dpeerlab/spectra

GPU tutorial?

Opened this issue · 3 comments

caitsimop commented

I'm looking to use the GPU implementation of Spectra as mentioned in issue #21.

When you load from Spectra import Spectra_gpu as spc_gpu you get this info:

Spectra GPU support is still under development. Raise any issues on github 
 
 Changes from v1: 
 (1) GPU support [see tutorial] 
 (2) minibatching for local parameters and data 
 Note that minibatching may affect optimization results 
 Code will eventually be merged into spectra.py

Is there a tutorial hiding somewhere? I'm especially interested in the minibatching.

Thanks!

russellkune commented

Hi Cat, I'm working this. keeping it open

  • ❤️1
kousaa commented

Hello,

thanks for developing this great tool! Just adding to Cat's comment above, it would be great if in your upcoming tutorial there is a section on how to utilize the GPU on the new Apple silicon OS.

Thank you!
Anastasia

andrewjkwok commented

Hello, wanted to check if there were any updates on GPU use?

One thing I noted was that various arguments for the CPU est_spectra function don't seem to be needed / used for the gpu version i.e.
label_factors, overlap_threshold, clean_gs and min_gs_num.

Is there any specific reason for this?