act_pusht_baseline_100k

Stock LeRobot ACT config on PushT: ~0% success with default inference settings; the starting point of the tuning journey.

Part of a 6-week robot-learning training block (https://github.com/JaHeRoth/robot-learning). Evaluated on gym-pusht at n_action_steps=16; avg_sum_imputed_reward imputes 0.95/step to horizon 300 after success.