Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/Rane2021/yolov5_train_cpp_inference
yolov5训练和c++推理代码,效果出色
https://github.com/Rane2021/yolov5_train_cpp_inference
Last synced: 12 days ago
JSON representation
yolov5训练和c++推理代码,效果出色
- Host: GitHub
- URL: https://github.com/Rane2021/yolov5_train_cpp_inference
- Owner: Rane2021
- Created: 2022-06-29T01:44:28.000Z (almost 2 years ago)
- Default Branch: master
- Last Pushed: 2023-12-09T00:23:13.000Z (7 months ago)
- Last Synced: 2024-02-29T05:35:23.318Z (4 months ago)
- Language: Python
- Size: 269 KB
- Stars: 5
- Watchers: 1
- Forks: 1
- Open Issues: 4
-
Metadata Files:
- Readme: README.md
Lists
- awesome-yolo-object-detection - Rane2021/yolov5_train_cpp_inference
README
# requirements
目前使用 source activate pytorch1.8#### update 20231206
```
# 修复由于版本更新出现的库不兼容问题
pip install -r requirement.txt
```# 检查错误的图像:
使用:python 02_val.py
修改地方:
```apl
1、data/04_dumei_coco128_test.yaml / 03_fanyingshi_coco128_test.yaml / 05_chumuqi_coco128_test.yaml 中的 test 图像路径
这里还要注意下类别,有的修改过好多次2、修改 02_val.py line 346: --data && --weights
3、python 02_val.py
```# 更新训练集
1、通过验证整理有问题的图像,整理出来,通过 10_coco_label_tool 重新标注数据;
辅助脚本:runs/vip_train_2023-3/01_get_tesed_data.py
2、将新数据并入最近的一版本模型训练数据集中;# 重新训练模型
- train chumuqi:
```shell
python 01_train.py --data data/05_chumuqi_coco128.yaml --weights weights/yolov5s.pt --batch-size 16 --epochs 200 --imgsz 640
```- train dumei:
```shell
python 01_train.py --data data/04_dumei_coco128.yaml --weights weights/yolov5s.pt --batch-size 16 --epochs 200 --imgsz 640 --single-cls```
- train fangyingshi:
```shell```
# 导出模型
```shell
python 03_export.py --weights runs/vip_train_2023-3/exp_2023-3-24_dumei/weights/best.pt
```