1.軟件準(zhǔn)備
- anaconda3 https://www.anaconda.com/products/individual
- visual studio2017社區(qū)版 https://visualstudio.microsoft.com/zh-hans/downloads/
- cuda11.1 https://developer.nvidia.com/cuda-10.0-download-archive
- cudnn ## 目前沒(méi)有x64 牍汹，需要登陸才能下載https://developer.nvidia.com/rdp/cudnn-archive
- 迅雷下載
血淚史度苔，直接官網(wǎng)下載cuda和cudnn然低，速度慢的想死都哭，等待了一天都沒(méi)有下載下來(lái)局荚，把下載鏈接復(fù)制到迅雷谆奥，不用會(huì)員半個(gè)小時(shí)下完
2.anaconda安裝及cuda，cudnn安裝請(qǐng)參考http://www.reibang.com/p/915e6b3cdd29
3. 創(chuàng)建tensorflow-gpu環(huán)境
>>> conda info --env #查看已經(jīng)安裝的環(huán)境
>>> conda remove --name tsgpu --all #刪除上一個(gè)安裝錯(cuò)誤的環(huán)境沦童，或者不想要的環(huán)境
image.png
>>> conda create --name tensorgpu python=3.6
image.png
image.png
>>> conda activate tensorgpu
>>> conda install tensorflow-gpu==2.1.0
image.png
>>>python -m pip install --upgrade pip #更新pip撑碴，不然后續(xù)會(huì)報(bào)錯(cuò)
查看安裝版本及路徑
>>> tf.__version__
>>>tf.__path__
image.png
查看可用gpu個(gè)數(shù)
>>> import tensorflow as tf
>>> print("Num GPUs Available: ", len(tf.config.experimental.list_physical_devices('GPU')))
image.png
>>> import tensorflow as tf
>>> tf.compat.v1.disable_eager_execution()
>>> sess=tf.compat.v1.Session()
>>> hello=tf.constant("hello")
>>> print(sess.run(hello))
b'hello'
image.png
參考文章如下：
anaconda下安裝tensorflow-gpu http://www.reibang.com/p/915e6b3cdd29
Anaconda安裝Tensorflow-gpu https://www.cnblogs.com/jshmztl/p/13306837.html
