search

jakowenko / double-take

Unified UI and API for processing and training images for facial recognition.
https://hub.docker.com/r/jakowenko/double-take
MIT License
1.19k stars 93 forks source link

[FEAT] Support for Windows GPU version of Deepstack #224

Open guarddog13 opened 2 years ago

guarddog13 commented 2 years ago

Is your feature request related to a problem? Please describe. No

Describe the solution you'd like As the title says support for GPU of Deepstack on Windows.

jakowenko commented 2 years ago

Is this related to the DeepStack addon? I can look into publishing the GPU version as well.

guarddog13 commented 2 years ago

Yes sorry that's what I meant. I can get the docker GPU version running and double take sees that it is attached but it returns nothing when I try to click on matches in Double Take. Thank you so much!!

guarddog13 commented 2 years ago

@jakowenko I now use deepstack on a separate Ubuntu machine that has an Nvidia GPU. I set up deepstack gpu up on this machine with docker. Does doubletake not support the GPU version of deepstack? When I attempt to train with double take i get a 500 error...training with the same picture using deepstack non-gpu gives a positive result.

FYI Double take and Frigate GPU works fine.

jakowenko commented 2 years ago

@jakowenko I now use deepstack on a separate Ubuntu machine that has an Nvidia GPU. I set up deepstack gpu up on this machine with docker. Does doubletake not support the GPU version of deepstack? When I attempt to train with double take i get a 500 error...training with the same picture using deepstack non-gpu gives a positive result.

FYI Double take and Frigate GPU works fine.

Hm...do you get back anything else besides the 500? It's most likely something with DeepStack. The API should work the same across their different versions. Can you replicate the 500 with their examples on their docs? May need to open an issue with them and see if they know what's going on.