[Linux] Ubuntu20.04 version configuration pytorch environment 2023.09.05 [Tutorial]
Article directory
1. Install Anaconda virtual environment manager
First, go to the Anaconda official website to download the installation file of the Linux version Anaconda3-2023.07-2-Linux-x86_64.sh
, enter the installation file path, and run the following script to install it.
bash Anaconda3-2023.07-2-Linux-x86_64.sh
Installation requires reading the user agreement, just keep pressing enter
it. When you reach the current page, you need to output yes
for confirmation to start the installation.
Then follow the prompts to complete the installation.
2. Create a virtual environment and activate it
Anaconda's basic environment management syntax is as follows
创建虚拟环境
conda create -n <your-virtualenv-name> python=3.8
激活虚拟环境
conda activate <your-virtualenv-name>
取消激活虚拟环境
deactivate
移除虚拟环境
conda env remove <your-virtualenv-name>
查看所有虚拟环境
conda env list
After the installation is complete, you can create a virtual environment. After creating the virtual environment, we first activate the environment. For example, I created a virtual environment metaRL
and then conda activate metaRL
activated it using
3. Anaconda cancels the default activation of the virtual environment
If we don’t want to activate the anaconda virtual environment every time we start the terminal, we can cancel the default activation, open the terminal and enter the following statement
conda config --set auto_activate_base false
4. Install Pytorch
First go to the Pytorch official website , then select the model of your device and copy the corresponding installation statement to install it.
For example, the installation statement I copied here is
pip3 install torch torchvision torchaudio
Then wait for the installation to complete
after a long wait
Finally, the installation is completed, and a successful installation Suceessfully built lit
prompt appears.
4. Test pytorch
After the installation is successful, we need to test whether it can be used normally. Open the terminal and start python in the terminal.
python
Import pytorch
import pytorch
View torch version
print(torch.__version__)
Check if cuda is available
print(torch.cuda.is_available())
The following results indicate that the installation is successful.