YOLOv8模型类别ID与对应名称查询及多YOLO模型类别对比咨询

YOLOv8官方预训练的目标检测模型默认使用COCO 80类数据集，具体的类别ID与名称对应如下：

• ID 0: person
• ID 1: bicycle
• ID 2: car
• ID 3: motorcycle
• ID 4: airplane
• ID 5: bus
• ID 6: train
• ID 7: truck
• ID 8: boat
• ID 9: traffic light
• ID 10: fire hydrant
• ID 11: stop sign
• ID 12: parking meter
• ID 13: bench
• ID 14: bird
• ID 15: cat
• ID 16: dog
• ID 17: horse
• ID 18: sheep
• ID 19: cow
• ID 20: elephant
• ID 21: bear
• ID 22: zebra
• ID 23: giraffe
• ID 24: backpack
• ID 25: umbrella
• ID 26: handbag
• ID 27: tie
• ID 28: suitcase
• ID 29: frisbee
• ID 30: skis
• ID 31: snowboard
• ID 32: sports ball
• ID 33: kite
• ID 34: baseball bat
• ID 35: baseball glove
• ID 36: skateboard
• ID 37: surfboard
• ID 38: tennis racket
• ID 39: bottle
• ID 40: wine glass
• ID 41: cup
• ID 42: fork
• ID 43: knife
• ID 44: spoon
• ID 45: bowl
• ID 46: banana
• ID 47: apple
• ID 48: sandwich
• ID 49: orange
• ID 50: broccoli
• ID 51: carrot
• ID 52: hot dog
• ID 53: pizza
• ID 54: donut
• ID 55: cake
• ID 56: chair
• ID 57: couch
• ID 58: potted plant
• ID 59: bed
• ID 60: dining table
• ID 61: toilet
• ID 62: tv
• ID 63: laptop
• ID 64: mouse
• ID 65: remote
• ID 66: keyboard
• ID 67: cell phone
• ID 68: microwave
• ID 69: oven
• ID 70: toaster
• ID 71: sink
• ID 72: refrigerator
• ID 73: book
• ID 74: clock
• ID 75: vase
• ID 76: scissors
• ID 77: teddy bear
• ID 78: hair drier
• ID 79: toothbrush

YOLOv3、YOLOv5、YOLOv7、YOLOv8的官方预训练版本均基于COCO 80类数据集开发，因此类别数量（80个）、类别ID与名称的对应关系完全一致。

若用户基于自定义数据集训练上述模型，类别数量、ID及对应名称会根据自定义标注规则发生变化，此时不同模型之间的类别体系无统一一致性。

内容的提问来源于stack exchange，提问作者Mohamed Safrullah Mohomed