Software Engineering Courses (SECourses)•17mo ago

@Dr. Furkan Gözükara 20 epochs 0 resemblance to the training characters (multiple people) using reg

@Dr. Furkan Gözükara 20 epochs 0 resemblance to the training characters (multiple people) using regularization class person (people of all ages and genders) TE and unet lr 1e-4, Epoch 45 I'm getting some resemblance, so I increased now lr of the TE to 9e-4, lets see

Ddsienra @Dr. Furkan Gözükara 20 epochs 0 resemblance to the training characters (multip...

Furkan Gözükara SECourses•8/30/24, 3:58 PM

you need to train like 200 to see more resemblance . did you look my grids on patreon for reg images?

Furkan Gözükara SECourses•8/30/24, 3:58 PM

recording runpod and massed compute tutorial atm

FFurkan Gözükara SECourses recording runpod and massed compute tutorial atm 😄

dsienraOP•8/30/24, 4:02 PM

Yes I did, mixed faces, I'm playing with the TE lr to see its effects, I want to know if something fix or minimize the bleeding, also I'm not using class on my characters

dsienraOP•8/30/24, 4:02 PM

I'm just experimenting

dsienraOP•8/30/24, 4:04 PM

TE lr 9e-4 is quite aggressive but may be is what it needs idk

Furkan Gözükara SECourses•8/30/24, 4:35 PM

nice let me know

FFurkan Gözükara SECourses nice let me know

uisato•8/30/24, 4:57 PM

Do you happen to know why there's little to no improvement when using Torch 2.5? Said it a couple of messages ago.

uisato•8/30/24, 4:58 PM

Double-checked, I'm using Torch 2.5. [Turned off the validating requirements part in gui.bat]

daroth•8/30/24, 5:00 PM

wonder if this is a LORA trained on a single pic:
https://civitai.com/models/702897/crazy-face-peter-parker-meme-face

nicearoni•8/30/24, 5:51 PM

long ass post was so long, converted it to image haha

1241-photo_of_a_marching_band_practicing_outs-Fluxflux1-dev-fp8-1547391564.png

1247-photo_of_a_marching_band_practicing_outs-Fluxflux1-dev-fp8-1547391564.png

1252-a_group_of_students_wearing_red_blue_y-Fluxflux1-dev-fp8-1565880735.png

1251-a_group_of_students_wearing_red_blue_y-Fluxflux1-dev-fp8-1565880735.png

1304-a_marching_band_rehearsal_is_taking_plac-Fluxflux1-dev-fp8-1565880735.png

1254-a_marching_band_rehearsal_is_taking_plac-Fluxflux1-dev-fp8-1565880735.png

Nnicearoni long ass post was so long, converted it to image haha

Dreamer@TDFilmStudio•8/30/24, 6:00 PM

since you trained on 512X512 dataset, will your outputs be affected if you go for higher resolution image generation like full HD?

Uuisato Double-checked, I'm using Torch 2.5. [Turned off the validating requirements par...

uisato•8/30/24, 6:01 PM

Also, does anybody have a Comfy workflow for FLUX + LORAs optimized for RTX 3090? I'm getting 15.79s/it, which strikes me as there's something wrong with my workflow/configuration.

uisato•8/30/24, 6:01 PM

Using this one.

MarkuryFLUX_LoRA_v2_-_FLUX__LORA.json17.59KB

DDreamer@TDFilmStudio since you trained on 512X512 dataset, will your outputs be affected if you go fo...

nicearoni•8/30/24, 6:02 PM

all of these outputs were 1024x1024 so I don't think so? I'll try some higher resolutions now

Nnicearoni all of these outputs were 1024x1024 so I don't think so? I'll try some higher re...

Dreamer@TDFilmStudio•8/30/24, 6:05 PM

Thanks for sharing your findings. I finished with my first character Flux Lora training today. I want to try out product lora next. Do you think it would be different from your style lora experiment? Why did you remove the word style from the target trained folder name? Does that impact the final quality?

DDreamer@TDFilmStudio Thanks for sharing your findings. I finished with my first character Flux Lora t...

nicearoni•8/30/24, 6:10 PM

I removed it from the Class prompt textbox in the kohya GUI as well after I clicked 'prepare training data' and 'copy info to respective fields' and renamed the folder (not sure that textbox is used outside of preparing\copying, just wanted to be sure). I removed it because of the convos around here about training on just 'ohwx' instead of 'ohwx man'. In comparing the 25 epoch from both trainings I did, I preferred the second one that only had the instance prompt and no captions.

Nnicearoni I removed it from the Class prompt textbox in the kohya GUI as well after I clic...

Dreamer@TDFilmStudio•8/30/24, 6:11 PM

Thats really interesting. Also reduces the prep steps. Captioning is tedious job

Uuisato Also, does anybody have a Comfy workflow for FLUX + LORAs optimized for RTX 3090...

nicearoni•8/30/24, 6:13 PM

yeah that def seems high... here's what I'm getting in SwarmUI (which of course uses Comfy as its backend) for 30 steps per 1024x1024 image in Swarm with a 2GB LoRA I trained myself...

DDreamer@TDFilmStudio Thats really interesting. Also reduces the prep steps. Captioning is tedious job

nicearoni•8/30/24, 6:15 PM

definitely agree on captioning being tedious! I decided to try without after hearing some people say it didn't help or even hurt character likeness, and that flux apparently has its own internal captioning somehow when training?

Dreamer@TDFilmStudio•8/30/24, 6:17 PM

Sharing my first batch of flux lora.

message.txt8.82KB

DDreamer@TDFilmStudio Sharing my first batch of flux lora.

nicearoni•8/30/24, 6:19 PM

dude these are

!! thanks for sharing - how many pics in your training dataset and how many training steps?

Nnicearoni dude these are 🔥 !! thanks for sharing - how many pics in your training dataset...

Dreamer@TDFilmStudio•8/30/24, 6:21 PM

Thanks. I had 16 pics. Rest of the configs I just imported Doc's 'rank 3' slow version

Dreamer@TDFilmStudio•8/30/24, 6:21 PM

Still let me confirm

Nnicearoni dude these are 🔥 !! thanks for sharing - how many pics in your training dataset...

Dreamer@TDFilmStudio•8/30/24, 6:27 PM

I think 16 pics gave me 3200 steps, took a little less than 4 hours with 1024X1024 max resolution for my base images on my 4090

Uuisato Using this one.

uisato•8/30/24, 7:41 PM

Solved it! Had to update NVIDIA drivers and turnoff shared memory. Turned on the --normalvram argument on Comfy. Getting 3.2 it/s

3090-is-really-slow-at-flux-standart-comfyui-workflow-v0-1wog5nzpskid1.webp

daroth•8/30/24, 7:44 PM

for generating with Flux, do you guys use Forge WebUI or ComfyUI?

Ddaroth for generating with Flux, do you guys use Forge WebUI or ComfyUI?

J2_Nedo•8/30/24, 7:45 PM

Forge and SwarmUI

J2_Nedo•8/30/24, 7:45 PM

SwarmUI is missing in your List

daroth•8/30/24, 7:47 PM

substitue it with other

daroth•8/30/24, 7:55 PM

�

𝔸𝖨ΞӾ•8/30/24, 8:10 PM

Check this out: https://www.reddit.com/r/StableDiffusion/comments/1f523bd/good_flux_loras_can_be_less_than_45mb_128_dim/

From the StableDiffusion community on Reddit: Good Flux LoRAs can b...

Explore this post and more from the StableDiffusion community

�

𝔸𝖨ΞӾ•8/30/24, 8:10 PM

Pretty interesting, only trained specific layers and the Lora filesize was only 4.5mb lol

�

𝔸𝖨ΞӾ•8/30/24, 8:13 PM

I wish I understood this more and what tools to use to train the specific layers they mentioned here:

This LoRA was trained on two layers single_transformer_blocks.7.proj_out and single_transformer_blocks.20.proj_out

It is possible to train on blocks.7.proj_outlayer only if you don't want the dataset style transfer, the size depending on the dim can be below 4.5mb (bf16), the model is powerful enough to offer advantages such as single layer training, so let's do that.

A Flux LoRA with 99% face likeness with great flexibility can be as small as 580kb (single layer, dim 16)

nicearoni•8/30/24, 8:13 PM

that's crazy

nicearoni•8/30/24, 8:15 PM

read the thread - no mention of how speed is affected only training on a couple layers

�

𝔸𝖨ΞӾ•8/30/24, 8:15 PM

I know someone like the Dr. could crack the code I bet lol

Михаил Кислов•8/30/24, 8:29 PM

Google translator.

Hi all! Successfully trained Lora for Flux. But when using it in Forge, it is ignored.

That is, with the same seed there will be the same result both when using Lora and without it. At the same time, in the console you can see that he is loading it.

Has anyone encountered the problem? How to solve it?

nicearoni•8/30/24, 8:38 PM

I need to reinstall because my Forge wasn't giving consistent results either. Then the console started saying something about patching the Loras? I could generate regular flux images okay in Forge but when I try to use my Lora, my 3090's dedicated memory filled and started using shared memory, so a single image was taking minutes to generate, and then it didn't look like the Lora had applied as expected - or as it does in Swarm where my images come out amazing usig Lora's I trained with Doc's configs. So I probably need a fresh reinstall of Forge. I'll test again and let you know.