Software Engineering Courses (SECourses)•3y ago

Hi, can anybody help me? I just watched one of your videos (Realistic Photos By Kohya LoRA Stable Di

Hi, can anybody help me? I just watched one of your videos (Realistic Photos By Kohya LoRA Stable Diffusion Training) and it says to just type man or woman to generate regularization images. When I type woman 90% of the images are nudes, however, I noticed that all your man images had clothes will this affect my outcome?

IIsaac Brock how can people have so many epochs? i do 4k steps in 3 epochs is this wrong?

Furkan Gözükara SECourses•5/1/23, 10:49 AM

are you doing this on kohya ss? what is your repeating count?

KKhang Hi Dr.Furkan I have watched most of your Dreambooth videos but I didn't see anyt...

Furkan Gözükara SECourses•5/1/23, 10:49 AM

i prefer to not use it if possible because i remember there once was a bug with it which prevented training

P1pSl1p•5/1/23, 11:28 AM

rpg_v4 + controlnet + multi_frame video rendering

P1pSl1p•5/1/23, 11:28 AM

from your tutorial https://youtu.be/kmT-z2lqEPQ

YouTubeSECourses

Video To Anime - Generate An EPIC Animation From Your Phone Recordi...

RunPod: https://bit.ly/RunPodIO. Discord : https://bit.ly/SECoursesDiscord. Turn your real video into an animation with just 1 click by using Stable Diffusion for free. If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on

https://www.patreon.com/SECourses

Playlist of Sta...

P1pSl1p•5/1/23, 11:28 AM

using a custom LoRa and going for more photoreal

P1pSl1p•5/1/23, 11:29 AM

working on a new render now to add more facial expression using two controlnets

PP1pSl1p working on a new render now to add more facial expression using two controlnets

rafraf•5/1/23, 11:52 AM

So good

P1pSl1p•5/1/23, 11:54 AM

thank you. it's been a loooong day.

P1pSl1p•5/1/23, 11:58 AM

it'll be 6 hours till the render is done for the new video. fingers crossed for a touch of temporal coherence

CChary Hi, can anybody help me? I just watched one of your videos (Realistic Photos By ...

TheWhykiki•5/1/23, 11:58 AM

Hey had the same yesterday after watching the tutorial. Have you looked at the prompt of the generated images?
I saw only by chance later that there were words in my prompt that I did not enter

This was due to the extension Unprompted, which I once installed. This added keywords like nude and modelshoot

TheWhykiki•5/1/23, 11:59 AM

Even after deactivation the prompt was added, so I had to restart automatic

PP1pSl1p it'll be 6 hours till the render is done for the new video. fingers crossed for ...

rafraf•5/1/23, 12:55 PM

what kind of progress are you making with temporal coherence, is this a holy grail, or do you see smoothness in the horizon? Just wondering if its readily available provided time is invested. But what is the timeframe, a couple of days, weeks? mad scientist dedication?

rafraf•5/1/23, 12:55 PM

Just wondering if worth getting into for myself

Furkan Gözükara SECourses•5/1/23, 1:15 PM

automatic1111 released a new release that is good

Rrafraf what kind of progress are you making with temporal coherence, is this a holy gra...

P1pSl1p•5/1/23, 1:22 PM

I'm getting okay results. Really hit and miss depending on how complex the generation is. I'm by no means an expert at this stuff so it's a lot of trial and error. As for when we'll achieve true temporal coherence? Could be tomorrow could be next year. But I'm sure there's people much smarter than me working on it right now. I'm trying to learn it because I'm setting goals for myself with stable diffusion. Rather than just endlessly generating images. I want to progress with the tech.

mayen•5/1/23, 1:25 PM

I keep getting errors like,

CUDA out of memory. Tried to allocate 20.00 MiB (GPU 0; 10.76 GiB total capacity; 4.29 GiB already allocated; 10.12 MiB free; 4.46 GiB reserved in total by PyTorch)

I have a 8gb 3070ti, is there any reason for me not to be able to use Konya correctly?

PP1pSl1p I'm getting okay results. Really hit and miss depending on how complex the gener...

rafraf•5/1/23, 1:25 PM

good to know. it's good to focus on things, to not get caught up in the noise of everything that is happening

Rrafraf good to know. it's good to focus on things, to not get caught up in the noise of...

P1pSl1p•5/1/23, 1:39 PM

For sure

FFurkan Gözükara SECourses automatic1111 released a new release that is good

P1pSl1p•5/1/23, 1:41 PM

Is this the fabled update that supports torch2.0?

PP1pSl1p Is this the fabled update that supports torch2.0?

Furkan Gözükara SECourses•5/1/23, 1:43 PM

yep

Furkan Gözükara SECourses•5/1/23, 1:43 PM

it was already supporting

Furkan Gözükara SECourses•5/1/23, 1:43 PM

i already have shown multiple times

Furkan Gözükara SECourses•5/1/23, 1:43 PM

but now out of box

FFurkan Gözükara SECourses it was already supporting

P1pSl1p•5/1/23, 1:44 PM

Yeah. I followed one of your tutorials that had torch2.0. I got it set up and then a git pull reverted me right back. Lol. Didn't have the heart to go through that again. I'll update in the morning when my render is done.

PP1pSl1p Yeah. I followed one of your tutorials that had torch2.0. I got it set up and th...

Furkan Gözükara SECourses•5/1/23, 1:49 PM

ah i see what you mean. i dont do automatic updates D:

Furkan Gözükara SECourses•5/1/23, 1:49 PM

would you guys be interested in deep voice cloning?

TheWhykiki•5/1/23, 1:50 PM

I'm pulling my hair out here, where does the calculation come from that you see in the text file video????

https://youtu.be/mnCY8uM7E50?t=1153

YouTubeSECourses

Transform Your Selfie into a Stunning AI Avatar with Stable Diffusi...

Our Discord : https://discord.gg/HbqgGaZVmr. How to do free Stable Diffusion DreamBooth training on Google Colab for free. If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on

https://www.patreon.com/SECourses

Playlist of Stable Diffusion Tutorials, Automatic1111 and Goo...

TTheWhykiki I'm pulling my hair out here, where does the calculation come from that you see ...

Furkan Gözükara SECourses•5/1/23, 1:51 PM

it is totally based on experience of users

Furkan Gözükara SECourses•5/1/23, 1:51 PM

but the best way is

TheWhykiki•5/1/23, 1:51 PM

I tried to get ChatGPT to explain dreambooth_training.py to me.
So far this worked well.
Then I asked what values I should put here, starting from 15 images of mine:

lr_warmup_steps, num_class_images, sample_batch_size, max_train_steps,save_interval

After analyzing all the code, I thought she should know on what bases the calculation is based. But apparently it doesn't

According to the calculation in the video it should be 15*12=180. But ChatGPT says 15...

Furthermore she says:

The number of class images --num_class_images should be set to 12, because this is the number of training images you have for your class.

The --sample_batch_size depends on how much GPU memory you have available. A batch size of 4 is relatively small and should be fine for most GPUs, but it depends on your specific configuration.

The --max_train_steps should be at least large enough to go through all the training frames once. Since you have 12 training frames and use a batch size of 1, you should set at least 12 steps, e.g. --max_train_steps=12.

The --save_interval depends on how often you want to save the model weights. A good rule is to save every few epochs. Since you only have 12 training frames, you might not want to train too many epochs to avoid overfitting. For example, if you want to train 3 epochs, you can use --max_train_steps=36 and --save_interval=12 to save the model after each epoch.

Furkan Gözükara SECourses•5/1/23, 1:51 PM

having checkpoints and comparing them

TTheWhykiki I tried to get ChatGPT to explain dreambooth_training.py to me. So far this work...

Furkan Gözükara SECourses•5/1/23, 1:52 PM

what gpu do you have?

TheWhykiki•5/1/23, 1:52 PM

Trying the Google Colab currently.

TheWhykiki•5/1/23, 1:53 PM

On my local machines a RTX2080 Ti and on my Laptop RTX3070 8GB. I searching for a good eGPU

TheWhykiki•5/1/23, 1:53 PM

On Colab its the T4 - I used the link from the video

P1pSl1p•5/1/23, 1:54 PM

3070 8gb is pretty good for local training. Just make sure you have a good cooling pad for the laptop.

TheWhykiki•5/1/23, 1:54 PM

I can use Dreambooth on the RTX2080 - I tested it, but my results all where bad

Not because of the machine but because of my knowledge

P1pSl1p•5/1/23, 1:54 PM

What are you trying to train? Yourself?

TheWhykiki•5/1/23, 1:55 PM

Yes at first myself and my son. And afterwards some Web UI

TTheWhykiki I can use Dreambooth on the RTX2080 - I tested it, but my results all where bad ...

Furkan Gözükara SECourses•5/1/23, 1:55 PM

well you have much better options

Furkan Gözükara SECourses•5/1/23, 1:55 PM

with rtx 2080 because you can easily save checkpoints

Furkan Gözükara SECourses•5/1/23, 1:55 PM

and compare them

Furkan Gözükara SECourses•5/1/23, 1:56 PM

have you installed dreambooth?

Furkan Gözükara SECourses•5/1/23, 1:56 PM

i can quickly install dreambooth on your machine

TTheWhykiki Yes at first myself and my son. And afterwards some Web UI

P1pSl1p•5/1/23, 1:56 PM

I've had much better success with KohyaSS when training people.

https://youtu.be/TpuDOsuKIBo

YouTubeSECourses

Generate Studio Quality Realistic Photos By Kohya LoRA Stable Diffu...

#Kohya SS web GUI DreamBooth #LoRA training full tutorial. You don't need technical knowledge to follow this tutorial. In this tutorial I have explained how to generate professional photo studio quality portrait / self images for free with Stable Diffusion training.

Our Discord server