Giters

microsoft / nn-Meter

A DNN inference latency prediction toolkit for accurately modeling and predicting the latency on diverse edge devices.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

I wonder if nn-Meter can be applied to other network structures?

nuts-bottles opened this issue · comments

commented

I modify the original neural models in the code, but it returns
"ValueError: Unsupported Model Name: modresnet50_0 in torchvision.
Supporting list: resnet18, alexnet, vgg16, squeezenet, densenet161, inception_v3, googlenet, shufflenet_v2, mobilenet_v2, Resnext50_32x4d wide_resnet50_2, mnasnet"

commented

Hi, it seems you used nn-Meter with wrong model_type. We support latency prediction for CNN structure. If your model is CNN, you can refer to this guidance to use nn-Meter. I think the python binding usage is more suitable in your case.

commented

I find that there are some requirements for the definition of the model structure in NII model.
The code works well, thank you.