When I’m trying to install flash-attn inside a virtual environment, the build process, starts eating up all the memory and eventually crashes the whole system.
what is the correct way to install flash-attn for the jetson orin boards?
Hello,
Thanks for visiting the NVIDIA Developer forums! Your topic will be best served in the Jetson category.
I will move this post over for visibility.
Cheers,
Tom
Hi,
For the JetPack 6.1 environment, you can find the prebuilt package in the below link directly:
https://pypi.jetson-ai-lab.dev/jp6/cu126
Thanks.
thank you for your response.
now I can install it but as soon as I want to use it it gives me error about not being able to find a name in libcuda.
Hi,
Could you share what kind of command that triggers the issue?
We can import the package on JetPack 6.1 without issues.
$ python3
Python 3.10.12 (main, Sep 11 2024, 15:47:36) [GCC 11.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import flash_attn
>>> flash_attn.__version__
'2.6.3'
Thanks.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.