Software Engineering Courses (SECourses)•3y ago

File "/workspace/stable-diffusion-webui/extensions/sd_dreambooth_extension/dreambooth/diff_to_sd.py"

File "/workspace/stable-diffusion-webui/extensions/sd_dreambooth_extension/dreambooth/diff_to_sd.py", line 28, in <module>
from dreambooth.dreambooth import shared as shared # noqa
ModuleNotFoundError: No module named 'dreambooth.dreambooth'

CCA_BALROG File "/workspace/stable-diffusion-webui/extensions/sd_dreambooth_extension/dream...

Furkan Gözükara SECourses•3/7/23, 7:16 PM

you must be missing a part

Furkan Gözükara SECourses•3/7/23, 7:16 PM

i sent a dm to you

Khalifa•3/7/23, 7:54 PM

hey guys, I followed the tutorial for style transfer in control net, I downloaded models, but I dont see them in my preprocessor only in models

Khalifa•3/7/23, 7:54 PM

Khalifa•3/7/23, 7:56 PM

also the yaml's didnt Dl with the models?

KKhalifa Click to see attachment

Furkan Gözükara SECourses•3/7/23, 7:59 PM

do git pull on the extension

Furkan Gözükara SECourses•3/7/23, 7:59 PM

it downloads yaml files

Khalifa•3/7/23, 7:59 PM

oh alright mb

Khalifa•3/7/23, 8:02 PM

hmm I see them now but they still don't work

Khalifa•3/7/23, 8:02 PM

Furkan Gözükara SECourses•3/7/23, 8:04 PM

show your all setup

Furkan Gözükara SECourses•3/7/23, 8:04 PM

and model folders

Furkan Gözükara SECourses•3/7/23, 8:04 PM

screenshots

Khalifa•3/7/23, 8:06 PM

could the problem be I have models folder in different folder?

Khalifa•3/7/23, 8:06 PM

I just move yaml?

KKhalifa could the problem be I have models folder in different folder?

Furkan Gözükara SECourses•3/7/23, 8:06 PM

of course problem

Furkan Gözükara SECourses•3/7/23, 8:06 PM

please follow video

Khalifa•3/7/23, 8:12 PM

works now ty

Furkan Gözükara SECourses•3/7/23, 9:02 PM

48 GB card failed

Furkan Gözükara SECourses•3/7/23, 9:03 PM

torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB (GPU 0; 47.54 GiB total capacity; 42.79 GiB already allocated; 133.12 MiB free; 46.06 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF

Furkan Gözükara SECourses•3/7/23, 9:05 PM

gonna test on 80 gb lets see

Sandstormleader•3/7/23, 9:12 PM

Furkan Gözükara SECourses•3/7/23, 9:53 PM

80 failed too :d

Furkan Gözükara SECourses•3/7/23, 9:53 PM

FFurkan Gözükara SECourses torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 4.00 GiB (GPU...

Haashim•3/7/23, 10:07 PM

This do to dataset size ?

HHaashim This do to dataset size ?

Furkan Gözükara SECourses•3/7/23, 10:07 PM

i think it is due to extension bug

Furkan Gözükara SECourses•3/7/23, 10:07 PM

i opened a thread

Furkan Gözükara SECourses•3/7/23, 10:08 PM

but this machine consumes a lot credits

FFurkan Gözükara SECourses i think it is due to extension bug

Haashim•3/7/23, 10:08 PM

This is why I am looking at an alternative to the extension for training.
I am not good enough with code to keep up with fixes

Furkan Gözükara SECourses•3/7/23, 10:09 PM

true

dave5555•3/7/23, 10:12 PM

i have been using stable tuner for on vast or auto1111 on vast for training, currently both repos are broken for my setup

dave5555•3/7/23, 10:13 PM

can anyone help me translate the google colab kohya notebook to a jupyter notebook i can run on vast?

Ddave5555 i have been using stable tuner for on vast or auto1111 on vast for training, cu...

Haashim•3/7/23, 10:14 PM

Stabletuner is broken? Or cast giving issues?

Haashim•3/7/23, 10:14 PM

I was looking at everydream next possibly on Runpod

dave5555•3/7/23, 10:15 PM

i need to get pull stable tuner local to try the notebook export but dev mentioned in his discord he needed to fix something about notebooks

dave5555•3/7/23, 10:16 PM

i have a working notebook from stable tuner, i wish i could just add in offset noise

dave5555•3/7/23, 10:17 PM

is that possible to just add a little code to change the noise offset to my notebook?

Furkan Gözükara SECourses•3/7/23, 10:17 PM

why dont you use dreambooth?

Furkan Gözükara SECourses•3/7/23, 10:17 PM

on automatic1111

Furkan Gözükara SECourses•3/7/23, 10:17 PM

it works great

dave5555•3/7/23, 10:18 PM

it works good for lora on my 1080, but i want to use vast to not tie up my comp , and also to do normal finetune not lora

dave5555•3/7/23, 10:18 PM

i have been able to run auto1111 on vast up until now i get error

dave5555•3/7/23, 10:19 PM

Traceback (most recent call last):
File "/workspace/sd/launch.py", line 361, in <module>
start()
File "/workspace/sd/launch.py", line 356, in start
webui.webui()
File "/workspace/sd/webui.py", line 240, in webui
app.add_middleware(GZipMiddleware, minimum_size=1000)
File "/workspace/env/lib/python3.10/site-packages/starlette/applications.py", line 135, in add_middleware
raise RuntimeError("Cannot add middleware after an application has started")
RuntimeError: Cannot add middleware after an application has started
[11:57 AM]
trying to run auto1111 on vast , its always worked before , but havent done it in a while ....... it loads the model weights , creates the remote gradio, all looks like usual , but then it adds on this code and kills the port

dave5555•3/7/23, 10:19 PM

i found a thread about this error and people mention changing fast api

dave5555•3/7/23, 10:20 PM

the requirements.txt has correct version

dave5555•3/7/23, 10:20 PM

i tried pip reinstall fast api and still get error

Ddave5555 i need to get pull stable tuner local to try the notebook export but dev mention...

Haashim•3/7/23, 10:21 PM

Have you noticed a difference in results using stabletuner vs a1111 for training ?

dave5555•3/7/23, 10:23 PM

its hard to say , some people in stable tuner mention its better quality, im not sure ...... for me the notebook runs more reliably and quicker setup than installing auto1111 to vast so thats why i used it

File "/workspace/stable-diffusion-webui/extensions/sd_dreambooth_extension/dreambooth/diff_to_sd.py"

Similar Threads

Similar Threads

Similar Threads