FergasunFergie
  • Joined on 2023-09-04
FergasunFergie commented on issue mrq/ai-voice-cloning#484 2024-04-16 17:47:35 +00:00
So is this abandoned?

Looking through my install notes.... I also had to do this at one point:

pip uninstall whisper pip install openai-whisper

FergasunFergie commented on issue mrq/ai-voice-cloning#484 2024-04-16 17:43:00 +00:00
So is this abandoned?

One more thing, I also have a few programs installed directly:

TorToiSe 2.4.5 D:\APPLICATIONS\ai-voice-cloning-2\ai-voice-cloning\modules\tortoise-tts dlas …

FergasunFergie commented on issue mrq/ai-voice-cloning#484 2024-04-16 17:41:19 +00:00
So is this abandoned?

Here is my requirements.txt:

git+https://github.com/openai/whisper.git

more-itertools ffmpeg-python gradio<=3.23.0 music-tag voicefixer psutil phonemizer pydantic==1.10.11 websockets…

FergasunFergie commented on issue mrq/ai-voice-cloning#484 2024-04-15 06:08:52 +00:00
So is this abandoned?

I suspect it is, yes. Is there a command that can see what versions of software are installed? I have a functional copy of the software on my computer (I haven't trained in awhile...). I could…

FergasunFergie commented on issue mrq/ai-voice-cloning#466 2024-01-11 06:04:00 +00:00
"Unsupported audio format provided: .pth"

@Atoli Your .pth files should stay in the "models" folder.

For instance \ai-voice-cloning\training\white_female3a\finetune\models It is looking for .wav files in the voices folder. …

FergasunFergie opened issue mrq/ai-voice-cloning#419 2023-10-19 18:39:52 +00:00
Double Counting Epochs?
FergasunFergie commented on issue mrq/ai-voice-cloning#248 2023-10-01 02:08:21 +00:00
Overfitting with large datasets

@DoctorPopi Re: are there background sounds or noises? Nope. I have found that running any samples through a musical vocal remover actually has helped clarify my audio samples. I already…

FergasunFergie commented on issue mrq/ai-voice-cloning#248 2023-09-26 06:36:53 +00:00
Overfitting with large datasets

@DoctorPopi How does it sound though? I am not sure how VALL-E sounds - I have stuck with using Tortoise TTS engine. My datasets are about 30 to 40.

It seems that for as well as Tortoise…

FergasunFergie closed issue mrq/ai-voice-cloning#390 2023-09-20 23:58:11 +00:00
Is Training Console Output Broken?
FergasunFergie commented on issue mrq/ai-voice-cloning#390 2023-09-19 09:08:01 +00:00
Is Training Console Output Broken?

I am training again and it's working again.

FergasunFergie commented on issue mrq/ai-voice-cloning#361 2023-09-18 23:59:58 +00:00
American imposter

@MrMustachio43 What I'm going to say may seem intuitive, because I don't know that it's explicitly written up in any of the documentation. You need to go to settings and load the finetuned…

FergasunFergie opened issue mrq/ai-voice-cloning#390 2023-09-18 07:02:13 +00:00
Is Training Console Output Broken?
FergasunFergie commented on issue mrq/ai-voice-cloning#384 2023-09-16 02:12:58 +00:00
Why so many models? And about a thousand of other questions :)

@DoctorPopi I was using 20 short clips, and it seemed like one off clip ruined the whole voice. So back to the drawing board. I am wondering if I should have just pressed forward with more…

FergasunFergie closed issue mrq/ai-voice-cloning#377 2023-09-13 06:12:46 +00:00
Attempting to run training -- libcudart.so not t found
FergasunFergie commented on issue mrq/ai-voice-cloning#377 2023-09-13 06:12:45 +00:00
Attempting to run training -- libcudart.so not t found

Reinstalled under a fresh environment. Not quite sure if the error was precisely related to not having even numbers? I'll never know, but had a clean install this time. Red flag should have…

FergasunFergie commented on issue mrq/ai-voice-cloning#379 2023-09-13 03:14:07 +00:00
Sharing a German fine-tuned model and Latin-1 tokenizer

I don't know German, but this sounds fantastic. I am wondering how whisper translation to German (not sure if that is one of the options) would turn out.

Question -- based on the Alice in…

FergasunFergie commented on issue mrq/ai-voice-cloning#377 2023-09-12 07:10:43 +00:00
Attempting to run training -- libcudart.so not t found

This is probably another artifact of my unelegant install with base tortoise-tts environment. I am going to reinstall, now that I think I understand a bit more.

FergasunFergie opened issue mrq/ai-voice-cloning#377 2023-09-12 06:25:45 +00:00
Attempting to run training -- libcudart.so not t found
FergasunFergie commented on issue mrq/ai-voice-cloning#361 2023-09-12 00:38:56 +00:00
American imposter

I set my voice chunks to 512 or 256 -- but I think the key is Temperature to 1 or very high -- as someone said about 0.75. I realized this when watching people use the tool via Youtube.

I…

FergasunFergie commented on issue mrq/ai-voice-cloning#160 2023-09-10 21:50:35 +00:00
Can't train a single good model

Use a small subset then.

With a small subset (8 clips of ~4 seconds each):

1 chunk: https://vocaroo.com/15lY8pR1WRhb 2 chunks: https://vocaroo.com/19R30vtl8gjn 4 chunks:…