DeepMimic: example-guided deep reinforcement learning of physics-based character skills




  • Paper topic: Animation and Simulation
  • Software type: Code
  • Able to run a replicability test: True
  • Replicability score: 5
  • Paper listed in the Graphics Replicability Stamp Initiative
  • Software language: C/C++, Python
  • License: MIT
  • Build mechanism: Makefile, Not applicable (python, Matlab..)
  • Dependencies: libbullet/eigen3/opengl/freeglut/glew/swig/openmpi/tensorflow
  • Documentation score {0,1,2}: 1
  • Reviewer: Julie Digne <>
  • Time spent for the test (build->first run, timeout at 100min): 30min

Source code information


I was able to run and visualize pretrained policies + learned a policy training "humanoid3d_spinkick_args.txt

If you want to contribute with another review, please follow these instructions.

Please consider to cut/paste/edit the raw JSON data attached to this paper.