I am wondering if anyone has benchmarked model performance on intel hardware as a backend using TVM and comparing it with openVINO. I am interested in looking at how each performs for couple of models like resnet and llama2 for example?
I am wondering if anyone has benchmarked model performance on intel hardware as a backend using TVM and comparing it with openVINO. I am interested in looking at how each performs for couple of models like resnet and llama2 for example?