Software Engineering Courses (SECourses)•16mo ago

@Dr. Furkan Gözükara can models do like different font styles? or maybe do like font wildcards for m

@Dr. Furkan Gözükara can models do like different font styles? or maybe do like font wildcards for making alogo?

Ddxqb why do you offload if you have 48 gb?

Furkan Gözükara SECourses•9/18/24, 5:11 PM

multi gpu doesnt work even with offload

Furkan Gözükara SECourses•9/18/24, 5:11 PM

multi gpu traning overheard

MMarrvelous Please let me know if this isn't the right place for questions... I followed th...

Furkan Gözükara SECourses•9/18/24, 5:12 PM

hello welcome

Furkan Gözükara SECourses•9/18/24, 5:12 PM

swarmui developer wants you to ask in their discord

Furkan Gözükara SECourses•9/18/24, 5:12 PM

https://discord.gg/smdGbypw

Furkan Gözükara SECourses•9/18/24, 5:12 PM

please ask there

LLW24Ai Hello, can you tell me how to reduce the scales of a trained lora flux? At my ex...

Furkan Gözükara SECourses•9/18/24, 5:12 PM

fp16

Furkan Gözükara SECourses•9/18/24, 5:12 PM

other than that it will reduce quality

Furkan Gözükara SECourses•9/18/24, 5:12 PM

lora network rank 64 32 16

ZZono50 @Dr. Furkan Gözükara can models do like different font styles? or maybe do like ...

Furkan Gözükara SECourses•9/18/24, 5:12 PM

i think flux can do yes

dsienra•9/18/24, 5:19 PM

@Dr. Furkan Gözükara koha training is running much faster with the new version

dsienra•9/18/24, 5:19 PM

there is some news about this?

FFurkan Gözükara SECourses i think flux can do yes

LW24Ai•9/18/24, 5:19 PM

One user wrote to me that my workout has problems due to the weight of 1gb. For some reason, his models are 18 mb each. I have seen such models with a similar weight, but they did not bring me high-quality results and differences

LLW24Ai One user wrote to me that my workout has problems due to the weight of 1gb. For ...

Furkan Gözükara SECourses•9/18/24, 5:31 PM

single layer

Furkan Gözükara SECourses•9/18/24, 5:31 PM

reduces quality

Furkan Gözükara SECourses•9/18/24, 5:31 PM

they dont test anything properly to compare

Ddsienra @Dr. Furkan Gözükara koha training is running much faster with the new version

Furkan Gözükara SECourses•9/18/24, 5:31 PM

which version? torch 2.5?

LLW24Ai One user wrote to me that my workout has problems due to the weight of 1gb. For ...

Furkan Gözükara SECourses•9/18/24, 5:34 PM

you can use lora resize of kohya and see quality difference

FFurkan Gözükara SECourses you can use lora resize of kohya and see quality difference

LW24Ai•9/18/24, 5:36 PM

thanks

Zono50OP•9/18/24, 5:40 PM

@Dr. Furkan Gözükara on supir app, if guidance scale needs to be 2, where do i set that? Text guidance scale?

FFurkan Gözükara SECourses which version? torch 2.5?

dsienra•9/18/24, 5:53 PM

I'm not sure, I updated to train individual layers but with not a good success, now I training all layers and is super fast

dsienra•9/18/24, 5:53 PM

rn_image_picker_lib_temp_2cd97714-df2b-4b42-93c9-2f5e4e32f5c0.jpg

Ddsienra Click to see attachment

dxqb•9/18/24, 5:54 PM

resolution? bs? card?

dsienra•9/18/24, 5:56 PM

dsienra•9/18/24, 5:56 PM

1024x1024

dsienra•9/18/24, 5:57 PM

dsienra•9/18/24, 5:58 PM

dsienra•9/18/24, 5:58 PM

1024x1024 rtx 3090

dsienra•9/18/24, 5:59 PM

this is normal?

dxqb•9/18/24, 6:01 PM

no, for 1024 that's fast

dsienra•9/18/24, 6:02 PM

the checkpoint is 1.18gb so I'm sure I'm training all the layers

dxqb•9/18/24, 6:04 PM

what did you have beefore the pytorch upgrade?

dsienra•9/18/24, 6:07 PM

false alarm sorry

dsienra•9/18/24, 6:07 PM

my dataset es low ress

dsienra•9/18/24, 6:07 PM

thats why

dsienra•9/18/24, 6:08 PM

I have bukets enable

dxqb•9/18/24, 6:08 PM

ah yeah, no upscale on kohya. resolution makes a major difference on flux

dxqb•9/18/24, 6:12 PM

what resolution was it actually?

dsienra•9/18/24, 6:14 PM

some 768 some 512

dsienra•9/18/24, 6:15 PM

not all 1:1 diferent aspect ratios

FFurkan Gözükara SECourses please ask there

Marrvelous•9/18/24, 6:18 PM

Thanks. I posted there but there is a long conversation already happening so probably missed. I'll try posting again later

ZZono50 @Dr. Furkan Gözükara on supir app, if guidance scale needs to be 2, where do i ...

Furkan Gözükara SECourses•9/18/24, 8:10 PM

yes

Ddsienra Click to see attachment

Furkan Gözükara SECourses•9/18/24, 8:10 PM

good speed

Ddsienra Click to see attachment

Furkan Gözükara SECourses•9/18/24, 8:10 PM

nope

Ddsienra false alarm sorry

Furkan Gözükara SECourses•9/18/24, 8:11 PM

yep i was gonna tell this :d

TheJayHole•9/18/24, 8:39 PM

New supporter here. Wanted to say thanks. Your info is great. Couple of questions for you. I'm doing LoRA training on Flux Really, really slow on my 3090 (Windows with latest torch 2.5 speedup), 51 images in LoRA. This may have been answered somewhere, but previously with this SDXL LoRA I would do 32 repeats, 51 images, 10 epoch (16320 steps). This was a little high admittedly. The config you had would be 10200 steps Are epochs better than repeats with FLUX1.dev? Does this lower the number required? Just looking to understand that a little better. As for the slowness I'm 4735 steps and running on 1.5-2 days. I did update torch. Pretty much the same speed for me. Might have to try on RunPod. Any thoughts/advice are appreciated.

TTheJayHole New supporter here. Wanted to say thanks. Your info is great. Couple of quest...

Furkan Gözükara SECourses•9/18/24, 8:53 PM

welcome

Furkan Gözükara SECourses•9/18/24, 8:53 PM

what is your step speed?

Furkan Gözükara SECourses•9/18/24, 8:53 PM

you can use Rank_4_17250MB_4_85_Second_IT.json it is pretty decent speed

@Dr. Furkan Gözükara can models do like different font styles? or maybe do like font wildcards for m

Similar Threads

Similar Threads

Similar Threads