[object Object]

← back to Exo

make readme clearer with linux nvidia

2c5d05538cdaa6a2ef416aa2434ba3f17e1d80a9 · 2024-11-24 00:18:26 +0400 · Alex Cheema

Files touched

Diff

commit 2c5d05538cdaa6a2ef416aa2434ba3f17e1d80a9
Author: Alex Cheema <alexcheema123@gmail.com>
Date:   Sun Nov 24 00:18:26 2024 +0400

    make readme clearer with linux nvidia
---
 README.md | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/README.md b/README.md
index 550b81a6..c7e0eec8 100644
--- a/README.md
+++ b/README.md
@@ -67,10 +67,10 @@ The current recommended way to install exo is from source.
 ### Prerequisites
 
 - Python>=3.12.0 is required because of [issues with asyncio](https://github.com/exo-explore/exo/issues/5) in previous versions.
-- Linux (with NVIDIA card):
-  - NVIDIA driver (test with `nvidia-smi`)
-  - CUDA (https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#cuda-cross-platform-installation) (test with `nvcc --version`)
-  - cuDNN (https://developer.nvidia.com/cudnn-downloads) (test with [link](https://docs.nvidia.com/deeplearning/cudnn/latest/installation/linux.html#verifying-the-install-on-linux:~:text=at%20a%20time.-,Verifying%20the%20Install%20on%20Linux,Test%20passed!,-Upgrading%20From%20Older))
+- For Linux with NVIDIA GPU support (Linux-only, skip if not using Linux or NVIDIA):
+  - NVIDIA driver - verify with `nvidia-smi`
+  - CUDA toolkit - install from [NVIDIA CUDA guide](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#cuda-cross-platform-installation), verify with `nvcc --version`
+  - cuDNN library - download from [NVIDIA cuDNN page](https://developer.nvidia.com/cudnn-downloads), verify installation by following [these steps](https://docs.nvidia.com/deeplearning/cudnn/latest/installation/linux.html#verifying-the-install-on-linux:~:text=at%20a%20time.-,Verifying%20the%20Install%20on%20Linux,Test%20passed!,-Upgrading%20From%20Older)
 
 ### Hardware Requirements
 

← f5fdacd2 16.0.0 for all circleci jobs  ·  back to Exo  ·  backend endpoint now uses SSE to send each model as its load e16170cf →