1.GitHub 下载yolov5

2.环境：win10，visual studio 2015 ，anaconda2.7，python=3.7，cuda=10.2 ,pytorch=1.5(GPU)

3.更改镜像-清华

4.安装pytorch

5.安装requirements.txt:报错。解决方案：

git+https://github.com/philferriere/cocoapi.git#subdirectory=PythonAPI

用这个替换掉原有语句

6.下载预训练模型，YOLOv5l.pt。

7.直接运行detect文件，报错。解决方案：

1.python detect.py --source ./inference/images/bus.jpg --weights ./weights/yolov5s.pt --conf 0.4 --device cuda:0

2.这里报错：找不到cuda

3.进入到torch_utils.py,

def select_device(device='', apex=False, batch_size=None):

# device = 'cpu' or '0' or '0,1,2,3'

cpu_request = device.lower() == 'cpu'

if device and not cpu_request: # if device requested other than 'cpu'

os.environ['CUDA_VISIBLE_DEVICES'] = device # set environment variable

assert torch.cuda.is_available(), 'CUDA unavailable, invalid device %s requested' % device # check availablity

.

.

.

加入一个语句

def select_device(device='', apex=False, batch_size=None):

# device = 'cpu' or '0' or '0,1,2,3'

print(torch.cuda.is_available())

cpu_request = device.lower() == 'cpu'

if device and not cpu_request: # if device requested other than 'cpu'

os.environ['CUDA_VISIBLE_DEVICES'] = device # set environment variable

assert torch.cuda.is_available(), 'CUDA unavailable, invalid device %s requested' % device # check availablity

这样就可以找到cuda

于是得到结果：

接下来，是制作自己的小型数据集，训练和测试

0.下载coco2017，手动选40张图片和对应标签。

1.目录结构：

##数据集和代码平行目录

yolo的格式是 xxx.jpg和对应的 xxx.txt,共2个文件

txt中存放的是标注好的人、狗、摩托车的坐标；还有所属类别的序号

2.自制yaml文件

train: ../lee/train/images

val: ../lee/valid/images

test: ../lee/test/images

# number of classes

nc: 80

# class names

names: ['person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus', 'train', 'truck', 'boat', 'traffic light',

'fire hydrant', 'stop sign', 'parking meter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow',

'elephant', 'bear', 'zebra', 'giraffe', 'backpack', 'umbrella', 'handbag', 'tie', 'suitcase', 'frisbee',

'skis', 'snowboard', 'sports ball', 'kite', 'baseball bat', 'baseball glove', 'skateboard', 'surfboard',

'tennis racket', 'bottle', 'wine glass', 'cup', 'fork', 'knife', 'spoon', 'bowl', 'banana', 'apple',

'sandwich', 'orange', 'broccoli', 'carrot', 'hot dog', 'pizza', 'donut', 'cake', 'chair', 'couch',

'potted plant', 'bed', 'dining table', 'toilet', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cell phone',

'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'book', 'clock', 'vase', 'scissors', 'teddy bear',

'hair drier', 'toothbrush']

运行指令：

python train.py --img 640 --batch 16 --epochs 10 --data ./data/lee.yaml --cfg ./models/yolov5s.yaml --device cuda:0

结果如下

(yolov5_py37_torch1.5) D:\yolo\yolov5-master>python train.py --img 640 --batch 16 --epochs 10 --data ./data/lee.yaml --cfg ./models/yolov5s.yaml --device cuda:0

Namespace(batch_size=16, bucket='', cache_images=False, cfg='./models/yolov5s.yaml', data='./data/lee.yaml', device='cuda:0', epochs=10, evolve=False, hyp='', img_size=[640], multi_scale=F

alse, name='', noautoanchor=False, nosave=False, notest=False, rect=False, resume=False, single_cls=False, weights='')

True

Using CUDA Apex device0 _CudaDeviceProperties(name='GeForce RTX 2060 SUPER', total_memory=8192MB)

Start Tensorboard with "tensorboard --logdir=runs", view at http://localhost:6006/

Hyperparameters {'optimizer': 'SGD', 'lr0': 0.01, 'momentum': 0.937, 'weight_decay': 0.0005, 'giou': 0.05, 'cls': 0.58, 'cls_pw': 1.0, 'obj': 1.0, 'obj_pw': 1.0, 'iou_t': 0.2, 'anchor_t':

4.0, 'fl_gamma': 0.0, 'hsv_h': 0.014, 'hsv_s': 0.68, 'hsv_v': 0.36, 'degrees': 0.0, 'translate': 0.0, 'scale': 0.5, 'shear': 0.0}

from n params module arguments

0 -1 1 3520 models.common.Focus [3, 32, 3]

1 -1 1 18560 models.common.Conv [32, 64, 3, 2]

2 -1 1 19904 models.common.BottleneckCSP [64, 64, 1]

3 -1 1 73984 models.common.Conv [64, 128, 3, 2]

4 -1 1 161152 models.common.BottleneckCSP [128, 128, 3]

5 -1 1 295424 models.common.Conv [128, 256, 3, 2]

6 -1 1 641792 models.common.BottleneckCSP [256, 256, 3]

7 -1 1 1180672 models.common.Conv [256, 512, 3, 2]

8 -1 1 656896 models.common.SPP [512, 512, [5, 9, 13]]

9 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False]

10 -1 1 131584 models.common.Conv [512, 256, 1, 1]

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

12 [-1, 6] 1 0 models.common.Concat [1]

13 -1 1 378624 models.common.BottleneckCSP [512, 256, 1, False]

14 -1 1 33024 models.common.Conv [256, 128, 1, 1]

15 -1 1 0 torch.nn.modules.upsampling.Upsample [None, 2, 'nearest']

16 [-1, 4] 1 0 models.common.Concat [1]

17 -1 1 95104 models.common.BottleneckCSP [256, 128, 1, False]

18 -1 1 32895 torch.nn.modules.conv.Conv2d [128, 255, 1, 1]

19 -2 1 147712 models.common.Conv [128, 128, 3, 2]

20 [-1, 14] 1 0 models.common.Concat [1]

21 -1 1 313088 models.common.BottleneckCSP [256, 256, 1, False]

22 -1 1 65535 torch.nn.modules.conv.Conv2d [256, 255, 1, 1]

23 -2 1 590336 models.common.Conv [256, 256, 3, 2]

24 [-1, 10] 1 0 models.common.Concat [1]

25 -1 1 1248768 models.common.BottleneckCSP [512, 512, 1, False]

26 -1 1 130815 torch.nn.modules.conv.Conv2d [512, 255, 1, 1]

27 [-1, 22, 18] 1 0 models.yolo.Detect [80, [[116, 90, 156, 198, 373, 326], [30, 61, 62, 45, 59, 119], [10, 13, 16, 30, 33, 23]]]

Model Summary: 191 layers, 7.46816e+06 parameters, 7.46816e+06 gradients

Optimizer groups: 62 .bias, 70 conv.weight, 59 other

Scanning labels ..\lee\train\labels.cache (39 found, 0 missing, 0 empty, 0 duplicate, for 39 images): 100%|████████████████████████████████████████

█████| 39/39 [00:00<00:00, 19317.18it/s]

Scanning labels ..\lee\valid\labels.cache (29 found, 0 missing, 0 empty, 0 duplicate, for 29 images): 100%|████████████████████████████████████████

█████| 29/29 [00:00<00:00, 29281.37it/s]

Analyzing anchors... Best Possible Recall (BPR) = 0.9935

Image sizes 640 train, 640 test

Using 8 dataloader workers

Starting training for 10 epochs...

Epoch gpu_mem GIoU obj cls total targets img_size

0/9 2.93G 0.1186 0.08896 0.1206 0.3281 242 640: 33%|███████████████████████▎ | 1/3 [

0/9 2.95G 0.1167 0.08447 0.1214 0.3225 161 640: 33%|███████████████████████▎ | 1/3 [

0/9 2.95G 0.1167 0.08447 0.1214 0.3225 161 640: 67%|██████████████████████████████████████████████▋

0/9 2.95G 0.1169 0.08754 0.121 0.3255 105 640: 67%|██████████████████████████████████████████████▋

0/9 2.95G 0.1169 0.08754 0.121 0.3255 105 640: 100%|██████████████████████████████████████████████████

0/9 2.95G 0.1169 0.08754 0.121 0.3255 105 640: 100%|██████████████████████████████████████████████████

████████████████████| 3/3 [00:04<00:00, 1.61s/it]

注意：下载的数据集中，有几张图片的txt文件有问题，训练时会报错

比如：KeyError: ‘…\lee\train\images\000000000009.jpg’

可以直接删除相应的图片和txt文件

然后，会在这个weights文件夹中，生成新的xxx.pt权重文件

win10这步会有问题，由于代码是针对Linux写的。

打开train.py文件

'runs/exp' //这个是Linux的目录结构，需要改成win10的

# Train

if not opt.evolve:

tb_writer = SummaryWriter(log_dir=increment_dir('runs/exp', opt.name))

print('Start Tensorboard with "tensorboard --logdir=runs", view at http://localhost:6006/')

train(hyp)

改成：

# Train

if not opt.evolve:

tb_writer = SummaryWriter(log_dir=increment_dir('..\\yolo\\yolov5-master\\runs\\exp', opt.name))

print('Start Tensorboard with "tensorboard --logdir=runs", view at http://localhost:6006/')

train(hyp)

最后是训练完成后，用自己的权重测试图片

命令：

python detect.py --source ../lee/test/images --weights ./weights/myweights.pt

(yolov5_py37_torch1.5) D:\yolo\yolov5-master>python detect.py --source ../lee/test/images --device cuda:0 --weights ./weights/yolov5s.pt

Namespace(agnostic_nms=False, augment=False, classes=None, conf_thres=0.4, device='cuda:0', img_size=640, iou_thres=0.5, output='inference/output', save_txt=False, source='../lee/test/images', update=False, vi

ew_img=False, weights=['./weights/myweights.pt'])

True

Using CUDA device0 _CudaDeviceProperties(name='GeForce RTX 2060 SUPER', total_memory=8192MB)

Fusing layers... Model Summary: 140 layers, 7.45958e+06 parameters, 7.45958e+06 gradients

image 1/29 D:\yolo\lee\test\images\000000000389.jpg: 512x640 12 persons, 2 ties, Done. (0.034s)

image 2/29 D:\yolo\lee\test\images\000000000394.jpg: 640x640 1 frisbees, Done. (0.017s)

image 3/29 D:\yolo\lee\test\images\000000000395.jpg: 640x640 7 persons, 1 cell phones, Done. (0.015s)

image 4/29 D:\yolo\lee\test\images\000000000397.jpg: 512x640 1 persons, 1 pizzas, 1 chairs, Done. (0.017s)

image 5/29 D:\yolo\lee\test\images\000000000400.jpg: 640x640 1 dogs, Done. (0.015s)

image 6/29 D:\yolo\lee\test\images\000000000404.jpg: 640x448 1 boats, Done. (0.017s)

image 7/29 D:\yolo\lee\test\images\000000000415.jpg: 640x384 1 persons, 1 tennis rackets, Done. (0.016s)

image 8/29 D:\yolo\lee\test\images\000000000419.jpg: 512x640 2 persons, 2 cars, 1 tennis rackets, Done. (0.017s)

image 9/29 D:\yolo\lee\test\images\000000000428.jpg: 384x640 1 persons, 1 cakes, Done. (0.017s)

image 10/29 D:\yolo\lee\test\images\000000000431.jpg: 448x640 1 persons, 1 sports balls, 1 tennis rackets, Done. (0.016s)

image 11/29 D:\yolo\lee\test\images\000000000436.jpg: 640x448 1 persons, 1 bottles, 1 donuts, Done. (0.016s)

image 12/29 D:\yolo\lee\test\images\000000000438.jpg: 512x640 17 donuts, Done. (0.016s)

image 13/29 D:\yolo\lee\test\images\000000000443.jpg: 512x640 1 remotes, Done. (0.016s)

image 14/29 D:\yolo\lee\test\images\000000000446.jpg: 640x512 1 persons, 1 couchs, 1 potted plants, Done. (0.016s)

image 15/29 D:\yolo\lee\test\images\000000000450.jpg: 512x640 1 wine glasss, 3 cups, 1 pizzas, Done. (0.017s)

image 16/29 D:\yolo\lee\test\images\000000000459.jpg: 640x576 1 persons, 1 ties, Done. (0.018s)

image 17/29 D:\yolo\lee\test\images\000000000471.jpg: 448x640 1 buss, Done. (0.016s)

image 18/29 D:\yolo\lee\test\images\000000000472.jpg: 256x640 1 airplanes, Done. (0.016s)

image 19/29 D:\yolo\lee\test\images\000000000474.jpg: 640x448 1 persons, 1 baseball gloves, Done. (0.016s)

image 20/29 D:\yolo\lee\test\images\000000000486.jpg: 448x640 1 bowls, 1 refrigerators, Done. (0.016s)

image 21/29 D:\yolo\lee\test\images\000000000488.jpg: 448x640 4 persons, 1 sports balls, 3 baseball gloves, Done. (0.015s)

image 22/29 D:\yolo\lee\test\images\000000000490.jpg: 640x640 1 dogs, 1 skateboards, 1 chairs, Done. (0.015s)

image 23/29 D:\yolo\lee\test\images\000000000491.jpg: 448x640 1 birds, 1 teddy bears, Done. (0.016s)

image 24/29 D:\yolo\lee\test\images\000000000502.jpg: 448x640 1 bears, Done. (0.015s)

image 25/29 D:\yolo\lee\test\images\000000000510.jpg: 512x640 1 persons, 1 benchs, Done. (0.016s)

image 26/29 D:\yolo\lee\test\images\000000000514.jpg: 640x384 1 beds, Done. (0.016s)

image 27/29 D:\yolo\lee\test\images\000000000520.jpg: 512x640 2 persons, 7 birds, Done. (0.017s)

image 28/29 D:\yolo\lee\test\images\000000000529.jpg: 640x448 1 persons, 1 motorcycles, Done. (0.015s)

image 29/29 D:\yolo\lee\test\images\000000000531.jpg: 512x640 9 persons, 1 tennis rackets, Done. (0.016s)

Results saved to D:\yolo\yolov5-master\inference/output

Done. (1.946s)

可以在…\yolov5-master\inference\output中看到

来源：oschina

链接：https://my.oschina.net/u/4391448/blog/4404703