Help with Inference

[rakshav-flwls]

Hey ! I am trying Speech Separation on 5 speaker and i have a single wav file , how can i use svoice to infere on single wav audio ?

Or the only way to use is to train the model and then evaluate ? or we can also use trained model to infer on single wav file ?

[muhammad-ahmed-ghani]

Hey ! I am trying Speech Separation on 5 speaker and i have a single wav file , how can i use svoice to infere on single wav audio ?

Or the only way to use is to train the model and then evaluate ? or we can also use trained model to infer on single wav file ?

For 5 speakers you need to train the model and then do the inference as there is no 5 source separation pre-trained model available in my knowledge. For 3 speakers you can use this model: https://huggingface.co/speechbrain/sepformer-libri3mix