GPU/python環境配置與驗證。
(1)GPU加速型實例安裝NVIDIA GPU驅動及CUDA工具包:https://support.huaweicloud.com/usermanual-ecs/zh-cn_topic_0149470468.html#ZH-CN_TOPIC_0149470468__section1034245773916
(2)華爲雲linux服務器部署TensorFlow-gpu全攻略:https://www.cnblogs.com/zxyza/p/10535939.html
(3) Ubuntu安裝Anaconda3: https://www.jianshu.com/p/d9fb4e65483c
(4)添加環境變量: vim ~/.bashrc
export PATH="/root/anaconda3/bin:$PATH"
export PATH=/usr/local/cuda/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
export CUDA_HOME=/usr/local/cuda
(5)source ~/.bashrc
(6)創建虛擬環境:
conda create -n py37 python=3.7
進入環境
source activate py37
conda activate py37
退出環境
source deactivate
conda deactivate
(7)source activate py37
(8)安裝tensorflow-gpu:pip install tensorflow-gpu==1.13.1 -i https://pypi.tuna.tsinghua.edu.cn/simple
(9)測試:
import tensorflow as tf
import os
# os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
print('GPU>>>>>>', tf.test.is_gpu_available())
a = tf.constant(2.0)
b = tf.constant(4.0)
print(a + b)
(10) 結果:
GPU>>>>>> True
Tensor("add:0", shape=(), dtype=float32)
(11) 不同版本torch安裝:
conda install pytorch==1.2.0 torchvision==0.4.0 cudatoolkit=10.0
conda install pytorch==1.4.0 torchvision==0.5.0 cudatoolkit=10.0
conda install pytorch==1.5.0 torchvision==0.6.0 cudatoolkit=10.1
上述命令直接安裝太太太慢了,可以通過更換conda源來加速下載。
# 修改conda配置
vim .condarc
# 在配置鍾添加清華源
channels:
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/menpo/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/bioconda/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/msys2/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge
- default
show_channel_urls: true
# 安裝pytorch和對應版本的cudatoolkit
conda install pytorch=1.4.0 torchvision cudatoolkit=10.1