Deep sparse compile_model vs onnxruntime

Difference between deep sparse compile_model engine vs onnxruntime?

is compile_model is better than onnxruntime…if so how…is it the inference speed increases?

Hi @amir
“Better” is relative depending on a variety of different variables. It could depend on the network or operator support for example. For example, Neural Magic has several blog posts that describe situations where using deepsparse compile_model is very much advantageous.

Do you have an example use case that helps better illustrate your situation?
Thank you,
Jeannie / Neural Magic