Skip to content

Latest commit

 

History

History
 
 

shapestacks_renderer

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
README.md
README.md
 
 
record_scenario.py
record_scenario.py
 
 

README.md

Recording ShapeStacks Scenarios

After activating the ShapeStacks virtual environment, run the recording script as follows:

(venv) $ LD_PRELOAD=/usr/lib/nvidia-384/libOpenGL.so python3 record_scenario.py \
  --mjmodel_path ${SHAPESTACKS_CODE_HOME}/data/shapestacks_example/mjcf/env_ccs-easy-h=5-vcom=2-vpsf=0-v=1.xml \
  --record_path ${SHAPESTACKS_CODE_HOME}/data/shapestacks_example/recordings/env_ccs-easy-h=5-vcom=2-vpsf=0-v=1 \
  --mjsim_time 4 \
  --fps 8 \
  --max_frames 32 \
  --cameras cam_1 cam_2 \
  --res 224 224 \
  --formats rgb

This will run the simulation forward from the beginning for 4 seconds and record RGB renderings with 8 FPS (up to 32 frames) at a resolution of 224 x 224 pixels.

After the maximum simulation time, the recording script will also report if the stack has collapsed (topmost shape has moved) or not.