Hi guys!

I have trained some fillers, but there is a problem with the fillers and the silence, sometimes on the results are fillers where I havent said nothing. So I have read something about its posible to absorb small noise with silence increasing a parameter on the configuration file, but I cant find this parameter.

Does someone know about this parameter?

Thanks

--- (Edited on 4/27/2017 5:39 am [GMT-0500] by kibkib) ---

Depends on the decoder largerly, you didn't mention which decoder do you use.

--- (Edited on 5/1/2017 10:20 [GMT+0400] by nsh) ---