yolov7-keypoint-customization

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

yolov7-keypoint-customization

(★ 11)

Revision of official yolov7-pose to support custom dataset for keypoint detection

yolov7-keypoint-customization 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# yolov7-keypoint-customization
Revision of official yolov7-pose to support custom dataset for keypoint detection. 
## Statement
"yolov7-keypoint-customization" is for convinient use of keypoint detection. The revision that added custom keypoint support was based on the official yolov7-pose project. 
See https://github.com/WongKinYiu/yolov7 for full official project please.
## Get started
### Configuration
Before training, you need to revise the configuration in ```./data/custom.yaml``` and ```./cfg/yolov7-w6-pose-custom.yaml```. 
In ```./data/custom.yaml```, you need to edit your dataset path, class number and class label as follows: 
```
train: ../root_path/your_traget_path/yolo_labels/train.txt 
val: ../root_path/your_traget_path/yolo_labels/val.txt 
test: ../root_path/your_traget_path/yolo_labels/val.txt

nc: 1 #num of class 
names: [ 'custom'] # label of each class 
```
And the train.txt/val.txt should be filled with corresponding image path (Assumed you have completed the preparation of dataset). 
In ```./cfg/yolov7-w6-pose-custom.yaml```, just need to revise the nkpt (number of keypoint in each sample) and nc (class_num). 
### Running
Now let's start to train your custom yolov7-keypoint model. The pre-training weight can be obtained in official page, see https://github.com/WongKinYiu/yolov7/tree/pose 
#### DP mode
```
python train_keypoint.py --weights ./yolov7-w6-person.pt --data data/custom.yaml --device 0,1,2,3 --hyp data/hyp.pose.yaml --cfg cfg/yolov7-w6-pose-custom.yaml --name yolov7-keypoint --kpt-label [--epoch] [--batch] [--linear-lr]
```
#### DDP mode
```
python -m torch.distributed.launch --nproc_per_node 4 --master_port 9527 train_keypoint.py --workers 8 --device 0,1,2,3 --sync-bn --batch-size 128 --data data/custom.yaml --img 640 640 --cfg cfg/yolov7-w6-pose-custom.yaml --weights '' --name yolov7-keypoint --hyp data/hyp.pose.yaml
```
#### Detecting
```
python detect.py --weights ./[your_best].pt --source [your_testset_path] --kpt-label
```
## Acknowledgements
https://github.com/WongKinYiu/yolov7/tree/pose 
https://github.com/WongKinYiu/yolov7 
## Further
Any problems about this, welcome to communicate with me. 
Email: chenjisen123@foxmail.com

yolov7-keypoint-customization

yolov7-keypoint-customization 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

yolov7-keypoint-customization

yolov7-keypoint-customization 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README