for inference

[johney0301]

Hello.
thank you for your help about separation research field.
It is very fantastic work.

I want to ask you about inference.
I try this command,
$ python run.py --model SepReformer_Base_WSJ0 --engine-mode test_save --out-wav-dir outputs/

but I don't know where command my input folder
is it right command like..
$ python run.py --model SepReformer_Base_WSJ0 --engine-mode test_save inputs/ --out-wav-dir outputs/

please help me.
thank you.

[MohamedBarghush]

For inference:

In the run.py file, you can find the --sample-file command, in which you can use to give the model the file you want to infer the model on, in that case, the command would be:
$ python run.py --model SepReformer_Base_WSJ0 --engine-mode infer_sample --sample-file sample_wav/sample_WSJ_mixture.wav

and you don't really need to mention the output file, because it will, by default place the output in the folder where the sample you want to infer is.

Note: The --engine-mode test and --engine-mode test_save are for testing purposes on a dataset, not on a single sample like inference