fileClassification: wav file not found! (-1, -1, -1)

[preniqivjosa]

I am using pyAudioAnalysis library for voice classification app. The training works well but when I test the model for an unknown audio (by using this command: aT.file_classification("audio.wav", "model-name","model")) the model sometimes recognise the .wav file and gives the predictions, but most of the times (for the same tested audios) gives this output
fileClassification: wav file not found!
(-1, -1, -1).