coco128 训练

原创

已于 2025-07-03 00:54:07 修改 · 1.2k 阅读

26 ·

CC 4.0 BY-SA版权

文章标签：

#深度学习 #机器学习 #人工智能

于 2025-06-24 22:57:29 首次发布

1 环境准备

Pytorch学习 -优快云博客

2 数据集

2.1 配置数据集

打开YOLOv10项目，

找到 yolov10\ultralytics\cfg\datasets\coco128.yaml：

数据集配置文件数据集可自动下载也可先下载好放到指定目录然后配置好下面的路径

2.1.1 路径配置

相对路径/或绝对路径建议相对路径

如果自定义数据集需要修改此文件或者copy一份改写

2.1.2 分类配置

80分类

2.2 数据结构解剖

图像分布：128张图片平均包含7.26个标注实例
类别统计：实际出现32个类别（占80类的40%），前5大类别为：
1. person (17.8%)
2. car (9.2%)
3. chair (7.1%)
4. bottle (5.3%)
5. dining table (4.9%)

2.3 标注格式

原始COCO JSON转YOLO TXT格式示例：如需要

# 转换前（COCO JSON）
{"bbox": [183,24,94,120], "category_id": 18}

# 转换后（YOLO TXT）
18 0.35 0.42 0.12 0.15  # 类别ID 中心x 中心y 宽度 高度

2. 训练

2.1 训练参数

根据硬件调整 epochs/batch/device参数

train

yolo detect train data=coco128.yaml model=yolov10n.yaml epochs=100 batch=32 imgsz=640 device=0（GPU设备修改）

或


from ultralytics import YOLO

# Load YOLOv10n model from scratch
model = YOLO("yolov10n.yaml")

# Train the model
model.train(data="coco8.yaml", epochs=100, batch=32,imgsz=640)

2.2 超参数优化方案

# hyp.scratch.yaml 关键参数调整
lr0: 0.01  → 0.005   # 降低初始学习率
warmup_epochs: 3 → 5  # 延长热身期
mixup: 0.0 → 0.1      # 启用混合增强

2.3 数据增强策略

# dataset.yaml 增强配置
augmentations:
  - hsv_h: 0.015      # 色相增强
  - hsv_s: 0.7        # 饱和度增强  
  - flipud: 0.5       # 垂直翻转概率
  - mosaic: 1.0       # 马赛克增强

3. 训练过程

epochs= 100 通常是个经验值

epochs参数指模型遍历整个训练数据集的轮次数，其范围需根据任务复杂度、数据集规模及资源条件动态调整。‌

实验场景‌：简单任务或小数据集可能只需50-150轮；复杂任务（如高精度检测）可能需要500-1000+轮.

YOLOv5默认训练轮数为300；YOLOv8/v9/v10参考值通常为100-500。

3.1 epochs 调整依据

‌因素‌	‌建议调整方向‌	‌依据说明‌
‌数据集规模‌	数据量少 → 降低轮数（如100-200）	避免过拟合
‌任务复杂度‌	小目标检测/密集场景 → 增加轮数	需更充分学习特征
‌资源限制‌	显存不足 → 减小批次并增加轮数	平衡训练稳定性与收敛性
‌早停机制‌	验证集性能饱和时提前终止	防止无效训练

3.2 过程日志

yolov10) PS F:\ai\yolov\v10\yolov10> & C:/Users/qiu-t/anaconda3/envs/yolov10/python.exe f:/ai/yolov/v10/yolov10/coco_128.py
New https://pypi.org/project/ultralytics/8.3.158 available 😃 Update with 'pip install -U ultralytics'
Ultralytics YOLOv8.1.34 🚀 Python-3.12.11 torch-2.7.1+cu128 CUDA:0 (NVIDIA GeForce RTX 5060 Laptop GPU, 8151MiB)
engine\trainer: task=detect, mode=train, model=yolov10n.yaml, data=coco128.yaml, epochs=100, time=None, patience=100, batch=32, imgsz=640, save=True, save_period=-1, val_period=1, cache=False, device=None, workers=8, project=None, name=train6, exist_ok=False, pretrained=True, optimizer=auto, verbose=True, seed=0, deterministic=True, single_cls=False, rect=False, cos_lr=False, close_mosaic=10, resume=False, amp=True, fraction=1.0, profile=False, freeze=None, multi_scale=False, overlap_mask=True, mask_ratio=4, dropout=0.0, val=True, split=val, save_json=False, save_hybrid=False, conf=None, iou=0.7, max_det=300, half=False, dnn=False, plots=True, source=None, vid_stride=1, stream_buffer=False, visualize=False, augment=False, agnostic_nms=False, classes=None, retina_masks=False, embed=None, show=False, save_frames=False, save_txt=False, save_conf=False, save_crop=False, show_labels=True, show_conf=True, show_boxes=True, line_width=None, format=torchscript, keras=False, optimize=False, int8=False, dynamic=False, simplify=False, opset=None, workspace=4, nms=False, lr0=0.01, lrf=0.01, momentum=0.937, weight_decay=0.0005, warmup_epochs=3.0, warmup_momentum=0.8, warmup_bias_lr=0.1, box=7.5, cls=0.5, dfl=1.5, pose=12.0, kobj=1.0, label_smoothing=0.0, nbs=64, hsv_h=0.015, hsv_s=0.7, hsv_v=0.4, degrees=0.0, translate=0.1, scale=0.5, shear=0.0, perspective=0.0, flipud=0.0, fliplr=0.5, bgr=0.0, mosaic=1.0, mixup=0.0, copy_paste=0.0, auto_augment=randaugment, erasing=0.4, crop_fraction=1.0, cfg=None, tracker=botsort.yaml, save_dir=f:\ai\yolov\v10\yolov10\runs\detect\train6

                   from  n    params  module                                       arguments                    

  0                  -1  1       464  ultralytics.nn.modules.conv.Conv             [3, 16, 3, 2]                

  1                  -1  1      4672  ultralytics.nn.modules.conv.Conv             [16, 32, 3, 2]               

  2                  -1  1      7360  ultralytics.nn.modules.block.C2f             [32, 32, 1, True]            

  3                  -1  1     18560  ultralytics.nn.modules.conv.Conv             [32, 64, 3, 2]               

  4                  -1  2     49664  ultralytics.nn.modules.block.C2f             [64, 64, 2, True]            

  5                  -1  1      9856  ultralytics.nn.modules.block.SCDown          [64, 128, 3, 2]              

  6                  -1  2    197632  ultralytics.nn.modules.block.C2f             [128, 128, 2, True]          

  7                  -1  1     36096  ultralytics.nn.modules.block.SCDown          [128, 256, 3, 2]             

  8                  -1  1    460288  ultralytics.nn.modules.block.C2f             [256, 256, 1, True]          

  9                  -1  1    164608  ultralytics.nn.modules.block.SPPF            [256, 256, 5]                

 10                  -1  1    249728  ultralytics.nn.modules.block.PSA             [256, 256]                   

 11                  -1  1         0  torch.nn.modules.upsampling.Upsample         [None, 2, 'nearest']         

 12             [-1, 6]  1         0  ultralytics.nn.modules.conv.Concat           [1]                          

 13                  -1  1    148224  ultralytics.nn.modules.block.C2f             [384, 128, 1]                

 14                  -1  1         0  torch.nn.modules.upsampling.Upsample         [None, 2, 'nearest']         

 15             [-1, 4]  1         0  ultralytics.nn.modules.conv.Concat           [1]                          

 16                  -1  1     37248  ultralytics.nn.modules.block.C2f             [192, 64, 1]                 

 17                  -1  1     36992  ultralytics.nn.modules.conv.Conv             [64, 64, 3, 2]               

 18            [-1, 13]  1         0  ultralytics.nn.modules.conv.Concat           [1]                          

 19                  -1  1    123648  ultralytics.nn.modules.block.C2f             [192, 128, 1]                

 20                  -1  1     18048  ultralytics.nn.modules.block.SCDown          [128, 128, 3, 2]             

 21            [-1, 10]  1         0  ultralytics.nn.modules.conv.Concat           [1]                          

 22                  -1  1    282624  ultralytics.nn.modules.block.C2fCIB          [384, 256, 1, True, True]    

 23        [16, 19, 22]  1    929808  ultralytics.nn.modules.head.v10Detect        [80, [64, 128, 256]]         

YOLOv10n summary: 385 layers, 2775520 parameters, 2775504 gradients, 8.7 GFLOPs

Freezing layer 'model.23.dfl.conv.weight'
AMP: running Automatic Mixed Precision (AMP) checks with YOLOv8n...
f:\ai\yolov\v10\yolov10\ultralytics\utils\checks.py:641: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
  with torch.cuda.amp.autocast(True):
AMP: checks passed ✅
f:\ai\yolov\v10\yolov10\ultralytics\engine\trainer.py:276: FutureWarning: `torch.cuda.amp.GradScaler(args...)` is deprecated. Please use `torch.amp.GradScaler('cuda', args...)` instead.
  self.scaler = torch.cuda.amp.GradScaler(enabled=self.amp)
train: Scanning F:\ai\yolov\v10\datasets\coco128\labels\train2017.cache... 126 images, 2 backgrounds, 0 corrupt
val: Scanning F:\ai\yolov\v10\datasets\coco128\labels\train2017.cache... 126 images, 2 backgrounds, 0 corrupt: 
Plotting labels to f:\ai\yolov\v10\yolov10\runs\detect\train6\labels.jpg... 
optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically...
optimizer: AdamW(lr=0.000119, momentum=0.9) with parameter groups 95 weight(decay=0.0), 108 weight(decay=0.0005), 107 bias(decay=0.0)
Image sizes 640 train, 640 val
Using 8 dataloader workers
Logging results to f:\ai\yolov\v10\yolov10\runs\detect\train6
Starting training for 100 epochs...

      Epoch    GPU_mem     box_om     cls_om     dfl_om     box_oo     cls_oo     dfl_oo  Instances       Size  
      1/100      6.73G        3.5      5.731      4.271      3.308      7.837      4.339        536        640:
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 2/2  
                   all        128        929   7.78e-07   0.000805   7.04e-07   2.76e-07

      Epoch    GPU_mem     box_om     cls_om     dfl_om     box_oo     cls_oo     dfl_oo  Instances       Size  
      2/100      6.65G      3.605      5.761      4.263      3.433      8.211      4.346        364        640:
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 2/2  
                   all        128        929    7.6e-07   0.000805   1.24e-06   3.95e-07

      Epoch    GPU_mem     box_om     cls_om     dfl_om     box_oo     cls_oo     dfl_oo  Instances       Size  
      3/100      7.08G      3.537      5.722       4.25      3.357      7.648      4.332        427        640:
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 2/2  
                   all        128        929   7.59e-07   0.000805   9.74e-07   3.26e-07

      Epoch    GPU_mem     box_om     cls_om     dfl_om     box_oo     cls_oo     dfl_oo  Instances       Size  
      4/100      6.73G      3.572      5.705      4.269      3.393      7.635      4.367        435        640:
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 2/2  
                   all        128        929    7.4e-07   0.000805   9.34e-07    3.5e-07

      Epoch    GPU_mem     box_om     cls_om     dfl_om     box_oo     cls_oo     dfl_oo  Instances       Size  
      5/100      6.98G      3.558      5.717      4.249      3.356       7.72      4.327        499        640:
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 2/2  
                   all        128        929    7.4e-07   0.000805   1.08e-06   3.56e-07

      Epoch    GPU_mem     box_om     cls_om     dfl_om     box_oo     cls_oo     dfl_oo  Instances       Size  
      6/100      6.83G      3.654      5.714      4.261      3.457      7.791      4.325        416        640:
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 2/2  
                   all        128        929    7.4e-07   0.000805   1.02e-06   3.75e-07

      Epoch    GPU_mem     box_om     cls_om     dfl_om     box_oo     cls_oo     dfl_oo  Instances       Size  
      7/100      6.73G      3.604      5.726      4.237       3.39      7.656      4.307        561        640:
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 2/2  
                   all        128        929    7.4e-07   0.000805   9.16e-07   3.45e-07

      Epoch    GPU_mem     box_om     cls_om     dfl_om     box_oo     cls_oo     dfl_oo  Instances       Size  
      8/100      7.08G      3.567      5.734      4.231      3.418      7.774      4.283        535        640:
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 2/2  
                   all        128        929   7.59e-07   0.000805   1.31e-06   4.26e-07

      Epoch    GPU_mem     box_om     cls_om     dfl_om     box_oo     cls_oo     dfl_oo  Instances       Size  
      9/100      6.73G      3.594       5.73      4.228      3.452      7.902      4.284        353        640:
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 2/2  
                   all        128        929   7.76e-07   0.000805   1.34e-06    4.3e-07

      Epoch    GPU_mem     box_om     cls_om     dfl_om     box_oo     cls_oo     dfl_oo  Instances       Siz