what is the correct CLI for inferencing my own policy on real robot?
Hello,
I registered my ur cobot inside robots and then did a successful data collection as well. I did some minor changes to lerobot_record.py to match with how my cobot sends its pose and gripper data. I trained my data using the lerbot/smolVLA base checkpoint. But, now the problem is real-time inference on real robot. There is no specific script to do inference. The documentation says to use lerobot_record.py itself. Then the problem is should I use my changed lerobot_record.py, or is there anything else that I need to change there? Further, dataset_repoID arguement at the time of inference must direct to the collected data, or a new folder to log the evaluation results? Appreciate if someone can assist me in here, as the documentation does not give clear information on this.