字體：小 中 大 |
|
|
|
| 2019/03/04 10:42:26瀏覽197｜回應0｜推薦0 | |
安裝ubuntu 16
ubuntu 16.04.4 LTS,不要auot login
安裝cuda
第一次開機後,連網, login進desktop,
Terminal / sudo apt-get update
Additional Drivers/Additional Drivers/Using NVIDIA binary driver,安裝後重新開機.
連網,login進desktop,
...
Additional Drivers/Additional Drivers/Using X.Org Xserver,安裝後重新開機.
...
在login畫面,按Alt+Ctrl+F1,進入指令模式.
sudo service lightdm stop ### stop Xserver
sudo bash cuda_10.0.130_410.48_linux.run
accept
Install NVIDIA Driver = y
Do you want to run nvidia-xconfig = n #裝了就完蛋了
...
安裝cuDNN
login進desktop
sudo dpkg -i libcudnn7_7.5.0.56-1+cuda10.0_amd64.deb
sudo dpkg -i libcudnn7-dev_7.5.0.56-1+cuda10.0_amd64.deb
安裝conda 37
login進desktop
sudo ./Anaconda3-2018.12-Linux-x86_64.sh
創建py36 env
new terminal
conda create -n py36 python=3.6
test:
conda env list
# conda environments:
#
py36 /home/aigcat/.conda/envs/py36
base * /home/aigcat/anaconda3
安裝tensorflow
source activate py36
(py36)
conda install tensorflow-gpu
test:
python
import tensorflow
安裝Keras 224
(py36)
pip install keras==2.2.4
安裝jupyter notebook
conda install ipykernel
conda install spyder notebook conda install numpy scipy pyyaml hdf5 h5py conda install matplotlib conda install mkl-service pip install pandas
#play DeepLearningFlappyBird
pip install pygame pip install opencv-python |
|
| ( 心情隨筆｜心情日記 ) |
