I just happened to use "voice blender" on the Applio Colab notebook recently and I wanted to ask which epochs and steps to list when sharing a merged voice model. For example: I merged two different path files based on which has the best loss (i.e. 210 epochs and 2100 steps based on the g/fm scale and 190 epochs and 1900 steps based on the g/mel scale).
#Which epochs and steps should I list when sharing a merged voice model?
soft dove
dawn swift
use the epoch count of the first model
soft dove
So I should list the one based off of the path file I used that has the best loss shown on the g/fm scale?
tall hearthBOT
Ayo? @soft dove level 15 !!! 
dawn swift
sure
soft dove
I wasn't sure to list which epoch and steps count out of the two path files. Although, path files based off of g/fm and g/total do have the same count sometimes, I guess I'll list the counts from that instead of the g/mel path file then. Thanks.
hoary idol
I wasn't sure to list which epoch and steps count out of the two path files. Alt...
why not both epochs?
soft dove
why not both epochs?
I don't know. Should I list both epoch and step counts from the two different path files I merged?
hoary idol
I don't know. Should I list both epoch and step counts from the two different pa...
why must be one of them?
soft dove
I just don't know which epoch and step counts I should list when sharing them in #1175430844685484042.
hoary idol
what you share is the merged one, not one of the original ones, I don't get why only one of the original models should be mentioned
if you want to share a Gogeta, why should you only mention either Goku or Vegeta?
soft dove
Well, the post guidelines for sharing a model mentioned that my post should include number of epochs and occasionally steps. The merged one I was sharing is that of the same person using the same dataset, extraction type and architect.
hoary idol
use the epoch count of the first model
I think you and staffs should reconsider the guideline for posting merged voices