-
Notifications
You must be signed in to change notification settings - Fork 208
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
added nvidia jetson hw support #59
base: main
Are you sure you want to change the base?
Conversation
Hey Zuowei, you will need to compile triton separately on your AGX Orin.
That is not covered in this PR but you will find relevant work on the
triton GitHub issues.
…On Fri, Aug 9, 2024, 04:17 Zuowei ***@***.***> wrote:
thank you for your post, i just install your modified causal-conv1d
successfully, but when i install your mamba-1p1p1 it shows INFO: pip is
looking at multiple versions of mamba-ssm to determine which version is
compatible with other requirements. This could take a while.
ERROR: Could not find a version that satisfies the requirement triton
(from mamba-ssm) (from versions: none)
ERROR: No matching distribution found for triton
i am using Jetson AGX Orin, and i can not just pip or conda install
triton... thank you
—
Reply to this email directly, view it on GitHub
<#59 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AH23AWX37N6EAQ53TMMZZQLZQR3LBAVCNFSM6AAAAABFWEX3ZSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDENZXGQYTGMRTGU>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
thank you so much for your reply! I just did a lot search on compiling triton on jetson, but i only found the tutorial of using triton server or installing 'tritonclient' package from whl, i didn't find how to compile 'triton' to be a package for my conda environment, maybe could you send me a link or sth, which can tell how to install it? i'm a rookie, that would help me a lot thank you sooo much!!! |
well, i just found your post under triton of openai triton-lang/triton#3505 i will try it first thank you @!!! |
Hello @Unrealluver
Thank you for releasing the code for Vim. This is a PR to support nvidia jetson GPUs. With Vim's efficiency, it would be really interesting to see their performance compared to convnets and transformers on this platform.
The
setup.py
has been modified for bothmamba_ssm
andcausal-conv1d
as seen in the below PRsstate-spaces/mamba#262
Dao-AILab/causal-conv1d#20
I have added Jetson support with CUDA compatibility > 70 as
triton
has a minimum CUDA compatibility of 70I have also modified
README.md
to fix a small typoLet me know if further changes are required