运行general_IRL时报错

githubkx commented 2 years ago

您好，很感谢您的文章与代码的分享。当我运行 python general_IRL.py 文件时，会出现以下错误：

Target Vehicle: 2617
Loading Data...
Timestep: 75
Timestep: 65
Timestep: 110
Timestep: 73
Timestep: 36
Timestep: 10
Timestep: 57
Timestep: 99
Timestep: 131
Timestep: 88
Traceback (most recent call last):
  File "G:/GitHub/Driving_IRL_NGSIM/general_IRL.py", line 94, in <module>
    max_v = np.max([traj[2] for traj in buffer_scene for buffer_scene in buffer], axis=0)
  File "<__array_function__ internals>", line 6, in amax
  File "C:\Users\Administrator\Anaconda3\envs\IRL\lib\site-packages\numpy\core\fromnumeric.py", line 2755, in amax
    keepdims=keepdims, initial=initial, where=where)
  File "C:\Users\Administrator\Anaconda3\envs\IRL\lib\site-packages\numpy\core\fromnumeric.py", line 86, in _wrapreduction
    return ufunc.reduce(obj, axis, dtype, out, **passkwargs)
ValueError: zero-size array to reduction operation maximum which has no identity

Process finished with exit code 1

在我debug时，发现代码第79行if terminated or features[-1] > 2.5: continue，所有的features[-1]值均在140左右，导致buffer恒为[]，因此无法得到max_v的值造成错误。请问您知道是什么原因引起的吗，另外运行personal_IRL.py也会造成这种错误。期待您的解答，谢谢。

Hello, thank you for sharing your article and code. When I run python general_IRL.py file, the following error occurs:

Target Vehicle: 2617
Loading Data...
Timestep: 75
Timestep: 65
Timestep: 110
Timestep: 73
Timestep: 36
Timestep: 10
Timestep: 57
Timestep: 99
Timestep: 131
Timestep: 88
Traceback (most recent call last):
  File "G:/GitHub/Driving_IRL_NGSIM/general_IRL.py", line 94, in <module>
    max_v = np.max([traj[2] for traj in buffer_scene for buffer_scene in buffer], axis=0)
  File "<__array_function__ internals>", line 6, in amax
  File "C:\Users\Administrator\Anaconda3\envs\IRL\lib\site-packages\numpy\core\fromnumeric.py", line 2755, in amax
    keepdims=keepdims, initial=initial, where=where)
  File "C:\Users\Administrator\Anaconda3\envs\IRL\lib\site-packages\numpy\core\fromnumeric.py", line 86, in _wrapreduction
    return ufunc.reduce(obj, axis, dtype, out, **passkwargs)
ValueError: zero-size array to reduction operation maximum which has no identity

Process finished with exit code 1

When I debug, I find that in line 79 of the code,if terminated or features [- 1] > 2.5: continue, the values of all features[-1] are about 140, resulting in a constant buffer= [], so maxv cannot be obtained. Do you know what caused it? Run `personal IRL. py ` can also cause this error. Looking forward to your answer, thank you.

MCZhi commented 2 years ago

Hello, thanks for your question. For some rare circumstances, due to the noise in the dataset, a target vehicle could be invalid because it could collide with others at the initial state. So I would like to suggest changing the vehicle_id under such circumstances. But I am not sure why the values of features[-1] are about 140 as I am unable to reproduce this mistake, mostly under 2.5. You may want to render the environment to see what might go wrong.

githubkx commented 2 years ago

Thank you for your answer. I have corrected the error，which is caused by data processing.

MCZhi / Driving-IRL-NGSIM

运行general_IRL时报错 #4