[NeurIPS 2020] Official PyTorch Implementation of "Residual Force Control for Agile Human Behavior Imitation and Extended Motion Synthesis". NeurIPS 2020.
Many thanks for sharing code of this very nice work!
I am insterested in the experiment II Extended Motion Synthesis on H3.6M dataset.
Will you release the code related to this experiment?
Btw, in the readme, the first RFC GIF result, the humanoid model changed direction in Yaw a lot. Is this due to the reward function, which does not foucs on global orientation (qr)?
Hi,
Thanks for sharing this excellent work.
I met some problems when runing the training code about multiprocessing. The training code "python motion_imitation/motion_im.py --cfg 0506 --num_threads <max_num_CPU_threads_you_have>" works well when I only use one cpu core. But it stuck when I set the num_threads over 1(such as 2,4,12), the problem seems to be happened in "agent.py": "pod, worker_memory, worker_logger = queue.get()" when the ppo algorithm starts sampling.
The multiprocessing process seems work okay in the first sampling iteration, but the queue start to be empty in the second iteration and then the whole training process stuck.
I'm not familiar with gym and multiprocessing, do you know how to fix it?
Many thanks for sharing code of the very interesting work!
I want to know if your work can imitate the various human 3D poses from a live video. i.e., let the robot to learn the actions of people in the video
I'm not very familiar with the filed of motion imitation. I thought that once a model is trained, it could mimic all the motion clips. But from your code, it seems like that a new model has to be trained for each motion clip.