【多人姿态估计】Alphapose_yolov8复现

1.环境

参考此处安装docker 、nvidia-docker

https://blog.csdn.net/qq_35975447/article/details/113248132

然后导入docker image

https://blog.csdn.net/qq_35975447/article/details/108662743

sudo docker load -i your_image_name.tar

 创建容器：

#!/bin/bash
nvidia-docker run -p 10001:22 -m 8GB -it --ipc=host --net=host \
		-v /your_local_path/:/your_env_path \
		-v /tmp/.X11-unix:/tmp/.X11-unix \
		-e DISPLAY=unix$DISPLAY \
		-e GDK_SCALE \
		-e GDK_DPI_SCALE \
		--name your_container_name your_images_name:your_images_version /bin/bash
 

2.解压代码

AlphaPose-Yolov8.zip

3.安装Alphapose

cd Alphapose_Yolov8

python setup.py build develop --user

4.运行demo

python scripts/demo_inference.py --cfg configs/coco/resnet/256x192_res50_lr1e-3_2x-dcn.yaml --checkpoint pretrained_models/fast_dcn_res50_256x192.pth --indir examples/demo/ --vis --showbox --save_img --pose_track --sp --vis_fast --detector yolov8

新开一个窗口：

xhost +

结果在AlphaPose-Yolov8/examples/res/vis目录下：