the difference between code and web demo when doing SS tasks

[25sunyao]

Is there any difference between the backend process of Speech Separation in the web demo and using the python demo.py script for separation? I noticed that the same audio produces significantly different results when separated.

[alibabasglab]

Hi, there shouldn't be noticeable difference. If you obtained different results from the web demo and 'python demo.py' script, please check if you have downloaded successfully the models to the directory './checkpoints/'.