How to use OpenVINO backend for Cellpose cyto2

MouseLand / cellpose

a generalist algorithm for cellular segmentation with human-in-the-loop capabilities

https://www.cellpose.org/

BSD 3-Clause "New" or "Revised" License

1.24k stars 359 forks source link

How to use OpenVINO backend for Cellpose cyto2 #953

Closed pvardanis closed 3 weeks ago

pvardanis commented 4 weeks ago

Hi,

I'm having a cyto2 model and would like to run an optimized inference for Intel with OpenVINO. In the docs, I see:

from cellpose.contrib import openvino_utils

model = models.CellposeModel(...)

model = openvino_utils.to_openvino(model)

However this is only available in versions 2.*. Even with such versions, it isn't quite clear how to use or I'm missing something in the docs.

carsen-stringer commented 3 weeks ago

since openvino does not support training, only inference, we decided the small speed-up was not worth maintaining the code. I would recommend using google colab for free GPU access as a faster option. but this should be the correct usage I believe, and then you use the model for inference (with the model.eval method)