ReadMe not very understandable

[yangezheng]

Hello,

I'm very interested at this project and I'm trying to replicate this result, now I have my own audio recording with 2 channel, but I have a little bit problem to follow the instruction in ReadMe, I don't know if I should follow the "Table 1 results" or "Real result used in the video" sections.

Should I just put my audio files in the real_examples folder and just run the file "./cascaded_cpu.sh $myaudio" command? What is the difference compare to "python evaluate_cascaded.py"?