Is it possible to split denoiser module?

[EuphoriaCelestial]

Hi, my previous Tactron 2 model has some high pitch noise in it, and I can remove it using denoiser in Waveglow, it work quite well.
Can I use that function without Waveglow? Because I am trying some other TTS core, like Fastspeech 2, and I want to use Waveglow's denoiser on their output audio as well
If no, please explain how this module work, I will try to recreate it. I tried to read the code but not yet fully understand the pipeline