DIS-A100-4090

Updated the environment from CUDA 10.2 to CUDA 11.8 ! Adapt to 4090 and A100 ！

Using this environment, ISNet can run on GPUs with Ampere architecture and earlier, such as the 30 series cards, 40 series cards, A100, A10, etc. Except for the H100, which requires a CUDA 12+ environment, CUDA 11.8 currently supports the vast majority of Nvidia graphics cards.

CUDA 11.8 Environment Configuration

(1) Clone this repo

git clone https://github.com/HUANGYming/DIS-A100-4090.git

Go to the DIS/ISNet folder

1. Installing a Conda Environment Using a YAML File.

conda env create -f environment_cu118.yaml

2. Installing a Conda Environment Using a TXT File.

pip install requirements_cu118.txt

(2) Only Download Configuration files

Go to the DIS/ISNet folder

Download environment_cu118.yaml or requirements_cu118.txt, then install by Conda or pip.

(3) Creating a Conda Environment from a Compressed Package.

In addition to installation from pip and conda sources, I have provided a conda environment compressed package. You can directly unzip it in the conda environment for use.

1. Download compressed package from onedrive

https://connecthkuhk-my.sharepoint.com/:u:/g/personal/huangym2_connect_hku_hk/EQSnmFcSxp9Iq5yOTs5z53UBVULNvaFH_N3uELSaZloObA?e=vxEhCA

2. Find the conda environment directory

conda info --envs

3. Enter into the conda environment directory

cd /path/to/file/envs/

3. Unzip the environment compressed package

unzip ISNet.tar.gz

Contact

If you have any questions or suggestions, please contact me at the following email: huangym2@connect.hku.hk

Acknowledgements

Special thanks to the members of the ISNet project team, whose hard work and innovative thinking laid the groundwork for the success of this project.