Hi All,
I'm now confused with a problem of forwarding the model in NCNN.
The model is resnet50 trained under PyTorch1.0. I delete all the keys with "num_batches_tracked"of the state_dict created by the code:torch.load(model_path).
It succeeded in converting the model, but failed when run:
ncnn::Mat feature;
ex.extract("Addmm_1", feature);
in my C++ codes.
exis a ncnn::Extractor, which can successfully extract all the layers' outputs except the last layer:Addmm_1.
Addmm_1 is converted from:
self.fc = nn.Linear(512 * block.expansion, feature_dimension)
in my Python codes.
Anyone have ever suffered the same problem?
Hi All,
I'm now confused with a problem of forwarding the model in NCNN.
The model is resnet50 trained under PyTorch1.0. I delete all the keys with
"num_batches_tracked"of thestate_dictcreated by the code:torch.load(model_path).It succeeded in converting the model, but failed when run:
in my C++ codes.
exis ancnn::Extractor, which can successfully extract all the layers' outputs except the last layer:Addmm_1.Addmm_1is converted from:in my Python codes.
Anyone have ever suffered the same problem?