search

brownvc / deep-synth

Public code release for our SIGGRAPH 2018 paper "Deep Convolutional Priors for Indoor Scene Synthesis"
Other
136 stars 33 forks source link

Can't load pre-trained models #3

Closed zyf12389 closed 5 years ago

zyf12389 commented 5 years ago

The pre-trained model that used in second stage(location) could't work well. The basemodel is Res101 in your code. Error is as follows.
RuntimeError: Error(s) in loading state_dict for ResNet: Unexpected key(s) in state_dict: "bn1.num_batches_tracked","layer1.0.bn1.num_batches_tracked", "layer1.0.bn2.num_batches_tracked", "layer1.0.bn3.num_batches_tracked", "layer1.0.downsample.1.num_batches_tracked", "layer1.1.bn1.num_batches_tracked", "layer1.1.bn2.num_batches_tracked"......

Looking forward to your reply. Thanks

zyf12389 commented 5 years ago

sorry. I found my torch's version is 0.4.0, maybe it's the version.