论文复现--RMPE: Regional Multi-person Pose Estimation
paper code home video
目录
paper code home video
Step1.创建新的nvidia-docker container名为MotionCatch
Step2.在此docker下安装anaconda3
出现问题1：提示网络连接失败
Checking Internet connectivity,Please make sure you are connected to the internet
Step3.按照安装官方文档在MotionCatch进行配置
1. 创建conda环境
出现问题2：
You may need to closed and restart your shell after running 'code init'
2. 安装PyTorch
3. 获取AlphaPose源码
4. 安装
出现问题3:Subprocess.CalledProcessedError:Command '[Which','g++']'returned non-zero exit status 1.
出现问题4:error No such file or directory :':usr/local/cuda:/usr/local/cuda/bin/nvcc'
出现问题5：
.../compiler_compat/id:cannot find -LOSMesa : No such file or directory
.../compiler_compat/id:cannot find -LGL : No such file or directory
.../compiler_compat/id:cannot find -LGLU : No such file or directory
出现问题6：Setup script exited with
Beginning with Matplotlib 3.6 ,Python 3.8 above is required
出现问题7:Import error:cannot import name 'get_installed_dostribution' from 'pip_internal.utils.misc'(/home/anaconda3/envs/alphapose/lib/python3.7/site-packages/pip/_internal/utils/misc.py)
Step4.目标检测模型配置
1.下载yolov3-spp.weights
2.下载这个模型halpe26_fast_res50_256x192.pth
出现问题8：
3.下载第二个模型halpe136_fast50_regression_256x192.pth
4.下载前两个multi_domain_fast50_dcn_combined_256x192.pt和 multi_domain_fast50_regression_256x192.pth
Step5.姿态追踪模型配置
1首先尝试第一个模型 下载这个
2 尝试第二个模型 下载第一个
Conda list
Step1.创建新的nvidia-docker container名为MotionCatch
nvidia-docker run -it --name MotionCatch nvidia/cuda:11.0.3-base-ubuntu18.04 /bin/bashStep2.在此docker下安装anaconda3
下载版本为Anaconda3-5.3.0-Linux-x86_64.sh 安装在MotionCatch内
docker inspect --format="{{.Id}}" MotionCatch # 查询该容器的container_id
docker cp /home/sqy/Software_Anzhuang/Anaconda3-5.3.0-Linux-x86_64.sh 24520982511f3b8d72adea23a0113817c49d8c991ade2edb3c6d6fa43305ca5e:/home# 移动到home下
bash Anaconda3-5.3.0-Linux-x86_64.sh出现问题1：提示网络连接失败
Checking Internet connectivity,Please make sure you are connected to the internet
解决方法:这是vscode相关的，不影响anaconda 直接输入no
安装结束，配置环境变量
apt-get update
apt-get upgrade
apt-get install vim
vi ~/.bashrc
export export PATH="/home/anaconda3/bin:$PATH"# conda init里有这条指令，复制下来添加到末尾,路径为安装路径
source ~/.bashrc检查安装是否成功
conda -V #显示conda 4.5.11注意，conda卸载只需要把anaconda3文件夹整个删除，再把bashrc里刚才添加的export语句删除即可
(Tips:中间我看了下磁盘空间，觉得snapd不爽 卸载了snapd，虽然这和论文没关系：)
sudo apt-get purge snapd
sudo apt autoremove --purge snapd
Step3.按照安装官方文档在MotionCatch进行配置
1. 创建conda环境
conda create -n alphapose python=3.7 -y
conda activate alphapose出现问题2：
You may need to closed and restart your shell after running 'code init'
原因：shell未激活,执行conda init是没用的
解决办法： source activate
(Tips：执行conda update conda后若出现conda info -e出错的情况
An unexpected error has occurred.conda has prepared the above report)
解决办法：把anaconda里的lib文件夹 libffi.so.7文件进行软连接更改 ln -s libffi.so.6 libffi.so.7即可
2. 安装PyTorch
conda install pytorch torchvision cudatoolkit=11.3 -c pytorch3. 获取AlphaPose源码
git clone https://github.com/MVIG-SJTU/AlphaPose.git
cd AlphaPose4. 安装
# 以下两条环境变量在宿主机和docker bashrc设置,添加完毕后生效
vi ~/.bashrc
export PATH="/usr/local/cuda/bin/:$PATH"
export LD_LIBRARY_PATH="/usr/local/cuda/lib64/:$LD_LIBRARY_PATH"
source ~/.bashrc
# 以下两条在docker安装
python -m pip install cython -i https://pypi.tuna.tsinghua.edu.cn/simple
sudo apt-get install libyaml-dev
# 仅在Ubuntu18.04进行此设置
locale-gen C.UTF-8
# 如果locale-gen指令未被查到
sudo apt-get install locales
export LANG=C.UTF-8
# 最后执行
python setup.py build develop
出现问题3:Subprocess.CalledProcessedError:Command '[Which','g++']'returned non-zero exit status 1.
出错原因：未安装g++
解决办法：apt-get install g++
出现问题4:error No such file or directory :':usr/local/cuda:/usr/local/cuda/bin/nvcc'
解决办法：在docker安装CUDA后，并执行export CUDA_HOME=/usr/local/cuda。重新执行python setup.py build develop
出现问题5：
.../compiler_compat/id:cannot find -LOSMesa : No such file or directory
.../compiler_compat/id:cannot find -LGL : No such file or directory
.../compiler_compat/id:cannot find -LGLU : No such file or directory
解决办法：安装缺失的三个动态库 并设置为当前虚拟环境的软连接
apt install libglu1-mesa
apt install -y libosmesa-dev
ln -s /usr/lib/x86_64-linux-gnu/libGL.so.1 /home/anaconda3/envs/alphapose/lib/libGL.so
ln -s /usr/lib/x86_64-linux-gnu/libGLU.so.1 /home/anaconda3/envs/alphapose/lib/libGLU.so
ln -s /usr/lib/x86_64-linux-gnu/libGLX_mesa.so.0.0.0 /home/anaconda3/envs/alphapose/lib/libOSMesa.so注意：rm -rf build后再执行 python setup.py build develop
出现问题6：Setup script exited with
Beginning with Matplotlib 3.6 ,Python 3.8 above is required
解决办法：缺什么安装什么，其他错误也是一样（注意加上清华源）
pip install matplotlib -i https://pypi.tuna.tsinghua.edu.cn/simple出现问题7:Import error:cannot import name 'get_installed_dostribution' from 'pip_internal.utils.misc'(/home/anaconda3/envs/alphapose/lib/python3.7/site-packages/pip/_internal/utils/misc.py)
解决办法：降级pip
python -m pip install pip==21.2 -i https://pypi.tuna.tsinghua.edu.cn/simple
好用的pip下载指令
pip install -i https://pypi.tuna.tsinghua.edu.cn/simple --trusted-host pypi.tuna.tsinghua.edu.cn numpy 编译完成： 
Step4.目标检测模型配置
1.下载yolov3-spp.weights
和 yolox-l
分别拷贝到detector/yolo/data和detector/yolox/data(需要自己创建一个data文件夹)
docker cp yolov3-spp.weights 24520982511f3b8d72adea23a0113817c49d8c991ade2edb3c6d6fa43305ca5e:/root/AlphaPose/detector/yolo/data
docker cp yolox_x.pth 24520982511f3b8d72adea23a0113817c49d8c991ade2edb3c6d6fa43305ca5e:/root/AlphaPose/detector/yolox/data
原始数据
2.下载这个模型halpe26_fast_res50_256x192.pth
(2 3 4的模型都放到/pretrained_models下)
docker cp halpe26_fast_res50_256x192.pth 24520982511f3b8d72adea23a0113817c49d8c991ade2edb3c6d6fa43305ca5e:/root/AlphaPose/pretrained_models
python scripts/demo_inference.py --cfg configs/halpe_26/resnet/256x192_res50_lr1e-3_1x.yaml --checkpoint pretrained_models/halpe26_fast_res50_256x192.pth --indir examples/demo/ --save_img(想用yolox-l做检测器，请在上述指令加上此参数：--detector yolox-x )
出现问题8：
ImportError: libgthread-2.0.so.0: cannot open shared object file: No such file or directory
解决方法：
apt-get install libglib2.0-0直接跑通
docker cp 24520982511f3b8d72adea23a0113817c49d8c991ade2edb3c6d6fa43305ca5e:/root/AlphaPose/examples/res/vis ./
查看结果：
3.下载第二个模型halpe136_fast50_regression_256x192.pth
docker cp halpe136_fast50_regression_256x192.pth 24520982511f3b8d72adea23a0113817c49d8c991ade2edb3c6d6fa43305ca5e:/root/AlphaPose/pretrained_models
python scripts/demo_inference.py --cfg configs/halpe_136/resnet/256x192_res50_lr1e-3_2x-regression.yaml --checkpoint pretrained_models/halpe136_fast50_regression_256x192.pth --indir examples/demo/ --save_img查看结果
4.下载前两个multi_domain_fast50_dcn_combined_256x192.pt和 multi_domain_fast50_regression_256x192.pth
docker cp multi_domain_fast50_dcn_combined_256x192.pth 24520982511f3b8d72adea23a0113817c49d8c991ade2edb3c6d6fa43305ca5e:/root/AlphaPose/pretrained_models
docker cp multi_domain_fast50_regression_256x192.pth 24520982511f3b8d72adea23a0113817c49d8c991ade2edb3c6d6fa43305ca5e:/root/AlphaPose/pretrained_models
python scripts/demo_inference.py --cfg configs/halpe_coco_wholebody_136/resnet/256x192_res50_lr1e-3_2x-regression.yaml --checkpoint pretrained_models/multi_domain_fast50_regression_256x192.pth --indir examples/demo/ --save_img
python scripts/demo_inference.py --cfg configs/halpe_coco_wholebody_136/resnet/256x192_res50_lr1e-3_2x-dcn-combined.yaml --checkpoint pretrained_models/multi_domain_fast50_dcn_combined_256x192.pth --indir examples/demo/ --save_img
Step5.姿态追踪模型配置
1首先尝试第一个模型 下载这个
模型移动到/root/AlphaPose/trackers/weights(自己创建weights目录)
docker cp osnet_ain_x1_0_msmt17_256x128_amsgrad_ep50_lr0.0015_coslr_b64_fb10_softmax_labsmth_flip_jitter.pth 24520982511f3b8d72adea23a0113817c49d8c991ade2edb3c6d6fa43305ca5e:/root/AlphaPose/trackers/weights
视频移动到 /root/AlphaPose/examples/video_demo
docker cp 1.mp4 24520982511f3b8d72adea23a0113817c49d8c991ade2edb3c6d6fa43305ca5e:/root/AlphaPose/examples/video_demo
(另外，可以更改/AlphaPose/trackers/tracker_cfg.py中的cfg.arch和cfg.loadmodel来更改模型)
跑模型
python scripts/demo_inference.py --cfg configs/halpe_26/resnet/256x192_res50_lr1e-3_1x.yaml --checkpoint pretrained_models/halpe26_fast_res50_256x192.pth --video examples/video_demo/1.mp4 --outdir examples/res/vis_video --save_video --pose_track出现问题9:执行卡住CUDA out of memory
可能的原因：内存不够
因为在执行之前可以执行的图片样例 出现报错
此问题尚未解决 但是可以通过宿主机重启docker重新执行之前的图片样例
2 尝试第二个模型 下载第一个
模型移动到/root/AlphaPose/detector/tracker/data(需要自己创建data)
docker cp jde.1088x608.uncertainty.pt 24520982511f3b8d72adea23a0113817c49d8c991ade2edb3c6d6fa43305ca5e:/root/AlphaPose/detector/tracker/datapython scripts/demo_inference.py --cfg configs/halpe_26/resnet/256x192_res50_lr1e-3_1x.yaml --checkpoint pretrained_models/halpe26_fast_res50_256x192.pth --video examples/video_demo/1.mp4 --outdir examples/res/vis_video --save_video --detector trackerConda list
# packages in environment at /home/anaconda3/envs/alphapose:
#
# Name Version Build Channel
_libgcc_mutex 0.1 main
_openmp_mutex 5.1 1_gnu
alphapose 0.5.0+a94dae6 dev_0
blas 1.0 mkl
brotlipy 0.7.0 py37h27cfd23_1003
bzip2 1.0.8 h7b6447c_0
ca-certificates 2022.07.19 h06a4308_0
certifi 2022.9.14 py37h06a4308_0
cffi 1.15.1 py37h74dc2b5_0
charset-normalizer 2.0.4 pyhd3eb1b0_0
chumpy 0.70 pypi_0 pypi
cryptography 37.0.1 py37h9ce1e76_0
cudatoolkit 11.3.1 h2bc3f7f_2
cycler 0.11.0 pypi_0 pypi
cython 0.29.32 pypi_0 pypi
cython-bbox 0.1.3 pypi_0 pypi
easydict 1.10 pypi_0 pypi
ffmpeg 4.3 hf484d3e_0 pytorch
fonttools 4.37.3 pypi_0 pypi
freetype 2.11.0 h70c0345_0
giflib 5.2.1 h7b6447c_0
gmp 6.2.1 h295c915_3
gnutls 3.6.15 he1e5248_0
halpecocotools 0.0.0 pypi_0 pypi
idna 3.3 pyhd3eb1b0_0
intel-openmp 2021.4.0 h06a4308_3561
jpeg 9e h7f8727e_0
jsonpatch 1.32 pypi_0 pypi
jsonpointer 2.3 pypi_0 pypi
kiwisolver 1.4.4 pypi_0 pypi
lame 3.100 h7b6447c_0
lcms2 2.12 h3be6417_0
ld_impl_linux-64 2.38 h1181459_1
lerc 3.0 h295c915_0
libdeflate 1.8 h7f8727e_5
libffi 3.3 he6710b0_2
libgcc-ng 11.2.0 h1234567_1
libgomp 11.2.0 h1234567_1
libiconv 1.16 h7f8727e_2
libidn2 2.3.2 h7f8727e_0
libpng 1.6.37 hbc83047_0
libstdcxx-ng 11.2.0 h1234567_1
libtasn1 4.16.0 h27cfd23_0
libtiff 4.4.0 hecacb30_0
libunistring 0.9.10 h27cfd23_0
libwebp 1.2.2 h55f646e_0
libwebp-base 1.2.2 h7f8727e_0
lz4-c 1.9.3 h295c915_1
matplotlib 3.5.3 pypi_0 pypi
mkl 2021.4.0 h06a4308_640
mkl-service 2.4.0 py37h7f8727e_0
mkl_fft 1.3.1 py37hd3c417c_0
mkl_random 1.2.2 py37h51133e4_0
munkres 1.1.4 pypi_0 pypi
natsort 8.2.0 pypi_0 pypi
ncurses 6.3 h5eee18b_3
nettle 3.7.3 hbbd107a_1
numpy 1.21.5 py37h6c91a56_3
numpy-base 1.21.5 py37ha15fc14_3
opencv-python 4.6.0.66 pypi_0 pypi
opendr 0.78 pypi_0 pypi
openh264 2.1.1 h4ff587b_0
openssl 1.1.1q h7f8727e_0
packaging 21.3 pypi_0 pypi
pillow 9.2.0 py37hace64e9_1
pip 21.2 pypi_0 pypi
protobuf 3.20.1 pypi_0 pypi
pycocotools 2.0.5 pypi_0 pypi
pycparser 2.21 pyhd3eb1b0_0
pyopenssl 22.0.0 pyhd3eb1b0_0
pyparsing 3.0.9 pypi_0 pypi
pysocks 1.7.1 py37_1
python 3.7.13 h12debd9_0
python-dateutil 2.8.2 pypi_0 pypi
pytorch 1.12.1 py3.7_cuda11.3_cudnn8.3.2_0 pytorch
pytorch-mutex 1.0 cuda pytorch
pyyaml 6.0 pypi_0 pypi
pyzmq 24.0.1 pypi_0 pypi
readline 8.1.2 h7f8727e_1
requests 2.28.1 py37h06a4308_0
scipy 1.7.3 pypi_0 pypi
setuptools 63.4.1 py37h06a4308_0
six 1.16.0 pyhd3eb1b0_1
sqlite 3.39.3 h5082296_0
tensorboardx 2.5.1 pypi_0 pypi
terminaltables 3.1.10 pypi_0 pypi
timm 0.1.20 pypi_0 pypi
tk 8.6.12 h1ccaba5_0
torchvision 0.13.1 py37_cu113 pytorch
tornado 6.2 pypi_0 pypi
tqdm 4.64.1 pypi_0 pypi
typing_extensions 4.3.0 py37h06a4308_0
urllib3 1.26.11 py37h06a4308_0
visdom 0.1.8.9 pypi_0 pypi
websocket-client 1.4.1 pypi_0 pypi
wheel 0.37.1 pyhd3eb1b0_0
xz 5.2.6 h5eee18b_0
zlib 1.2.12 h5eee18b_3
zstd 1.5.2 ha4553b6_0