Windows fatal exception: stack overflow error

[elchaima1234]

To calculate the PESQ between the original audio (input to the model) and the reconstructed audio (output of the model), I used the following code:

from pypesq import pesq
original_audio, _ = librosa.load("original_audio.wav", sr=FS)
reconstructed_audio, _ = librosa.load("reconstructed_audio.wav", sr=FS)
curr_pesq = pesq(original_audio, reconstructed_audio, FS)
print("PESQ:", curr_pesq)

Here is some necessary information about the two audio files (.wav):

Original Audio Length: 32193 Watermarked Audio Length: 32193 Sample Rate: 16000 Original audio type: float32, shape: (32193,) Watermarked audio type: float32, shape: (32193,)

However, every time I try to compute the PESQ, I encounter a "Windows fatal exception: stack overflow" error.

I have installed the required package, but I am unsure what the issue might be.

Has anyone experienced this problem before? Interestingly, it works on Google Colab, but when I use Spyder on my local machine, I get this error.