search

ChrisWu1997 / 2D-Motion-Retargeting

PyTorch implementation for our paper Learning Character-Agnostic Motion for Motion Retargeting in 2D, SIGGRAPH 2019
https://motionretargeting2d.github.io
MIT License
440 stars 86 forks source link

Extract latent representation information? #20

Closed chuanenlin closed 4 years ago

chuanenlin commented 4 years ago

From Fig. 9, 10, 11 of the paper, I saw that clustering visualizes the latent spaces distributions (view, skeleton, motion). I am wondering if it is possible to output the latent space data (e.g., camera-view angle in degrees of the subject for a specific frame)?

ChrisWu1997 commented 4 years ago

Hi, there are functions to visualize latent codes in cluster.py. You can look into that code for your own need.

chuanenlin commented 4 years ago

Thanks for the reply.

I tried running python cluster.py -n view to test and received the following error: line 126, in get_cluster_data all_data = torch.stack(all_data, dim=0) RuntimeError: expected a non-empty list of Tensors

Does this mean I first need to download the Mixamo dataset as described in mixamo_download_script.java, and if so, may I ask where should the .fbx files be saved and are there any specific formatting (e.g., of directory names) required?

ChrisWu1997 commented 4 years ago

Yes, you miss the data. But there is no need to download the original fbx, we provide processed data for direct usage.