• https://git.ecker.tech/ aims to provide a place to share my efforts while maintaining true ownership of my code, as I do not trust GitHub.

    XMR: 4B9TQdkAkBFYrbj5ztvTx89e5LpucPeTSPzemCihdDi9EBnx7btn8RDNZTBz2zihWsjMnDkzn5As1LU6gLv3KQy8BLsZ8SG

  • Joined on 2022-10-10
ecker closed issue ecker/ai-voice-cloning#110 2023-03-10 23:23:20 +00:00
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
ecker commented on issue ecker/ai-voice-cloning#103 2023-03-10 23:19:58 +00:00
! RETRAIN YOUR MODELS !

not gonna lie, i dont really understand the graphs... what is good and what is bad lol?

mrq/ai-voice-cloning#82 (comment)

ecker commented on issue ecker/ai-voice-cloning#100 2023-03-10 23:18:25 +00:00
! POSSIBLE NEW ISSUES !

Can you use code for the speed improvements from the fast-tts github?

For tortoise itself? I'll comb through it in the event I missed anything substantial, since the last I looked at it…

ecker commented on issue ecker/ai-voice-cloning#109 2023-03-10 22:38:27 +00:00
Cannot reshape tensor / Training process hangs instead of closing

I need to validate it, but the training script should print a message when it's done training, and AIVC should use that as an indicator to kill the process. I need to validate it later.

ecker commented on issue ecker/ai-voice-cloning#107 2023-03-10 22:37:08 +00:00
./train.sh: line 12: 36088 Segmentation fault

Alright, I overhauled how metrics are relayed and stored. Tensorboard logging is no longer used, so no more segfaults.

I still need to do more validation to make sure nothing breaks, but it…

ecker pushed to master at ecker/ai-voice-cloning 2023-03-10 22:35:37 +00:00
7f2da0f5fb rewrote how AIVC gets training metrics (need to clean up later)
ecker pushed to master at ecker/DL-Art-School 2023-03-10 22:34:39 +00:00
b5c6acec9e gutted bloat loggers, now all my useful metrics update per step
ecker commented on issue ecker/ai-voice-cloning#109 2023-03-10 17:09:50 +00:00
Cannot reshape tensor / Training process hangs instead of closing

Yeah, I'm not sure what caused it to crop up for Windows training, but sometimes the training process will hang after completing the training and won't kill itself. You'll need to enter `tskill…

ecker commented on issue ecker/ai-voice-cloning#107 2023-03-10 17:06:20 +00:00
./train.sh: line 12: 36088 Segmentation fault

Busy all morning, forgot to follow up. The stack trace didn't prove much useful, as I forgot that it won't really expose anything in Python itself, oops.

How strange it's related to tensorboard.…

ecker commented on issue ecker/ai-voice-cloning#103 2023-03-10 15:39:01 +00:00
! RETRAIN YOUR MODELS !

What do you think is the best way to trasncribe japanese speech? Is there a japanese whisper model? Do you have to transcribe to katakana?

All three whisper implementations can transcribe to…

ecker commented on issue ecker/ai-voice-cloning#100 2023-03-10 14:25:47 +00:00
! POSSIBLE NEW ISSUES !

Can you use code for the speed improvements from the fast-tts github?

For tortoise itself? I'll comb through it in the event I missed anything substantial, since the last I looked at it was…

ecker commented on issue ecker/ai-voice-cloning#107 2023-03-10 14:20:37 +00:00
./train.sh: line 12: 36088 Segmentation fault

Oh nice, you were able to get me a gdb trace. I'll thumb through it when I get a chance, but for now I'll give some brief help until I can:

This is probably unrelated.

Yeah, that red block…

ecker pushed to master at ecker/ai-voice-cloning 2023-03-10 14:04:13 +00:00
df0edacc60 fix the cleanup actually only doing 2 despite requesting more than 2, surprised no one has pointed it out
ecker commented on issue ecker/ai-voice-cloning#108 2023-03-10 13:50:48 +00:00
Error when resetting to default settings

Fixed in commit 8e890d30237c2a402aad98c2fd82626a52df87c5.

Refreshing would have "fixed" it too, since it still clears the settings.

ecker closed issue ecker/ai-voice-cloning#108 2023-03-10 13:50:48 +00:00
Error when resetting to default settings
ecker pushed to master at ecker/ai-voice-cloning 2023-03-10 13:49:46 +00:00
8e890d3023 forgot to fix reset settings to use the new arg-agnostic way
ecker commented on issue ecker/ai-voice-cloning#103 2023-03-10 05:40:17 +00:00
! RETRAIN YOUR MODELS !

We could use a discussion tab for this git.

Gitea doesn't have any feature like that. Although that's on me for using it over a gitlab instance, but oh well.

Are we sure that large…

ecker commented on issue ecker/ai-voice-cloning#107 2023-03-10 04:39:38 +00:00
./train.sh: line 12: 36088 Segmentation fault

I wouldn't have a clue. I haven't ran into any segfaults from Python, so I can't debug it.

However, I see:

gen_lr_steps: [0, 0, 0, 0]

where it shouldn't be that. I doubt it's the…

ecker pushed to master at ecker/ai-voice-cloning 2023-03-10 04:27:59 +00:00
d250e0ec17 brain fried
ecker pushed to master at ecker/ai-voice-cloning 2023-03-10 04:22:51 +00:00
0b364b590e maybe don't --force-reinstall to try and force downgrading, it just forces everything to uninstall then reinstall