62": "tv",
@@ -105,6 +107,7 @@
     "67": "cell phone",
     "68": "microwave",
     "69": "oven",
+    "7": "truck",
     "70": "toaster",
     "71": "sink",
     "72": "refrigerator",
@@ -114,7 +117,9 @@
     "76": "scissors",
     "77": "teddy bear",
     "78": "hair drier",
-    "79": "toothbrush"
+    "79": "toothbrush",
+    "8": "boat",
+    "9": "traffic light"
   },
   "in_channels": [
     512,
@@ -123,86 +128,86 @@
   ],
   "initializer_range": 0.02,
   "label2id": {
-    "person": 0,
-    "bicycle": 1,
-    "car": 2,
-    "motorcycle": 3,
     "airplane": 4,
-    "bus": 5,
-    "train": 6,
-    "truck": 7,
-    "boat": 8,
-    "traffic light": 9,
-    "fire hydrant": 10,
-    "stop sign": 11,
-    "parking meter": 12,
-    "bench": 13,
-    "bird": 14,
-    "cat": 15,
-    "dog": 16,
-    "horse": 17,
-    "sheep": 18,
-    "cow": 19,
-    "elephant": 20,
-    "bear": 21,
-    "zebra": 22,
-    "giraffe": 23,
+    "apple": 47,
     "backpack": 24,
-    "umbrella": 25,
-    "handbag": 26,
-    "tie": 27,
-    "suitcase": 28,
-    "frisbee": 29,
-    "skis": 30,
-    "snowboard": 31,
-    "sports ball": 32,
-    "kite": 33,
+    "banana": 46,
     "baseball bat": 34,
     "baseball glove": 35,
-    "skateboard": 36,
-    "surfboard": 37,
-    "tennis racket": 38,
+    "bear": 21,
+    "bed": 59,
+    "bench": 13,
+    "bicycle": 1,
+    "bird": 14,
+    "boat": 8,
+    "book": 73,
     "bottle": 39,
-    "wine glass": 40,
-    "cup": 41,
-    "fork": 42,
-    "knife": 43,
-    "spoon": 44,
     "bowl": 45,
-    "banana": 46,
-    "apple": 47,
-    "sandwich": 48,
-    "orange": 49,
     "broccoli": 50,
-    "carrot": 51,
-    "hot dog": 52,
-    "pizza": 53,
-    "donut": 54,
+    "bus": 5,
     "cake": 55,
+    "car": 2,
+    "carrot": 51,
+    "cat": 15,
+    "cell phone": 67,
     "chair": 56,
+    "clock": 74,
     "couch": 57,
-    "potted plant": 58,
-    "bed": 59,
+    "cow": 19,
+    "cup": 41,
     "dining table": 60,
-    "toilet": 61,
-    "tv": 62,
-    "laptop": 63,
-    "mouse": 64,
-    "remote": 65,
+    "dog": 16,
+    "donut": 54,
+    "elephant": 20,
+    "fire hydrant": 10,
+    "fork": 42,
+    "frisbee": 29,
+    "giraffe": 23,
+    "hair drier": 78,
+    "handbag": 26,
+    "horse": 17,
+    "hot dog": 52,
     "keyboard": 66,
-    "cell phone": 67,
+    "kite": 33,
+    "knife": 43,
+    "laptop": 63,
     "microwave": 68,
+    "motorcycle": 3,
+    "mouse": 64,
+    "orange": 49,
     "oven": 69,
-    "toaster": 70,
-    "sink": 71,
+    "parking meter": 12,
+    "person": 0,
+    "pizza": 53,
+    "potted plant": 58,
     "refrigerator": 72,
-    "book": 73,
-    "clock": 74,
-    "vase": 75,
+    "remote": 65,
+    "sandwich": 48,
     "scissors": 76,
+    "sheep": 18,
+    "sink": 71,
+    "skateboard": 36,
+    "skis": 30,
+    "snowboard": 31,
+    "spoon": 44,
+    "sports ball": 32,
+    "stop sign": 11,
+    "suitcase": 28,
+    "surfboard": 37,
     "teddy bear": 77,
-    "hair drier": 78,
-    "toothbrush": 79
+    "tennis racket": 38,
+    "tie": 27,
+    "toaster": 70,
+    "toilet": 61,
+    "toothbrush": 79,
+    "traffic light": 9,
+    "train": 6,
+    "truck": 7,
+    "tv": 62,
+    "umbrella": 25,
+    "vase": 75,
+    "wine glass": 40,
+    "zebra": 22
   },
   "label_noise_ratio": 0.5,
   "learnt_init_query": false,
@@ -220,15 +225,9 @@
   "num_head": 8,
   "num_levels": 3,
   "num_queries": 300,
-  "out_indices": [
-    2,
-    3,
-    4
-  ],
   "pe_temperature": 10000,
   "position_embed_type": "sine",
-  "torch_dtype": "float32",
-  "transformers_version": "4.35.0.dev0",
+  "transformers_version": "4.36.0.dev0",
   "use_encoder_idx": [
     2
   ],