您好，我尝试训练自己找到的其他数据集，我将数据集标注转换成YOLO格式，文件夹结构如下，其中detection文件夹中是目标检测的YOLO格式标注，segement_area文件夹中是可行区域的YOLO格式标注，segment_lane是车道线的YOLO格式标注。 Hello, I am trying to train on another dataset I found. I have converted the dataset annotations into YOLO format. The folder structure is as follows: The detection folder contains YOLO format annotations for object detection, the segement_area folder contains YOLO format annotations for feasible areas, and the segment_lane folder contains YOLO format annotations for lane markings.

The id represent the correspondence relation

├─dataset root │ ├─images │ │ ├─train │ │ ├─val │ ├─detection │ │ ├─labels │ │ │ ├─train │ │ │ ├─val │ ├─segment_area │ │ ├─labels │ │ │ ├─train │ │ │ ├─val │ ├─segment_lane │ │ ├─labels │ │ │ ├─train │ │ │ ├─val

例如sement_area/labels/val中的一个txt文件内容是： For example, the content of a .txt file in sement_area/labels/val is: 10 0.1306242515625 0.7815792680555556 0.0 0.8370350055555555 0.0 0.9999362277777777 0.46675068984375 1.0009756875 0.4173604421875 0.9570732458333334 0.1978859921875 0.9548784416666667 0.19983560859375 0.8006421791666666 0.1306242515625 0.7815792680555556 10 0.38896105 0.8978049486111112 0.50300084921875 0.8734278319444444 0.6151037515625 0.9306165583333332 0.6131541359375 0.9999362277777777 0.49021105781249996 1.0009756875 0.38896105 0.8978049486111112

同时，我修改了datasets中的yaml文件： At the same time, I modified the YAML file in the datasets： ` path: D:\file\graduate\data\guoneng_dataset\datasets\datasets\YOLOformat # dataset root dir

train: images\train # train images for object detection (relative to 'path')

val: images\val # val images for object detection (relative to 'path')

test: images\val # test images for object detection (relative to 'path')

labels_list:

detection
segment_area
segment_lane

tnc: 7 # number of classes nc_list: [5,1,1] map: [None,{'9':'0'},{'10':'0'}]

names: 0: person # 1 1: rider # 2 2: car # 3 3: bus # 3 4: truck # 3 5: bike # 3 6: motorcycle # 3 7: traffic sign # 4 8: traffic light # 5 9: lane # Combine all lane-related categories into one 6 10: area # Combine all area-related categories into one 7 `

以及模型的配置yaml文件： and the model's configuration YAML file: ` tnc: 7

scales: n: [0.33, 0.25, 1024] # YOLOv8n summary: 225 layers, 3157200 parameters, 3157184 gradients, 8.9 GFLOPs s: [0.33, 0.50, 1024] # YOLOv8s summary: 225 layers, 11166560 parameters, 11166544 gradients, 28.8 GFLOPs m: [0.67, 0.75, 768] # YOLOv8m summary: 295 layers, 25902640 parameters, 25902624 gradients, 79.3 GFLOPs l: [1.00, 1.00, 512] # YOLOv8l summary: 365 layers, 43691520 parameters, 43691504 gradients, 165.7 GFLOPs x: [1.00, 1.25, 512] # YOLOv8x summary: 365 layers, 68229648 parameters, 68229632 gradients, 258.5 GFLOPs

scale: n

backbone:

[from, repeats, module, args]

[-1, 1, Conv, [64, 3, 2]] # 0-P1/2
[-1, 1, Conv, [128, 3, 2]] # 1-P2/4
[-1, 3, C2f, [128, True]]
[-1, 1, Conv, [256, 3, 2]] # 3-P3/8
[-1, 6, C2f, [256, True]]
[-1, 1, Conv, [512, 3, 2]] # 5-P4/16
[-1, 6, C2f, [512, True]]
[-1, 1, Conv, [1024, 3, 2]] # 7-P5/32
[-1, 3, C2f, [1024, True]]
[-1, 1, SPPF, [1024, 5]] # 9

head:

[-1, 1, nn.Upsample, [None, 2, 'nearest']]
[[-1, 6], 1, Concat, [1]] # cat backbone P4
[-1, 3, C2f, [512]] # 12
[-1, 1, nn.Upsample, [None, 2, 'nearest']]
[[-1, 4], 1, Concat, [1]] # cat backbone P3
[-1, 3, C2f, [256]] # 15 (P3/8-small)
[-1, 1, Conv, [256, 3, 2]]
[[-1, 12], 1, Concat, [1]] # cat head P4
[-1, 3, C2f, [512]] # 18 (P4/16-medium)
[-1, 1, Conv, [512, 3, 2]]
[[-1, 9], 1, Concat, [1]] # cat head P5
[-1, 3, C2f, [1024]] # 21 (P5/32-large)

lane
[9, 1, nn.Upsample, [None, 2, 'nearest']]
[[-1, 6], 1, Concat_dropout, [1]] # cat backbone P4
[-1, 3, C2f, [512]] # 24
[-1, 1, nn.Upsample, [None, 2, 'nearest']]
[[-1, 4], 1, Concat_dropout, [1]] # cat backbone P3
[-1, 3, C2f, [256]] # 27 (P3/8-small)
[-1, 1, nn.Upsample, [None, 2, 'nearest']] # for lane segmentation
[[-1, 2], 1, Concat_dropout, [1]] # cat backbone P2
[-1, 3, C2f, [128]] # 30 (P2)
[-1, 1, nn.Upsample, [None, 2, 'nearest']] #
[[-1, 0], 1, Concat_dropout, [1]] # cat backbone P1
[-1, 3, C2f, [64]] # 33 (P1)

drivable
[9, 1, nn.Upsample, [None, 2, 'nearest']]
[[-1, 6], 1, Concat_dropout, [1]] # cat backbone P4
[-1, 3, C2f, [512]] # 36
[-1, 1, nn.Upsample, [None, 2, 'nearest']]
[[-1, 4], 1, Concat_dropout, [1]] # cat backbone P3
[-1, 3, C2f, [256]] # 39 (P3/8-small)
[-1, 1, nn.Upsample, [None, 2, 'nearest']] # 30 for drivable segmentation
[[-1, 2], 1, Concat_dropout, [1]]
[-1, 3, C2f, [128]] # 42 (P2)
[-1, 1, nn.Upsample, [None, 2, 'nearest']] #
[[-1, 0], 1, Concat_dropout, [1]]
[-1, 3, C2f, [64]] # 45 (P1)

tasks

[[15, 18, 21], 1, Detect, [5]] # 36 Detect(P3, P4, P5)
[[45], 1, Segment, [1, 32, 256]] # 37 drivable-Segment [1,32,256] was not working, you should change the head.py
[[33], 1, Segment, [1, 32, 256]] # 38 lane-Segment [1,32,256] was not working, you should change the head.py `

并修改了train.py中相应的文件路径，将amp设置为false是因为如果不设置的话在检查amp的时候报错了： I also modified the file paths in train.py accordingly and set amp to False because an error occurred when checking amp if it wasn't set. `import sys sys.path.insert(0, r"D:\file\graduate\code\YOLOv8-multi-task-main\YOLOv8-multi-task-main\ultralytics") from ultralytics import YOLO

model = YOLO(r'D:\file\graduate\code\YOLOv8-multi-task-main\YOLOv8-multi-task-main\ultralytics\models\v8\yolov8-bdd-v4-n-myCarDataset.yaml', task='multi') # build a new model from YAML

model.train(data=r'D:\file\graduate\code\YOLOv8-multi-task-main\YOLOv8-multi-task-main\ultralytics\datasets\datasets_cars.yaml', batch=12, epochs=30, imgsz=(640,640), device=[0], name='yolopm', val=True, task='multi', classes=[0,1,2,3,4,5,6,7,8,9,10], combine_class=[2,3,4,9], single_cls=False,amp=False) `

不知道是不是数据集格式的问题还是代码有哪里没有修改，运行训练代码后出现报错： I'm not sure if it's an issue with the dataset format or if there's something in the code that hasn't been modified, but after running the training code, I encountered an error: WARNING Box and segment counts should be equal, but got len(segments) = 43, len(boxes) = 106. To resolve this only boxes will be used and all segments will be removed. To avoid this please supply either a detect or segment dataset, not a detect-segment mixed dataset. Traceback (most recent call last): File "ultralytics/train_myCarDataset.py", line 12, in model.train(data=r'D:\file\graduate\code\YOLOv8-multi-task-main\YOLOv8-multi-task-main\ultralytics\datasets\datasets_cars.yaml', File "d:\file\graduate\code\yolov8-multi-task-main\yolov8-multi-task-main\ultralytics\yolo\engine\model.py", line 390, in train self.trainer.train() File "d:\file\graduate\code\yolov8-multi-task-main\yolov8-multi-task-main\ultralytics\yolo\engine\trainer.py", line 195, in train self._do_train(world_size) File "d:\file\graduate\code\yolov8-multi-task-main\yolov8-multi-task-main\ultralytics\yolo\engine\trainer.py", line 292, in _do_train self._setup_train(world_size) File "d:\file\graduate\code\yolov8-multi-task-main\yolov8-multi-task-main\ultralytics\yolo\engine\trainer.py", line 266, in _setup_train self.test_loader = self.get_dataloader(self.testset, batch_size=batch_size 2, rank=-1, mode='val') File "d:\file\graduate\code\yolov8-multi-task-main\yolov8-multi-task-main\ultralytics\yolo\v8\DecSeg\train.py", line 71, in get_dataloader dataset = self.build_dataset(dataset_path, mode, batch_size) File "d:\file\graduate\code\yolov8-multi-task-main\yolov8-multi-task-main\ultralytics\yolo\v8\DecSeg\train.py", line 45, in build_dataset return build_yolo_dataset(self.args, img_path, batch, self.data, mode=mode, rect=mode == 'val', stride=gs) File "d:\file\graduate\code\yolov8-multi-task-main\yolov8-multi-task-main\ultralytics\yolo\data\build.py", line 74, in build_yolo_dataset return YOLODataset( File "d:\file\graduate\code\yolov8-multi-task-main\yolov8-multi-task-main\ultralytics\yolo\data\dataset.py", line 40, in init super().init(args, **kwargs) File "d:\file\graduate\code\yolov8-multi-task-main\yolov8-multi-task-main\ultralytics\yolo\data\base.py", line 83, in init self.set_rectangle() File "d:\file\graduate\code\yolov8-multi-task-main\yolov8-multi-task-main\ultralytics\yolo\data\base.py", line 228, in set_rectangle self.im_files = [self.im_files[i] for i in irect] File "d:\file\graduate\code\yolov8-multi-task-main\yolov8-multi-task-main\ultralytics\yolo\data\base.py", line 228, in self.im_files = [self.im_files[i] for i in irect] IndexError: list index out of range

想请教一下这个报错是哪里的问题，应该怎么解决呢？谢谢！ I would like to ask about this error—what might be causing it, and how can it be fixed? Thank you!

JiayuanWang-JW / YOLOv8-multi-task

Traning Problem-训练其他数据集出现报错：IndexError: list index out of range #98

The id represent the correspondence relation

[from, repeats, module, args]