After running the batch this error appears
#Applio tensorboard error
slate widget
foggy maple
the compiled version should have no issues with run-tensorboard.bat
you should not be using any other .bat files (from the doc) for applio, that's was for original rvc
@slate widget
slate widget
so should i just do the steps to install tensorboard
bleak plaza
so should i just do the steps to install tensorboard
you can simply do this:
in cmd:
`` pip install tensorboard ''
then, in a folder where you have your model's files, you wanna paste my tensorboard batch
then you can run it as so
Paste in the path and hit enter ( or few times if cmd froze )
pro tip; you can open up terminal in that folder by typing in cmd in the address bar and hit enter:
foggy maple
so should i just do the steps to install tensorboard
you dont need to do any steps, tensorboard comes included
slate widget
you can simply do this:
in cmd:
`` pip install tensorboard ''
then, in a folder ...
when i ran the pip command it said it doesnt recognize it
bleak plaza
when i ran the pip command it said it doesnt recognize it
you tried what noobies said?
slate widget
yeah
bleak plaza
you should only go my route if what he said doesn't work
@foggy maple Unless you wanna help him?
bleak plaza
when i ran the pip command it said it doesnt recognize it
In any case, that just means you have no python installed / path is configured wrong or not configured at all
foggy maple
if your fork has the same requirements, then there's nothing else is needed
run-tensorboard.bat start tensorboard with the path to logs
bleak plaza
if your fork has the same requirements, then there's nothing else is needed
No because what I sent is independent
not relying on venvs
that's why
anyways.. you can try to help him maybe
foggy maple
independent one requires python installed, then it can just be run from a commandline
bleak plaza
In any case, that just means you have no python installed / path is configured w...
ye then
foggy maple
tensorboard --logdir=G:\Applio\logs\Singing_40k
slate widget
you want me to run that in cmd?
foggy maple
that's just an example how you start it if you had tensorboard installed into the main python repo
slate widget
when i meant steps i also just meant this
foggy maple
tensorboard venv batch creates a virtual environment you dont need
but when you start tensorboard it does tell you what URL to use
slate widget
so what am i exactly supposed to do
foggy maple
so what am i exactly supposed to do
that depends on what you've done
if you ran tensorvenv.bat, use it again
slate widget
when i tried doing that it just also said this
slate widget
after installing python and running it again this also happens
foggy maple
fkinh hell
open command prompt
do where python
if your output has 1st line as C:\Users\user\AppData\Local\Programs\Python\Python311\python.exe
then next command is pip install tensorboard
if you get windowsapps version 1st, either remove this from envrironment variables path or move them down
slate widget
where do i also get to the edit environment variable from the environemtn variable area
ignore the message above i figured it out and when i did both commands and checked if there was or was not a windowsapp and then ran tensor again it said the same thing and if you need more of the top this is it
slate widget
still doesnt work
bleak plaza
still doesnt work
tried my bat?
slate widget
tried my bat?
yeah
bleak plaza
which python ver did you install?
also what numpy?
pip show numpy
pip show tensorboard
to check the versions
slate widget
1.23.5 for numpy and 2.1.0 tensorboard
bleak plaza
Hmmm..
Imma suggest you to just yeet python ( that is, if you don't use anything else that needs it )
and get 3.10.6 version
and once you have all fresh, just pip install tensorboard
I mean, yea, that's one way to approach it if you can't debug the issue
slate widget
version 3.10.6 of?
slate widget
and how do i exactly do that
bleak plaza
and how do i exactly do that
you'd simply uninstall python
after you done it, test in cmd:
py
if it works, means it works
if it says it's not recognized:
https://phoenixnap.com/kb/add-python-to-path
and re-test
foggy maple
3.11 or 3.10 does not matter
slate widget
after doing everything it just says this
when i also ran tensorenv the output just got simplified
bleak plaza
after doing everything it just says this
did you check if python's recognized in the commandline first?
simply type in 'py'
if python works but pip doesn't:
py get-pip.py
bleak plaza
3.11 or 3.10 does not matter
well ye, let's just say I am sentimental towards 3.10.6
¯_(ツ)_/¯
slate widget
if python works but pip doesn't:
> py get-pip.py
python works but when i did the pip command this showed up
slate widget
still doesnt work
slate widget
slate widget
its just a part of it cant delete it
bleak plaza
it's not hard cmon, just type it
it's darn 3 words
well
did you add python to path?
slate widget
yeah
bleak plaza
show how you set it
@slate widget
the scripts part must be missing for you
if you have to, copy it as it is just replace my username with yours ofc
@slate widget Did you do it?
No offense, I'm quite busy rn so can't be sitting here all day long
slate widget
yeah i did
bleak plaza
does it work now?
slate widget
nope
bleak plaza
what does it say?
slate widget
the exact same thing as before
slate widget
foggy maple
so you dont have python installed and added to path
bleak plaza
bro, I told you to uninstall windows python and get the normal one
bleak plaza
ye
foggy maple
fix that
open new command prompt
verify that where python returns the right thing
after that pip install tensorboard
once that installs without errors
tensorboard --logdir=c:\wherever\your\logs\are
bleak plaza
he has windows one tho
foggy maple
does not matter
bleak plaza
does for aliases
foggy maple
normal python in path would cancel it
even better if he removes windowsapps from the path
bleak plaza
well in any case, hope you can help him
as I'm a lil busy rn
foggy maple
all this could've been avoided if he checked [x] add python to path during install
bleak plaza
thing is
windows store one doesn't ask for it, afaik
so all in all, relying on win store for essential things like that is a nogo
slate widget
still says same thing about doing pip install and the where python command and yes this is in the path variable
bleak plaza
go to your path
and confirm you have anything there
C:\Users\<username>\AppData\Local\Programs\Python\Python310\
slate widget
etc and share is missing
bleak plaza
python.exe is there?
slate widget
yeah
bleak plaza
which alias you tried, py or python, in cmd
slate widget
py
slate widget
yeah
bleak plaza
try:
py -m pip
slate widget
bleak plaza
try:
where pip
slate widget
slate widget
and also for the where pip still says same thing after command
slate widget
bleak plaza
now try:
tensorboard
in cmd
slate widget
bleak plaza
well
at this point I'll just say
completely oblitare python from your pc and reinstall all fresh, this time use normal installer and check " add to path "
slate widget
same version right?
bleak plaza
- uninstall from windows store
- control panel / programs
- in location
C:\Users\<YourUser>\AppData\Local\Programs\Python\remove all folders / files in there
3.1 - optional; restart pc
- get python 3.10.6 ( or really 3.11 but I'm a fan of 3.10 so
)
4.1. ensure " add to path " is checked - test in command line you can use " py "
- try again " pip install tensorboard "
And if all of that fails? goddamn heck I dunno. would be probs the most mysterious situation I've encountered so far
slate widget
the command did work this time but it still doesnt work for the env thing so do you want me to update this or no?
bleak plaza
the command did work this time but it still doesnt work for the env thing so do ...
you can if you want
other than that, it's fine
you can now use the tensorboard bat I provided before
for the record; just paste the path to wherever you keep tensorboard files / folders with such ( if you do comparisons ) and hit enter
gonna open up in browser automatically
slate widget
website doesnt work after pasting the path to it
bleak plaza
hit enter few times in the tensorboard cmd ( because if you click on cmd, it idles
might be it froze
then alt tab to browser and done
also, example. Just in case:
I typically like to nest folders with various runs' tensorboard files, in model's folder
( just ensure they're in different folders than eval. eval is always per-current run. other / past tensorboard files should not be placed in there else your graphs will screw up
example ^
Hope all's clear now
slate widget
after the path is put in the website doesnt work after reloading and clicking enter on the cmd also wont work because of the press any key to close at the bottom
bleak plaza
well, perhaps your ports are locked
@slate widget
right click on the bat and edit
then use this as port " 5000 " ( and if failed, try: 6006 <- which is tensorboard's default iirc )
I set 25565 as default cause of minecraft lol don't ask
slate widget
none of them worked
bleak plaza
oh my bad, this one too
slate widget
did try that
bleak plaza
did try that
then I honestly don't know
search up some guide on unlocking ports on pc, then try to unlock 25565 or 6006, maybe 5000
slate widget
the only reason i need tensorboard is just to check one thing and ill probably use tensorboard like that one time and never use it again
bleak plaza
the only reason i need tensorboard is just to check one thing and ill probably u...
what thing?
also, smh, you coulda' said it in the first place
cause then we'd all save the hassle and idk, could just open the file for you on my end
slate widget
Just check how many epochs I need for one single file and the only reason I didn’t say it in the first place because I didn’t expect it to do multiple days of troubleshooting just to not work
bleak plaza
Well, it is either what Noobies said, meaning that built-in tensor works without any issue
or you'd follow my approach which grants more freedom and flexibility in evaluations
Typically users have no issue installing tensorboard and using my script but I guess your case's special
So like you know, just use whatever applio has, as noobies said
slate widget
Yeah I did like first thing and last thing but still says the same exact thing
Also yeah i unlocked all the ports tried them all out and none of them worked so I pretty much can’t use tensorboard at all and the env version also still doesn’t work
bleak plaza
now that's what you call legit unlucky
foggy maple
make sure the bottom checkbox is checked, pick 'install now'
tested on a clean win10 install
slate widget
Was 100% checked off so I probably just can’t fix the issue at all
foggy maple
if you did not check it during install, you can always manually add the paths to the Path
slate widget
Because when I ran it after I finished adding everything it pretty much said something similar to what it said at the beginning which was something along the lines of this
slate widget
I doubt it because I just installed the version code gave me
foggy maple
there's no reason for pip install tensorboard to compile anything
slate widget
I mean you could technically just check how many epochs I need for my file and be done with all this
bleak plaza
just to make sure, you did pick this one, ye?
slate widget
Yeah I did
foggy maple
so are you gonna show the full error screenshot or nah?
slate widget
Well I’m just saying to save you the hassle just let me give one of you the file check the epochs and we’ll just be done
slate widget
also my applio just stopped working like it has the same error as the tensorboard applio
so im assuming something that was troubleshooted for tensorboard messed up applio
livid gulch
also my applio just stopped working like it has the same error as the tensorboar...
if you're cooked, you can try the kaggle one
just create a notebook from Applio_kaggle.ipynb (?) in the applio repo
slate widget
and what do i do after i created the notebook
livid gulch
and what do i do after i created the notebook
perhaps like this #✨│ai-help message
it is my tweaked one tho
slate widget
so what start am i putting the code in
foggy maple
how about you just wipe whatever you have installed, simply download the compiled version of applio and unzip it
no need to fiddle with python or other libraries
foggy maple
the irony of this screenshot is the folder name 'rvc'
if it can not find rvc module, it means the rvc folder is missing in Applio's folder
either you deleted it or you not extract the compiled version properly
slate widget
Also after doing a few of the past troubleshooting and reinstalling applio, tensorboard and applio finally work but when I launch tensorboard and go to the website it says this
foggy maple
Also after doing a few of the past troubleshooting and reinstalling applio, tens...
no logs in the logs folder?
slate widget
wdym by no logs in the logs folder
forest wigeon
wdym by no logs in the logs folder
I think he meant this.
slate widget
foggy maple
i meant
slate widget
yeah i dont have that
slate widget
none of the folders also have that file
slate widget
all i did was just get the link to tensorboard from the batch file and open the link up
foggy maple
so what log do you want to see?
slate widget
scalar
foggy maple
i mean you have no log
events.out file
where is it?
you train a model, training process creates a log file
slate widget
dont have that file
slate widget
you train a model, training process creates a log file
I haven’t trained a model yet because I need to know the epochs using tensorboard which doesn’t work
foggy maple
I haven’t trained a model yet because I need to know the epochs using tensorboar...
start traiting, it writes the log, tensorboard looks at the log
and makes charts, you look at the charts and figure out when to stop
that's how it works
slate widget
So use the applio trainer?
slate widget
and makes charts, you look at the charts and figure out when to stop
i have the events.out file now but when i attempt to use tensorboard it still has the same message
livid gulch
i have the events.out file now but when i attempt to use tensorboard it still ha...
show where the tfevents file is
also have you tried tensorVENV.bat as in the docs?
(sorry kinda lazy to scroll up here 
)
slate widget
It was in the eval folder which is basically this but replace the finetune with the name of my model
foggy maple
so you start tensorboard that points to logs
slate widget
also have you tried tensorVENV.bat as in the docs?
Also yeah I have tried that but it doesn’t work
foggy maple
it should pick up all logs in the folder
slate widget
so you start tensorboard that points to logs
Can you elaborate on what you mean by that
slate widget
this is also what it looks like even with the event file
foggy maple
try running from command line like tensorboard --inspect --logdir G:\logs
it should show an output of what' wrong
like no logs found
pastel halo
I have been having this problem too my model was training too quick
slate widget
so I guess its something to do with importing
foggy maple
this all makes zero sense
slate widget
So basically I can’t do anything about it?
foggy maple
I don't know how you've fucked up your windows install to that point and whether there's any way to unfuck it
how about this
open command prompt in C:
try to reproduce this
if the above works, you can save this .bat file in C:\tensorboard folder and use it to start it
you can edit the path to the log folder as needed
slate widget
When I do the bottom command it just says “—logdir is not recognized as an internal or external command, operable program or batch file.”
foggy maple
show me a screenshot
slate widget
slate widget
how would I edit the path for the .bat file
foggy maple
edit .bat in notepad
slate widget
Yeah I did do that but it still has the original path and the path was definitely saved because I checked back at it after I saw it didn’t work and it still had my path but it didn’t on tensorboard