Fbut first improve your training dataset
GThanks. ADetailer right?
Fyep
GDo you know what kind of bad images causes weird eyes in your model? I don't know how to improve my dataset further
Fimage quality focus lightning distance
F@helen i am checking your issue right now
F@helen here images they are working. there must be some step you skip. did you make container size 20 gb before install?
F
F
Fi tested both templates
HOK, I'll try again and check container size
thank you
thank youFyou are welcome.
Falso pick rtx 3090
DHey @Furkan Gözükara SECourses
Please I've been having this issue and don't know how to fix it
Could you help?
Please I've been having this issue and don't know how to fix it
Could you help?
DThe error has changed after editing the webui_user.bat file to this
Fjust replied you in qwen-flux-wan-such
Fi can't make use of roop/insightface/whatever in runpod, does anyone know why? it's banned or something? CUDA execution provider doesn't appear
Ftry auto installer working i tested yesterday : https://www.patreon.com/posts/auto-installer-84511510
Patreon
Get more from SECourses: Tutorials, Guides, Resources, Training, MidJourney, Voice Clone, TTS, ChatGPT, GPT, LLM, Scripts by Furkan Gözükara on Patreon
Fi'm not using roop. i created my own app using insightface
Fbut mine uses more or less the same stuff
Fit's just that i can't get the cuda execution provider here, dunno why
Fi see dont know
Fwell, good to know it should work xD
Ftrying to fine tune
FGitHub
I am trying to make training work on first official raw dataset of LJSpeech-1.1 My aim is fine tuning a single voice to get eleven labs level quality Then hopefully I will release this as a tutoria...
LHere's a fun little commandline client I made for ChatGPT:
minigpt.rb2.02KB
Ohii,anyone knows why ozen its transcribing my uploaded voice for training in different language instead of the speaking language as how the whisper is doing it
Oi have edited ozen.py and modified the default model to medium and also changed from cuda to cpu as i dont have a gpu at the moment.Thanks!
OLightning automatically upgraded your loaded checkpoint from v1.5.4 to v2.1.0. To apply the upgrade to your files permanently, run
Model was trained with pyannote.audio 0.0.1, yours is 3.0.1. Bad things might happen unless you revert pyannote.audio to 0.x.
python -m pytorch_lightning.utilities.upgrade_checkpoint C:\Users\Hp\.cache\torch\pyannote\models--pyannote--segmentation\snapshots\2ffce0501d0aecad81b43a06d538186e292d0070\pytorch_model.binModel was trained with pyannote.audio 0.0.1, yours is 3.0.1. Bad things might happen unless you revert pyannote.audio to 0.x.
Ocould be because of pyannote?
Fi dont know this
Fyou can edit files to change language
Faccurate
Ohmm,i will try to find out,as i said i set the model to medium so its ok with multilingual audios...i thinked maybe the training is done in english somehow and thats why the audio was transcribed in english, what about the traing proccess its working only if i have gpu right?as i tryed to start the training and i receive error that i there is no gpu
JHello.
Startup problem - Auto Windows Installer For SDXL Background Replacement for Product Images
A matching Triton is not available, some optimizations will not be enabled.
Error caught was: No module named 'triton'
Initializing captioner...
Initializing upscaler...
Initializing segmenter...
Initializing depth estimator...
Initializing depth ControlNet...
Initializing autoencoder...
Initializing SDXL pipeline...
Loading pipeline components...: 100%|████████████████████████████████████████████████████| 7/7 [00:08<00:00, 1.26s/it]
Traceback (most recent call last):
File "D:\pimages\background-replacement\app.py", line 150, in <module>
num_inference_steps = gr.Slider(precision=0,minimum=1, maximum=300,value=30, default=30,step=1, label="Num Inference Steps")
File "D:\pimages\background-replacement\venv\lib\site-packages\gradio\component_meta.py", line 146, in wrapper
return fn(self, **kwargs)
TypeError: Slider.init() got an unexpected keyword argument 'precision'
To continue, press any key. . .
help me solve the problem
Thank you
Startup problem - Auto Windows Installer For SDXL Background Replacement for Product Images
A matching Triton is not available, some optimizations will not be enabled.
Error caught was: No module named 'triton'
Initializing captioner...
Initializing upscaler...
Initializing segmenter...
Initializing depth estimator...
Initializing depth ControlNet...
Initializing autoencoder...
Initializing SDXL pipeline...
Loading pipeline components...: 100%|████████████████████████████████████████████████████| 7/7 [00:08<00:00, 1.26s/it]
Traceback (most recent call last):
File "D:\pimages\background-replacement\app.py", line 150, in <module>
num_inference_steps = gr.Slider(precision=0,minimum=1, maximum=300,value=30, default=30,step=1, label="Num Inference Steps")
File "D:\pimages\background-replacement\venv\lib\site-packages\gradio\component_meta.py", line 146, in wrapper
return fn(self, **kwargs)
TypeError: Slider.init() got an unexpected keyword argument 'precision'
To continue, press any key. . .
help me solve the problem
Thank you
J
Fwhat is your python version
Fcan you private message me the image you are trying your settings
Fi am gonna do a fresh install and test now
Fdoing a fresh install
J
F3.10.6 looking good
Fwhere did you install inside c drive?
Fah i think the reason is
Fgradio version
Jdrive D
Fyep
Fthat must be gradio
Fgoing to fix noıw