seo 网站描述长度,西安app定制开发公司,360安全浏览器官网入口,四平网站建设服务conda的一些命令以及创建环境的基本命令可参考#xff1a;Conda环境搭建以及激活 以及 conda 本地环境常用操作 前言 这里是梳理linux上在docker中使用conda#xff0c;以配置MLD-TResNet-L-AAM模型为例。论文笔记参考#xff1a;多标签分类论文笔记 | Combining Metric Lea… conda的一些命令以及创建环境的基本命令可参考Conda环境搭建以及激活 以及 conda 本地环境常用操作 前言 这里是梳理linux上在docker中使用conda以配置MLD-TResNet-L-AAM模型为例。论文笔记参考多标签分类论文笔记 | Combining Metric Learning and Attention Heads…MLD-TResNet-L-AAM/GATAAM) 文章目录 1. 安装anaconda2. 升级conda可选3. 安装cuda在conda中切换cuda的版本在conda中切换python版本收集运行环境回退conda里面包的版本4. 创建conda环境以下均在conda中进行5. 拉取代码6. 安装环境7. 训练命令 1. 安装anaconda
wget https://mirrors.tuna.tsinghua.edu.cn/anaconda/archive/Anaconda3-2021.11-Linux-x86_64.sh
或
curl -O https://mirrors.tuna.tsinghua.edu.cn/anaconda/archive/Anaconda3-2021.11-Linux-x86_64.shbash Anaconda3-2021.11-Linux-x86_64.sh配置环境
export PATH/root/anaconda3/bin:$PATH或者写入系统命令中开启即可启动 输入gedit ~/.bashrc命令打开文件在文件结尾输入以下语句保存。
export PATH/root/anaconda3/bin:$PATH然后
source ~/.bashrc2. 升级conda可选
conda update conda3. 安装cuda
apt-get updateapt-get install gcc gsh cuda_10.2.89_440.33.01_linux.run配置环境
输入gedit ~/.bashrc命令打开文件在文件结尾输入以下语句保存。
export PATH/usr/local/cuda/bin${PATH::${PATH}}
export LD_LIBRARY_PATH/usr/local/cuda/lib64${LD_LIBRARY_PATH::${LD_LIBRARY_PATH}}然后
source ~/.bashrc验证安装的cuda
nvcc -V安装cudnn
tar zxvf cudnn-10.2-linux-x64-v7.6.5.32.tgzchmod 666 /usr/local/cuda/includecp cudnn-*-archive/include/cudnn*.h /usr/local/cuda/include
cp cudnn-*-archive/lib/libcudnn* /usr/local/cuda/lib64
chmod ar /usr/local/cuda/include/cudnn*.h /usr/local/cuda/lib64/libcudnn*旧版验证这里是旧版
cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2新版验证
cat /usr/local/cuda/include/cudnn_version.h | grep CUDNN_MAJOR -A 2在conda中切换cuda的版本
conda search cudatoolkit --infoconda install cudatoolkitXXXXX在conda中切换python版本
conda search --full --name pythonconda install python3.8.18收集运行环境
pip freeze requirements.txt回退conda里面包的版本
conda list --revision查看历史版本然后运行接数字就好了比如
conda install --rev 1以下是配置论文代码
4. 创建conda环境以下均在conda中进行
conda create --name torchreid python3.7进入conda
conda activate torchreid或者
source activate torchreid5. 拉取代码
拉取
git clone https://github.com/openvinotoolkit/deep-object-reid.git切换分支
git checkout multilabel6. 安装环境
安装cudatoolkit(一定要先安装cuda)
conda install cudatoolkit10.2安装cudnn(这里直接安装即可会安装配套cuda的cudnn)
conda install cudnn配置环境
pip install -r requirements.txt遇到问题1
Collecting inplace_abnUsing cached inplace-abn-1.1.0.tar.gz (137 kB)Preparing metadata (setup.py) ... errorerror: subprocess-exited-with-error× python setup.py egg_info did not run successfully.│ exit code: 1╰─ [6 lines of output]Traceback (most recent call last):File string, line 36, in moduleFile pip-setuptools-caller, line 34, in moduleFile /tmp/pip-install-div2jd3n/inplace-abn_810be5ed00194c44a5bcc754bf5501e0/setup.py, line 4, in moduleimport torchModuleNotFoundError: No module named torch[end of output]note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed× Encountered error while generating package metadata.
╰─ See above for output.note: This is an issue with the package mentioned above, not pip.
hint: See above for details.解决方案
pip install --upgrade setuptools
python -m pip install --upgrade pip遇到问题2
ImportError: libGL.so.1: cannot open shared object file: No such file or directory解决方案
apt install libgl1-mesa-glx7. 训练命令
python tools/main.py --config-file configs/EfficientNetV2_small_gcn.yml --gpu-num 1 custom_datasets.roots [datasets/COCO/train.json, datasets/COCO/val.json] data.save_dir ./out/遇到问题3
网络问题两个文件无法下载分别是model.safetensors和pytorch_model.bin。需要科学上网
解决方案
在本地下好
https://huggingface.co/timm/tf_efficientnetv2_s.in21k/resolve/main/model.safetensorshttps://cdn-lfs.huggingface.co/repos/1a/54/1a5499191575630110693f1105f43e325ae7f696b9f8d34db19ab1309ce0aa15/09dc7ef3e90ec4570d22ae1af1c12cbc1aff590b2c68dec1cbd781340e5a8ccc?response-content-dispositionattachment%3Bfilename*%3DUTF-8%27%27pytorch_model.bin%3Bfilename%3D%22pytorch_model.bin%22%3Bresponse-content-typeapplication%2Foctet-streamExpires1697800394PolicyeyJTdGF0ZW1lbnQiOlt7IkNvbmRpdGlvbiI6eyJEYXRlTGVzc1RoYW4iOnsiQVdTOkVwb2NoVGltZSI6MTY5NzgwMDM5NH19LCJSZXNvdXJjZSI6Imh0dHBzOi8vY2RuLWxmcy5odWdnaW5nZmFjZS5jby9yZXBvcy8xYS81NC8xYTU0OTkxOTE1NzU2MzAxMTA2OTNmMTEwNWY0M2UzMjVhZTdmNjk2YjlmOGQzNGRiMTlhYjEzMDljZTBhYTE1LzA5ZGM3ZWYzZTkwZWM0NTcwZDIyYWUxYWYxYzEyY2JjMWFmZjU5MGIyYzY4ZGVjMWNiZDc4MTM0MGU1YThjY2M%7EcmVzcG9uc2UtY29udGVudC1kaXNwb3NpdGlvbj0qJnJlc3BvbnNlLWNvbnRlbnQtdHlwZT0qIn1dfQ__SignaturerdjIofPONLg5roVfXki%7EycLUzpfL0VQWI4RaT6-1mTyh5gJoXxHZna1Z97YbSuU9t-KFi2Q5iGguyjKsHhZzSlRv-zVd0lPhPGmuMGcE7npRrZnPNFB2V5XFYtho8fWzKVwreGujYmaFE7Ge-zkTivnO7YkqmNs%7EeH69FRmbp5xAqW0LbsEspr0aOkaxqCCVoFXgPyJBra4dVaITeremwJSD0OcW2F2zNCdbo0nGcCDzrYsRijiUIAXweXpESL10uLt776efgzGhrDtuKKdVD68Q-3voq5PdTWTQaVP2VCDjkb66QECoAoUDtRjGJtK41zNfBs-1N6RGMDUWMNRM4A__Key-Pair-IdKVTP0A1DKRTAX然后放在/root/.cache/huggingface/hub/models--timm--tf_efficientnetv2_s.in21k下。
遇到问题4
Couldnt apply path mapping to the remote file.解决方案 我碰到是因为远程没同步等待一会儿就好了。
其中还有网络不佳的问题 我手动下载强行改了路径先跑着 遇到问题5
AttributeError: module torch has no attribute frombuffer解决方案 发现readme中写pytorch版本要1.8.2而根据requiremets.txt装的是1.8.1的坑得很但是没找到但发现没装cudatoolkit现在进行安装。未果。还是先装1.8.2看看参考 https://pytorch.org/get-started/previous-versions/
pip3 install torch1.8.2 torchvision0.9.2 torchaudio0.8.2 --extra-index-url https://download.pytorch.org/whl/lts/1.8/cu102没用要装torch2.0
conda install pytorch2.0.1 torchvision0.15.2 torchaudio2.0.2 pytorch-cuda11.8 -c pytorch -c nvidia然后docker中的cuda也要换成11.8,cudnn要换成相应的conda中的cudatoolkit也要换成11.8
然后还要修改
vim /path/to/deep-object-reid-multilabel/torchreid/models/gcn.py vim /root/anaconda3/envs/torchreid/lib/python3.8/site-packages/randaugment/randaugment.py
遇到问题6
pth要下载网络不佳报错了。需要科学上网
解决方案
手动下载https://drive.google.com/uc?exportdownloadid1N0t0eShJS3L1cDiY8HTweKfPKJ55h141 然后放入并改名/root/.cache/torch/checkpoints/uc?exportdownloadid1N0t0eShJS3L1cDiY8HTweKfPKJ55h141.pth
遇到问题7
RuntimeError: DataLoader worker (pid 3273) is killed by signal: Bus error. It is possible that dataloaders workers are out of shared memory. Please try to raise your shared memory limit.解决方案
shared memory不够把环境打包重启docker设置shm2G
遇到问题6
然后报cuda内存不够
解决方案
调整batch size:
vim configs/EfficientNetV2_small_gcn.yml终于跑起来了脑壳嗡嗡的
注意
我这边跑的是COCO数据集要把coco2014的训练和验证数据集里的图片放到项目目录/datasets/COCO/images中
然后还有配置文件configs/EfficientNetV2_small_gcn.yml的voc都改成coco注意对应的大小写。
