Documentation Index
Fetch the complete documentation index at: https://docs.gpuhub.com/llms.txt
Use this file to discover all available pages before exploring further.
If you wish to reuse the dependency environment when renting other hosts, you can try to save all data on the system disk as an image. Please refer to the documentation on saving and loading Images.
Install Other Python Versions
All images come with Miniconda pre-installed.
# Create a virtual environment named my-env with Python version 3.7:
conda create -n my-env python=3.7
# Update the environment variables in .bashrc:
conda init bash && source /root/.bashrc
# Activate the created virtual environment my-env:
conda activate my-env
Install Other Frameworks
# PyTorch: Find the appropriate version from Previous PyTorch Versions. For example:
conda install pytorch==1.9.0 torchvision==0.10.0 torchaudio==0.9.0 cudatoolkit=11.3 -c pytorch -c conda-forge
# TensorFlow: Find the corresponding version link from Install TensorFlow with pip. For example:
pip install https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow_gpu-2.6.0-cp38-cp38-manylinux2010_x86_64.whl
Install Python Dependencies
# Using pip: For example,
pip install opencv-python scipy numpy Pillow
# Using conda: For example,
conda install numpy
# Searching for package names:
pip search xxxx
conda search xxxx
# Trick to view available versions if unsure: Write a random version number (e.g., 9.9),
# and pip will list all available versions in the error message:
pip install xxx=9.9
Install System Dependencies
# Example for installing zip:
apt-get update # Only needs to be run once, not every time
apt-get install -y zip
# Searching for package names:
apt-get update # Only needs to be run once, not every time
apt-cache search xxxxx
Install CUDA and Other Dependencies
For the installation of CUDA and other software, please refer to the documentation.
