Software Engineering Courses (SECourses)•3y ago

Hi all, can you tell me how to effectively train a neural network in this style?

NeuromastersOP•3/16/23, 2:48 PM

@Furkan Gözükara SECourses I make art after your lessons))

02404-3268174644-absurdist_realist_surrealist_lovecraftian_tech_noir_survival_action_adventure_retropunk_atompunk_clockpunk_biopunk_s.jpg

Zono50•3/16/23, 2:55 PM

how many epochs is good for small dataset of face photography

Zono50•3/16/23, 3:15 PM

quick question, once learning rate shows about 0.2, should we stop training?

Uurafegit i have a question, when i try to train my Dreambooth model, the following Error ...

Furkan Gözükara SECourses•3/16/23, 3:44 PM

that looks like a training dataset error

ZZono50 i'm sorry but at 2000 training steps, my subject should not look retarded like t...

Furkan Gözükara SECourses•3/16/23, 3:45 PM

can you try with our example training dataset. this doesnt look correct at all

ZZono50 heyyy, she looks human now

Furkan Gözükara SECourses•3/16/23, 3:45 PM

yes much better

NNeuromasters Hi all, can you tell me how to effectively train a neural network in this style?

Furkan Gözükara SECourses•3/16/23, 3:45 PM

of course

NNeuromasters @Furkan Gözükara SECourses I make art after your lessons))

Furkan Gözükara SECourses•3/16/23, 3:45 PM

awesome quality

ZZono50 how many epochs is good for small dataset of face photography

Furkan Gözükara SECourses•3/16/23, 3:46 PM

this totally depends on training dataset and learning rate. so you should compare checkpoints

Furkan Gözükara SECourses•3/16/23, 3:46 PM

now i need to come up with a good thumbnail for gpt4 video

FFurkan Gözükara SECourses that looks like a training dataset error

urafegit•3/16/23, 3:49 PM

What exactly do you mean by that? I have also tried it with a random test-dataset that i downloaded. Fileformat etc shouldnt be an issue right?

Uurafegit What exactly do you mean by that? I have also tried it with a random test-datase...

Furkan Gözükara SECourses•3/16/23, 3:51 PM

it can be. i saw 1 time it was not able to read training dataset files

Furkan Gözükara SECourses•3/16/23, 3:51 PM

i dont know maybe something in text file maybe image file

FFurkan Gözükara SECourses it can be. i saw 1 time it was not able to read training dataset files

urafegit•3/16/23, 3:52 PM

Allright thanks, ill try some other things. Also i am working on MacOs, which isnt really optimal.

Zono50•3/16/23, 4:11 PM

i wasn't using 0.005, i was using 0.05 learning rate which is probably why, changed it to 0.005 and now they look much better, should i stop before it gets to 0.2 learning rate

Zono50•3/16/23, 4:12 PM

1500 steps at a 0.005 learning rate loss Strength 0.1404, wondering if i need to lower my learning rate cause it's creeping up and not sure it's trained well enough yet

Zono50•3/16/23, 4:14 PM

this is her original picture, it's getting close

00000-0-333293005_134081732718361_6488269010978870673_n.png

Zono50•3/16/23, 4:18 PM

right now i'm at 2000 steps and my loss rate is about 0.1606, is that too high? should I lower my learning rate, or just stop the training and go from here?

Zono50•3/16/23, 4:19 PM

i'm assuming it's a sort of balancing act to get it perfect

Uurafegit Allright thanks, ill try some other things. Also i am working on MacOs, which is...

Furkan Gözükara SECourses•3/16/23, 4:19 PM

ah mac os. i wonder if that is the problem

ZZono50 i wasn't using 0.005, i was using 0.05 learning rate which is probably why, chan...

Furkan Gözükara SECourses•3/16/23, 4:20 PM

yes lower learning rate can be even better

I A•3/16/23, 5:50 PM

Hello everyone, I'm new here.

I A•3/16/23, 5:51 PM

Thnk you SEOCourses, you made the best, most extensive, and ONLY valid tutorials for explaining A1111, Runpod, training etc.

I A•3/16/23, 5:51 PM

Dr. Furkan, are you the man behind those tutorials?

FFurkan Gözükara SECourses of course

NeuromastersOP•3/16/23, 5:57 PM

where to start, what is your course suitable for this?

I A•3/16/23, 6:01 PM

I'm still facing the last problem on Runpod, but I can honestly say that these tutorials are the best, and the only ones that go into details STEP-BY-STEP and you learn so much, if you have the patience to watch the tutorials, pause, and exercise.

I A•3/16/23, 6:02 PM

I hope I can get this Runpod error fixed today, though.

EssaBalls•3/16/23, 6:05 PM

Someone have any idea what prompts I can use to have my subject be generated whole on one picture? My training pictures are fine and presents whole object. But when I'm trying to generate it is just zoom in or in a different perspective

I A•3/16/23, 6:10 PM

In my experience "full body shot" or "full body portrait" but you also need to change ratio, like 512x768. In general it is not easy for me either, most images are closeups.

II A Thnk you SEOCourses, you made the best, most extensive, and ONLY valid tutorials...

Furkan Gözükara SECourses•3/16/23, 6:30 PM

hello welcome

II A Dr. Furkan, are you the man behind those tutorials?

Furkan Gözükara SECourses•3/16/23, 6:30 PM

yes i am

NNeuromasters where to start, what is your course suitable for this?

Furkan Gözükara SECourses•3/16/23, 6:30 PM

start collecting training images

Furkan Gözükara SECourses•3/16/23, 6:30 PM

hopefully my style teaching video will be published this week

II A I hope I can get this Runpod error fixed today, though.

Furkan Gözükara SECourses•3/16/23, 6:30 PM

hi can you show the runpod error you are getting

I A•3/16/23, 6:34 PM

First of all thank you for all the great work, but especially for being available to give support to your viewers.

I A•3/16/23, 6:35 PM

Many others don't even bother replying to comments on YouTube, so we appreciate that.

II A Many others don't even bother replying to comments on YouTube, so we appreciate ...

Furkan Gözükara SECourses•3/16/23, 6:35 PM

thanks

I A•3/16/23, 6:36 PM

But just 5 minutes ago I think I found what I did wrong. I was updating requirements file, not requirements versions.

I A•3/16/23, 6:37 PM

I will do another session again, and let you know...

Zono50•3/16/23, 6:38 PM

if i do lower learning rate for longer, will i get better details on my training

ZZono50 if i do lower learning rate for longer, will i get better details on my training

Furkan Gözükara SECourses•3/16/23, 6:39 PM

likely

Furkan Gözükara SECourses•3/16/23, 6:39 PM

but it will take longer time

Zono50•3/16/23, 6:40 PM

like i got really good results, but she's like, "The eyes are a little off" and what not

Zono50•3/16/23, 6:40 PM

it's really close tho

Zono50•3/16/23, 6:41 PM

may train with 0.0005 next time or somethin

FFurkan Gözükara SECourses hi can you show the runpod error you are getting

I A•3/16/23, 6:44 PM

For the sake of reporting, this was the error I was getting. Exception initializing accelerator: Accelerator.init() got an unexpected keyword argument 'project_dir'

I A•3/16/23, 6:45 PM

Maybe it can be useful. In my case after many hours online, I got easily distracted and didn't see that there are two files: requirements & requirements versions

I A•3/16/23, 6:46 PM

If the training starts and finishes correctly, I will share my experience with you Dr Furkan.

I A•3/16/23, 7:07 PM

First update: In Dreambooth optimizer 8bit Adam is not there anymore. I have 1. Torch AdamW 2. Lion

ZZono50 may train with 0.0005 next time or somethin

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

yes this requires experimentation

Hi all, can you tell me how to effectively train a neural network in this style?

Similar Threads

Similar Threads

Similar Threads