Shih-En Wei

The network itself is nothing more than a CNN, so methods proposed for binarizing CNNs can be applied, for sure. I guess by performance you mean speed. Binarizing will cause...

In the updated caffe repo I can successfully compile with `CPU_ONLY := 1`. Please report again if same error still happens, thanks.

Apologies for the confusion. We've been discussing if we should keep that version, or provide one not beyond the scope of the paper. Will update soon.

The file is now back on: https://github.com/shihenw/convolutional-pose-machines-release/blob/master/testing/python/demo.ipynb

I'll add the setting I used in `genProto.py.` For the `pose_train_test.prototxt`, in fact the one used for LSP is the same as for MPI. And for FLIC, the only 2...

@shangwenxiang can you check are you using our version of caffe? When you have more than one version, make sure the environment variable $PYTHONPATH is pointing to the correct one.

There might be these difference: 1) Network itself: we just add a new model that runs faster and scores higher on MPI. Please pull repo and run `testing/get_model.sh` again. 2)...

You're right, the person detector here is not very strong. You can replace it to other person detector and use to result as the initialization of CPMs.

Can you check if your environment variables PATH and LD_LIBRARY to CUDA are correctly set?

Thanks for reporting! We'll fix this issue (and provide a python version as well).