Software Engineering Courses (SECourses)•15mo ago

also how do I switch fp16 models in flux without this message

BBrennenRB Hello all- I am looking for help with a project relating to flux and creating se...

xlads•10/18/24, 5:59 AM

Hello, sent you a request

xlads•10/18/24, 6:00 AM

@BrennenRB

PPelolisu Testing Style transfer function in new ui, is being born.

xlads•10/18/24, 6:00 AM

Interesting!

Jackie•10/18/24, 6:02 AM

I have a dataset of 25 images for this character, and I’m trying to train a LoRA for FLUX. I used the best config with Rank 4 and set the epoch to 100, which resulted in 2500 steps. However, the results from the prompt don’t look like the character I want to portray. Should I increase the epoch to 200? Also, I’m using captions generated by ChatGPT like this: 'GDZHBo_flx, a cartoon character with a large mustache and glasses, sitting with a playful expression as he appears to be gesturing something with his hands, adding a humorous and relaxed vibe.' Any advice?

Jackie•10/18/24, 6:07 AM

This is an example of the results I’m getting after training the character. The outputs don’t quite match the character I want

Ddruhl Can I have SwarmUI find models in my comfy unet folder? lol

druhl•10/18/24, 6:55 AM

This works (Saved me tons of space on home PC):

You should be able to change ModelRoot to whatever place on your system you have your models stored and, if the subdirs vary from what SwarmUI uses, modify them here. That should capture unet. It will look in {ModelRoot}/unet or {ModelRoot}/diffusion_models as needed.

AAIGambino dr for the style, what happens if you add in like I dont know 5,000 images for t...

Furkan Gözükara SECourses•10/18/24, 7:58 AM

sounds too much but can be tested

PPelolisu Testing Style transfer function in new ui, is being born.

Furkan Gözükara SECourses•10/18/24, 7:59 AM

nice

Ddeepblue_1337 and do you recommend masked training over "normal" training?

Furkan Gözükara SECourses•10/18/24, 7:59 AM

i dont recommend

Furkan Gözükara SECourses•10/18/24, 7:59 AM

get your training set better :d

Furkan Gözükara SECourses•10/18/24, 7:59 AM

but i plan to test

Furkan Gözükara SECourses•10/18/24, 7:59 AM

on flux too

NNick44 also how do I switch fp16 models in flux without this message

Furkan Gözükara SECourses•10/18/24, 8:00 AM

you can convert model to fp8

Furkan Gözükara SECourses•10/18/24, 8:00 AM

https://www.patreon.com/posts/how-to-convert-114003125

Patreon

How To Convert 24 GB FP16 Fine Tuning / DreamBooth FLUX Checkpoints...

Get more from SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts on Patreon

JJackie I have a dataset of 25 images for this character, and I’m trying to train a LoRA...

Furkan Gözükara SECourses•10/18/24, 8:00 AM

dont caption in details

Furkan Gözükara SECourses•10/18/24, 8:00 AM

it reduces likeliness

Furkan Gözükara SECourses•10/18/24, 8:01 AM

and you already dont have many images

FFurkan Gözükara SECourses and you already dont have many images

Jackie•10/18/24, 8:39 AM

thank, I will try with more images about 40 images, should I keep epoch 200 or lower epoch?

JJackie thank, I will try with more images about 40 images, should I keep epoch 200 or l...

Furkan Gözükara SECourses•10/18/24, 8:41 AM

200 good

Furkan Gözükara SECourses•10/18/24, 8:41 AM

so you can be sure it is trained enough

snowmonkey•10/18/24, 8:52 AM

I created a pod from a network volume and now want to store files in the volume

snowmonkey•10/18/24, 8:52 AM

where do I find the network volume inside the pod when I'm ssh'd in ?

Ssnowmonkey where do I find the network volume inside the pod when I'm ssh'd in ?

Furkan Gözükara SECourses•10/18/24, 9:03 AM

you can ask this in runpod server i never done

Furkan Gözükara SECourses•10/18/24, 9:03 AM

https://discord.gg/uJ3CXXHb

Furkan Gözükara SECourses•10/18/24, 9:04 AM

i just found a new error of bmaltais

Furkan Gözükara SECourses•10/18/24, 9:04 AM

never ending :d

Furkan Gözükara SECourses•10/18/24, 9:04 AM

https://github.com/bmaltais/kohya_ss/issues/2908

GitHub

Urgent fix needed ImportError: cannot import name 'cached_download'...

Hugging face hub has to be 0.25.2 or doesnt work man these errors always happens to me when recording a tutorial video @kohya-ss @bmaltais

FFurkan Gözükara SECourses dont caption in details

tazztone•10/18/24, 10:23 AM

i would disagree partially. u wanna caption clothing, background etc.
but not for what the lora should learn from the images

Ttazztone i would disagree partially. u wanna caption clothing, background etc. but not f...

Furkan Gözükara SECourses•10/18/24, 10:35 AM

well flux has internal captioning mechanism

Furkan Gözükara SECourses•10/18/24, 10:36 AM

that is why i dont recommned for flux

tazztone•10/18/24, 10:58 AM

it's just.... i trained same 10img lora of myself once without captions and once with minimal captions. and it's so hard to say which is better side by side... subjective.
i need some more xy grids

Ttazztone it's just.... i trained same 10img lora of myself once without captions and once...

Furkan Gözükara SECourses•10/18/24, 11:03 AM

true

FFurkan Gözükara SECourses well flux has internal captioning mechanism

tazztone•10/18/24, 11:15 AM

did BFL tell you? ^^

Ttazztone did BFL tell you? ^^

Furkan Gözükara SECourses•10/18/24, 11:31 AM

well it is discovered by several scripts developers

Furkan Gözükara SECourses•10/18/24, 11:31 AM

by analyzing its architecture

Furkan Gözükara SECourses•10/18/24, 11:31 AM

it generates image embeddings

Furkan Gözükara SECourses•10/18/24, 11:31 AM

which makes captioning effect

tazztone•10/18/24, 11:31 AM

interesting. thanks for sharing

ParadoxNeon•10/18/24, 12:40 PM

@Dr. Furkan Gözükara Hello everyone, I have a question about custom lora training on kohya using flux
I trained my face with 21 photos and 142 epochs as you mention on the equation on your video, which is... 3.000 / (number of photos) = number of epochs.
When I write the prompt on SwarmUI using Flux.1-dev-fp8. with 1024x1024 resolution the picture looks with low quality, also when the size is 1024+ seems to appear some artifact vertical lines on the picture (see the image) . I'm thinking to train again from zero a new lora with more images or use finetuning. What do you recommend?