Hello, firstly really appreciate the continuous development and maintaining of the repository. We are currently trying to run Real Robot Locomotion all on the perception board of the Booster T1. However, we assume that this is not explicitly supported, as the documentation on running onboard mostly refers to the Unitree G1. So, back in January, we ran into issues: namely that, because hsinference creates a python environment, made it so that the base booster_sdk on the perception board was not recognized inside of it. At the time, we circumvented the issue by bringing the booster_robotics_sdk_python.pyi and booster_robotics_sdk_python.cpython-310-x86_64-linux-gnu.so of the base instalation of the sdk on the robot to the python environment, and we were able to successfully deploy onboard. However, after we have made several of the official firmware updates on the robot and routinely upgraded the booster_sdk to their most recent versions, deployment now just never leaves PREP mode, whilst not showing any errors in the terminal. It still recognizes the booster_sdk package and its imports inside the hsinference environment, however we assume that there must exist some deprecated instance of some of the functions from the sdk that holosoma uses that was changed in the recent firmware or sdk updates. We just wanted to know if you were aware of this possible issue? Or if perhaps on your side everything is running smoothly.
Hello, firstly really appreciate the continuous development and maintaining of the repository. We are currently trying to run Real Robot Locomotion all on the perception board of the Booster T1. However, we assume that this is not explicitly supported, as the documentation on running onboard mostly refers to the Unitree G1. So, back in January, we ran into issues: namely that, because hsinference creates a python environment, made it so that the base booster_sdk on the perception board was not recognized inside of it. At the time, we circumvented the issue by bringing the
booster_robotics_sdk_python.pyiandbooster_robotics_sdk_python.cpython-310-x86_64-linux-gnu.soof the base instalation of the sdk on the robot to the python environment, and we were able to successfully deploy onboard. However, after we have made several of the official firmware updates on the robot and routinely upgraded the booster_sdk to their most recent versions, deployment now just never leaves PREP mode, whilst not showing any errors in the terminal. It still recognizes the booster_sdk package and its imports inside the hsinference environment, however we assume that there must exist some deprecated instance of some of the functions from the sdk that holosoma uses that was changed in the recent firmware or sdk updates. We just wanted to know if you were aware of this possible issue? Or if perhaps on your side everything is running smoothly.