Succes Rate always 0？
#1
by
LGH10086
- opened
Could anyone tell me why the success rate is always 0.Even the sum of reward is 0.
what's the meaning of "action = policy.select_aciton(batch)"?
the code in LeRobot is "action = policy.select_action(observation)"?
Same here. I got 0% SR after training π0 with 100k samples using lerobot/aloha_sim_insertion_human. When looking through the eval episodes, both arms got stuck at grasping the cubes. Any reasons behind this?