Skip to content
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

Linux support #12

Open
ZodiacFRA opened this issue Dec 11, 2024 · 7 comments
Open

Linux support #12

ZodiacFRA opened this issue Dec 11, 2024 · 7 comments

Comments

@ZodiacFRA
Copy link

Hi!

The readme states:

It extends the performance and feature set of WhisperKit from Apple platforms to Android and (soon) Linux.

I was wondering if you had a date for the Linux release? :)
Thanks for you work, have a great day

@bpkeene
Copy link
Contributor

bpkeene commented Dec 24, 2024

Hi ZodiacFRA,

We are adding initial (CPU) Linux support imminently, and appreciate your interest in the project :)

Please keep a look out in the coming days!

@ZodiacFRA
Copy link
Author

Thank you for your work :)
Can you share if you plan to use tflite models as well for the Linux version?

Merry Christmas

@keith4ever
Copy link
Collaborator

Hi @ZodiacFRA ,

For Linux version, we use the exact same model binary (.tflite) file, therefore you don't need to download it again.

@ZodiacFRA
Copy link
Author

Nice, do you happen to know if Whisper's Tflite models can be run on Qualcomm's DSP || HTP || GPU with Linux?

I tried to do it last year and it wasn't possible at the time because of the dynamic tensor shapes limitations, is that not longer the case?

@keith4ever
Copy link
Collaborator

As far as we know, it's either HTP (QNN) or just falling back to GPU which is accelerated by TFLite framework.
It may use DSP core but there is no way we can tell, as it is done in QualCom's library.

I do not know about dynamic tensor shape.. we're only using static shape with Whisper models, so please check it out.

@Alex2025Job
Copy link

Hi, any updates for linux release? is there a repo for it?

@keith4ever
Copy link
Collaborator

Hi @SongLi89 ,

The native Linux (x86) is supported. Please git pull to ToT main and try these:

make clean all
make env
make download-models
make build linux
make test linux

please note that ‘make build linux’ and ‘make test linux’ has to be done in Docker env.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants