It's not an easy task to get your ML/DL workstation environment up and running.If you are a windows user or only tried cloud platforms like google colab then this guide is certainly for you !
device: Lenovo Legion Y530-15ICH
GPU: GeForce GTX 1060 (6GB)
HDD type : GPT
RAM: 16GB
Any linux distro that is deep learning compatible can be used. you can also use windows but the installation is a little bit trickier. This guide contains steps to setup Ubuntu 20.04 for deep learning.
Installing Ubuntu 20.04:
I tried a dual boot for Ubuntu alongside windows and it worked.
steps:
1- Download ubuntu20.04 LTS from here link
2- Download rufus which is a usb bootloader from here link
3- Get an empty flash USB drive (4GB is fine) but (8-16 GB recommended).
3- Free some space for your new OS from the disk management in windows by shrinking the volume of one of your disks (make sure the new unallocated space suits your needs) this link may help link
4- Using rufus boot ubuntu on the flash drive (make sure you don't have any important files on it as it will be formatted. ) link
5- This step is tricky as it differs from device to another but the main goal is the same. disable secure booting and enable UEFI/ legacy booting depending on your device .follow this link here to check your deviice booting link
6- Go to the boot menu in your device and choose the usb drive .
7- follow the install instructions by the prompt windows. this may help link
steps:
1- From your applications menu search for software and updates
2- select the additional drivers tab
3- I chose nvidia-460 driver and the rest of this guide is based on it (you can choose other drivers but you have to choose other compatible CUDA toolkit and cuDNN ). click apply changes .
4- click ctrl+alt+T to open terminal and type reboot to restart your system
5- click ctrl+alt+T to open terminal and type nvidia-smi to check driver install you should expect something like this
warning never try sudo apt install cuda-toolkit as it's locked to cuda 10.1 which is not compatible with new drivers.
steps:
1- Go to link open terminal and use the best installation commands provided.
2- After finishing the installation it's important to add CUDA to your path. open terminal and type this command to open the bashrc file
sudo nano ~/.bashrccopy and add these lines to the end of the bashrc file
export PATH=/usr/local/cuda-11.2/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-11.2/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
export CUDA_HOME=/usr/local/cuda3- perform a system reboot
reboot4- verify cuda toolkit installation using this command
nvcc -Vyou should expect something like this
steps :
1-To download cuDNN you need to have a nvidia developer account which is easy to make. Then fill their 2-min survey and agree to the terms and conditions and choose archived versions
2- I tried the debian installation and it worked (you can try other installations too) for debian installations download runtime,developer and samples library as shown here
3- open terminal and type the following commands :
sudo apt-get install libfreeimage3 libfreeimage-dev
then
cd Downloads/
To install runtime library:
sudo dpkg -i libcudnn8_x.x.x-1+cudax.x_amd64.deb
(use tab to fill the x's with the version you downloaded)
To install the developer library, for example:
sudo dpkg -i libcudnn8-dev_8.x.x.x-1+cudax.x_amd64.deb
To install the code samples and the cuDNN library documentation, for example:
sudo dpkg -i libcudnn8-samples_8.x.x.x-1+cudax.x_amd64.deb
Procedure
Copy the cuDNN samples to a writable path.
$cp -r /usr/src/cudnn_samples_v8/ $HOME
Go to the writable path.
$ cd $HOME/cudnn_samples_v8/mnistCUDNN
Compile the mnistCUDNN sample.
$make clean && make
Run the mnistCUDNN sample.
$ ./mnistCUDNN
If cuDNN is properly installed and running on your Linux system, you will see a message similar to the following:
Test passed!
By far the hardest part should have passed the rest is the easy stuff.
Install Chrome
link
sudo dpkg -i google-chrome-stable_current_amd64.deb
Install Development Tools:
sudo apt install build-essential pkg-config cmake cmake-qt-gui ninja-build valgrind
Install Git:
sudo apt install git
git config --global user.name "Name"
git config --global user.email "name@domain.com"
Install Python 3:
sudo apt install python3 python3-wheel python3-pip python3-venv python3-dev python3-setuptools