requirements mismatch? about alltalk_tts

[erew123]

The requirements file is actually specifying greater than or equal to:

torch>=2.1.0+cu118

torch > (means greater than) = (or equal to) 2.1.0+cu118

So its demanding a minimum level of Torch 2.1.0 with cuda 11.8 or greater.

So torch 2.1.2 with cuda 12.1 is absolutely fine and the requirements wont complain at that, it will be a match for what the requirements file is requesting.

[erew123]

If you have a problem beyond that, please double check this list I'm assuming that's what you mean you've been through?

Id need to know what your start-up error is?

[erew123]

If you continue having problems, you're welcome to update AllTalk

and you can then generate me a diagnostics log file so I can help you figure out what's going on.

All the best!

[theinhumanity]

does this help?
diagnostics.log

[erew123]

Yes, thanks for that!

Your installed versions of packages are all good! No problems there I can see.

So 2x things I do spot. You're very low on system RAM. Im not sure how much of a problem this will be, however, using the Low VRAM function would probably be an issue for you, if you truly only have the 1.25GB's of memory free. As the LowVRAM option it uses 2GB of system ram (shouldn't be an issue if you are just using it normally) See my post below this one though re rebooting.

Im not sure if your system ram being so low would/could be an issue, but Im guessing you must have some things already loaded up.

INFO:root:System RAM: 1.25 GB available out of 7.97 GB total

As for the second thing, I'm assuming you didn't run these diagnostics while text-generation-webui was running?

INFO:root:Port Status: Port 7851 is in use.

As I see that port 7851 is already in use. Which would either mean AllTalk has started running its web service OR you actually have something else running on that port.

If Text-generation-webui with AllTalk wasn't running... then try following step 5 https://github.com/erew123/alltalk_tts#-alltalk-startup-warning-tts-subprocess-has-not-started-up-yet-will-keep-trying-for-60-seconds-maximum-please-wait-it-times-out-after-60-seconds

Otherwise, please let me know what the problem you get is?

[erew123]

Oh, by the way, if it had crashed in some way, that could explain why your System Ram was so full... possibly. So it may be worth a reboot of your computer and try again (though change the port number too).

[theinhumanity]

i figured it out! it turns out my ssd/ram/cpu were so slow it took 70s to load the model. Setting the timeout to 120s fixed the problem for me.

[erew123]

wow, 70 seconds... ok then! ok.. well... I guess that might explain someone else's problem that I've seen.

I guess Ill have to change a few bits around for the next release.... damn... that means I have to re-write bits of documentation :/ ahh well! hah!

Glad you are sorted though!