Closed J3inx closed 9 months ago
Possibly a duplicate of #147, can you check?
Possibly a duplicate of #147, can you check?
tring it now and so far it seems to be working, ill keep you updated on the progress but so far sorry about the duplicate, still kind of new to this stuff
Possibly a duplicate of #147, can you check?
tring it now and so far it seems to be working, ill keep you updated on the progress but so far sorry about the duplicate, still kind of new to this stuff
just tried using #147 and im still having the same problem, the tts installed correctly when doing it manually but when i try to run the install script to get all the other dependencies it still has the same problem
Does it get stuck on another dependency now then? Or is it still getting stuck on "installing TTS"
it's still getting stuck on installing tts
That's odd, are you sure it's installed? And on the latest version? It shouldn't try to reinstall when it's already installed.
That's odd, are you sure it's installed? And on the latest version? It shouldn't try to reinstall when it's already installed.
i'll run the manual tts install link again and attatch a screenshot of the output, but im pretty sure it's installed
everything is too big to put here but it's basically saying all the dependencies are already satisfied and that it is already installed
And that was inside of the venv? If it's installed in the venv it shouldn't try to reinstall?
And that was inside of the venv? If it's installed in the venv it shouldn't try to reinstall?
i tried installing it into the venv but it gave me permission errors, so I consulted chatgpt and it told me to use chmod to give the shell script the necessary permissions, then I ran the script and ran the install, and I'm pretty sure it worked, I may have accidentally installed it outside the venv so ill try uninstalling it and resintalling it while the venv is active and I don't already have a version installed
just uninstalled tts, activated the venv and installed it again, running the install.py again so will update you if it still doesnt install
so far nothing has changed, im on the same wifi I was the first time I installed tts even though it's real slow but there has been no change after running install.py
Are you running install.py or run.sh?
i'll try run.sh because i've only been running install.py
Are you running install.py or run.sh?
yeah im still getting the same problem, it's still going through the infinite tts install with no changes between run.sh and install.py
Can you activate the venv like in the issue mentioned in my first reply, and then running pip freeze
, and tell me what version of TTS is installed there, or if it's missing.
just did it, it says TTS==0.17.6
That is the latest version. I don't understand why it would try to reinstall it, as the package is installed, and doesn't care about which version is installed. It runs pip freeze
to check versions and installs, and it should have found TTS already in there, but as you're saying, it still tries to install it?
yeah, it will go into the venv, check for dependencies, then try and install TTS
Odd, i'm not sure why that would happen and why you're the first person who this happened to.
You could try deleting the venv
folder, and then running the run.sh
again. It's worth a shot, I've had it fix problems before that I wasn't sure about. The classic "have you tried turning it off and on again?" trick.
Odd, i'm not sure why that would happen and why you're the first person who this happened to. You could try deleting the
venv
folder, and then running therun.sh
again. It's worth a shot, I've had it fix problems before that I wasn't sure about. The classic "have you tried turning it off and on again?" trick.
deleted the venv folder and am running run.sh
, i'll give you an update in 10-20 mins if it doesn't work.
yeah run.sh
is still not working, same thing as install.py
it could be a problem with the terminal download thing, do you know if the terminal is the app that sends out the packets or if it is something else, because i can go into the activity manager on my mac and check the packet output to see if it is even downloading TTS at all, i did recently but it said 0 packets in and out.
It's running the install commands through subprocess.run()
, which launches it as a child process of the python process.
found a possibility of what it could be, when looking at the activity monitory i found a python
proccess but it doesnt seem to have any changes in packets sent/lost or size in total
I'll probably add a verbose flag tomorrow which will print the logs from running the commands for example. That should help with tracking down the issue.
I'll probably add a verbose flag tomorrow which will print the logs from running the commands for example. That should help with tracking down the issue.
that would be nice, it ran the whole night and is still stuck on the tts
(update it just failed the install and gave me a longwinded error code if you want me to paste that)
Yes, that error message could contain the info I need.
here is the code (this was after i deleted the venv to try the "turn off and turn back on" technique) also a part of it said there was no venv found but there was one earlier in the terminal history that i didnt copy down, so i either forgot i deleted it or something else
It appears that your internet timed out somewhere during the install. Or you were not able to reach pypi.org for another reason.
I do have plans to move TTS into an extension, as that would allow you to disable it, and it wouldn't need to install. That's not the best fix, but it would probably work.
It appears that your internet timed out somewhere during the install. Or you were not able to reach pypi.org for another reason.
I do have plans to move TTS into an extension, as that would allow you to disable it, and it wouldn't need to install. That's not the best fix, but it would probably work.
ok, so until then there isn't much i can do is there. ⎝◔╭╮◔⎠
Not that i know of, no
On the other hand, I'm starting work on moving it today, work will be done on this branch until it's ready to be merged.
The branch now has the TTS features removed, they've been moved to the extension's repository: https://github.com/gitmylo/audio-webui-TTS-extension
You can attempt switching to the TTS free branch and try to launch then. You'll be able to use audio-webui minus the Coqui TTS features.
just downloaded the branch, running the sh.run
now, so far it has skipped the tts and is working on downloading the numpy
segment, ill update you if it works
(it just finished downloading nump and moved onto pytorch so im hoping this is the end of this problem)
seems to be working, it's gone through numpy and torch so far and checking the activity monitor i see that it is constantly sending and receiving packets albeit at a slow rate due to my wifi. so for now im going to close the issue, thank you so much for your help! (╭☞◔◡◔)╭☞
running the run.sh
file and now once it gets to downloading pyworld
it immediately fails, it's had a couple failed installs so far but i keep restarting it and it keeps moving on, but i cant get past the pyworld
extension now,
Try running pip install wheel
in the venv, and also recommended in your global python install, as wheel is kind of important for a lot of packages.
Try running
pip install wheel
in the venv, and also recommended in your global python install, as wheel is kind of important for a lot of packages.
that worked, but it was kind of funky, i tried runing activateonly.sh
and used a script to check if im inside the venv, but i wasnt so i had to manually activate the venv and some other things using this
Ah yeah, I realize now that activate_only.sh
was using ./
instead of source
, it wasn't written very well, the windows side just creates a new cmd window with the environment active.
The other error might be related to some recent changes to huggingface_hub, as i had to patch some parts of it as the _check_disk_space method was not working on windows.
Actually, it might be because i removed the huggingface-hub requirement's specific version from the requirements, i'll add it back and see if that fixes it.
I've added the version-fixed huggingface-hub requirement back, and merged it into the TTS branch as well.
i'll try it out rq
also tried out the new activate_only.sh
and then ran the venv checker script and i dont think it's working still but i could easily be wrong
still didnt work
Try removing the venv folder and running the program again
Try removing the venv folder and running the program again
just started the process again after deleting the folder, i'll update you when it finishes or when anything happens
i did the wrong things earlier, i just ran it and it opened up to the webpage, so far seems to be working perfectly! i'll tell you if anything else happens
i keep getting an infinite loop on mac when running the install script. in order, it will find there is no venv, create a venv, find the packages it needs, and start installing the tts but never get past that, i left it running last night and there was no sign of progress after 8 hours of download, it doesn't seem like it is doing any sorts of downloading either
To Reproduce: be on mac air with an m2 chip on mac os somona 14.0 and run the python install script
Expected behavior: to run the install script and install all dependancis
Screenshot:
Additional context: i've tried running it a couple days ago and had success but ran into a roadblock with some other pieces because i had the wrong python installed, so i tried restarting the process to fix some things and got stuck in this problem