Hi everybody, I was considering whether to use Ultralytics library (interfacing python scripts to c#) or Sentis (native execution) for running inference on YOLO models. What would be the better choice overall, including performance and ease of development?
#Ultralytics vs Sentis
rich venture
devout python
Sentis requires the model to be in onnx format. If you can achieve that easily, it's probably best to use it, since you get an officially supported and documented package. This might not support latest techniques though, like quantization, cpu offloading and such.