search

PABannier / bark.cpp

Suno AI's Bark model in C/C++ for fast text-to-speech
MIT License
635 stars 49 forks source link

support GPU or not ? #73

Closed EricKong1985 closed 3 months ago

EricKong1985 commented 11 months ago

I have checked the project description that : The main goal of bark.cpp is to synthesize audio from a textual input with the Bark model in efficiently using only CPU.

Could I know the it support GPU or not ? I suppose that using GPU should be more faster that using CPU

PABannier commented 11 months ago

Hi @EricKong1985 ! We could support GPU with OpenCL. Currently, we are focusing on removing any bugs from the library and making it fast for CPU. If you're willing to add support for GPUs, I'm happy to guide you in the process.

PABannier commented 10 months ago

TODO: Support ggml-cuda for GPU acceleration

crackedpotato007 commented 7 months ago

Any updates on this?

PABannier commented 7 months ago

@arnav7633 I have this ongoing PR #124 to integrate the latest changes from encodec.cpp and ggml (which includes ggml-cuda). Currently, I do not have time to merge it, but I plan to do it by the end of the year.

PABannier commented 3 months ago

@crackedpotato007 @EricKong1985 The support for GPU has landed with CUDA and #139 . If you want to give it a try, i'd be happy to have your feedbacks

crackedpotato007 commented 1 month ago

Extremely sorry, for being... a few months late.

Had my uni exams so i was rarely checking my notifs or bothering to reply to them. How do i go about using the GPU? Running the interference in the examples section doesn't use my GPU at all