skipping keys containing lists, bc the robot gives nothing but tensors

lerobot/common/robot_devices/control_utils.py

@@ -109,6 +109,9 @@ def predict_action(observation, policy, device, use_amp):
     ):
         # Convert to pytorch format: channel first and float32 in [0,1] with batch dimension
         for name in observation:
+            if isinstance(observation[name], list):
+                continue  # VLA-like and multirobot policies include textual inputs in the observation
+
             if "image" in name:
                 observation[name] = observation[name].type(torch.float32) / 255
                 observation[name] = observation[name].permute(2, 0, 1).contiguous()