ValueError: zero-size array to reduction operation maximum which has no identity

[beizhengren]

Thanks for the sharing.

When I run the simulation command as
sh ./script/run_pluto_planner.sh pluto_planner nuplan_mini mini_demo_scenario last.ckpt new_output
the following error occurs:
ValueError: zero-size array to reduction operation maximum which has no identity

The detail of the error:

----------- Simulation failed: with the following trace:
Sequential:   0%|                                                                                             | 0/1 [00:47<?, ?it/s]Traceback (most recent call last):
  File "/home/wyz/workspace/learning/planing/nuplan-devkit/nuplan/planning/simulation/runner/executor.py", line 27, in run_simulation
    return sim_runner.run()
  File "/home/wyz/workspace/learning/planing/nuplan-devkit/nuplan/planning/simulation/runner/simulations_runner.py", line 113, in run
    trajectory = self.planner.compute_trajectory(planner_input)
  File "/home/wyz/workspace/learning/planing/nuplan-devkit/nuplan/planning/simulation/planner/abstract_planner.py", line 108, in compute_trajectory
    raise e
  File "/home/wyz/workspace/learning/planing/nuplan-devkit/nuplan/planning/simulation/planner/abstract_planner.py", line 105, in compute_trajectory
    trajectory = self.compute_planner_trajectory(current_input)
  File "/home/wyz/workspace/learning/planing/pluto/src/planners/pluto_planner.py", line 166, in compute_planner_trajectory
    planning_trajectory = self._run_planning_once(current_input)
  File "/home/wyz/workspace/learning/planing/pluto/src/planners/pluto_planner.py", line 200, in _run_planning_once
    candidate_trajectories, learning_based_score = self._trim_candidates(
  File "/home/wyz/workspace/learning/planing/pluto/src/planners/pluto_planner.py", line 286, in _trim_candidates
    sorted_probability = softmax(sorted_probability)
  File "/home/wyz/anaconda3/envs/pluto/lib/python3.9/site-packages/scipy/special/_logsumexp.py", line 232, in softmax
    x_max = np.amax(x, axis=axis, keepdims=True)
  File "/home/wyz/anaconda3/envs/pluto/lib/python3.9/site-packages/numpy/core/fromnumeric.py", line 2827, in amax
    return _wrapreduction(a, np.maximum, 'max', axis, None, out,
  File "/home/wyz/anaconda3/envs/pluto/lib/python3.9/site-packages/numpy/core/fromnumeric.py", line 88, in _wrapreduction
    return ufunc.reduce(obj, axis, dtype, out, **passkwargs)
ValueError: zero-size array to reduction operation maximum which has no identity
2024-12-10 22:03:06,974 WARNING {/home/wyz/workspace/learning/planing/nuplan-devkit/nuplan/planning/simulation/runner/executor.py:34}  Simulation failed with error:
 zero-size array to reduction operation maximum which has no identity
2024-12-10 22:03:06,974 WARNING {/home/wyz/workspace/learning/planing/nuplan-devkit/nuplan/planning/simulation/runner/executor.py:38}  
Failed simulation [log,token]:
 [2021.08.17.16.57.11_veh-08_01200_01636, 9e30155b8bb55fd9]

Here is my local dir

What is causing this problem, and how should it be resolved?
Thanks all.

[zachytong]

looks like the sorted_probability array passed to the softmax function is empty. This could happen if the _trim_candidates method in pluto_planner.py is not generating any candidate trajectories or if the array is being filtered incorrectly. maybe you should check sorted_probability before softmax it and find ways to handle it.

[Tobias-lt]

I also have the same problem, may I ask that have you solved the problem? And could you please let me know how to solve it? Thanks.

[lhcstation]

@Tobias-lt @beizhengren

hi, i encountered the same issue, and i follow this #21, upgraded numpy to 1.26.4

A new error occurred when i used numpy==1.26.4: "module 'numpy' has no attribute 'bool'", i followed this #2557, replaced all dtype=np.bool to dtype=np.bool_ in pluto_feature_builder.py

Then everything seemed to be OK

By the way, if you found no output video, maybe you should follow this #9

BEST REGARDS :)