Yes

okay thanks guys

so yeah how do i know that i trained a good model in the pipeline

the sample inference graph in tensorboard should be a solid orange line

oh

okay

like a diagonal orange line i mean

like this for example

and what is tensor board

it's a log monitoring tool for tensorflow

oh yeah like the legacy one

mm okay thanks!

my old tacotron 2 pipeline model isn't working in new synthesizer

should i train it again from the start or use the existing model?

If it worked before, it should work still Make sure you're inputting the speaker id properly (speaker id of 0 would be 0:1 because of the synthesis notebook's speaker weighting option)

Hey can someone help me I’m trying to see the state of a voice I’m training someone told me to use synthesis notebook but I don’t know a thing about it and I think they’re too bothered to explain lol which I understand tldr: I want to to know if my voice is trained enough, how can I do that? Using tacotron 2 pipeline

I tried training using 44khz audio and got low pitched voice, what am I doing wrong?

I changed the sample_rate at the config for 44000hz

I might (and probably am) wrong but isn’t 44khz 44100 hertz

yes

Oops

But well, I don't think it would make the audio low pitched like that, just a little bit

im trying to do learning decay but the graphs dont show up, how do i lnow that my staircases are solid?

like where do i look at my graphs

oh wait nvm i think ik how

i think its with tensorboard but it wont load for me

You need to run the tensorboard cell and wait for it to load If it doesn't load, check if your adblocker is blocking it

that 2nd things prob right

@{K EY1} (Kei) im running the training cell and its still not showing the graphs

(also just realized i dont have an ad blocker for some reason)