Software Engineering Courses (SECourses)•3y ago

i made those images with dreamshaper 6 using the terran lora i trained. haven't had great success o

i made those images with dreamshaper 6 using the terran lora i trained. haven't had great success on the first 3 versions of it, but dadaptation 3 helped this one quite a lot

flowwolfOP•6/1/23, 8:09 PM

trained the lora using the base 1.5

Fflowwolf i made those images with dreamshaper 6 using the terran lora i trained. haven't...

Furkan Gözükara SECourses•6/1/23, 8:12 PM

i see thanks

Furkan Gözükara SECourses•6/1/23, 8:12 PM

what is dadaptation 3

flowwolfOP•6/1/23, 8:13 PM

they added support in the latest kohya update. optimizer from facebook research

Fflowwolf they added support in the latest kohya update. optimizer from facebook research

Furkan Gözükara SECourses•6/1/23, 8:14 PM

can you show screenshot

flowwolfOP•6/1/23, 8:15 PM

https://github.com/facebookresearch/dadaptation https://github.com/kohya-ss/sd-scripts#25-may-2023-20230525

what do you want the screenshot of?

GitHub

GitHub - facebookresearch/dadaptation: D-Adaptation for SGD, Adam a...

D-Adaptation for SGD, Adam and AdaGrad. Contribute to facebookresearch/dadaptation development by creating an account on GitHub.

GitHub

GitHub - kohya-ss/sd-scripts

Contribute to kohya-ss/sd-scripts development by creating an account on GitHub.

Fflowwolf https://github.com/facebookresearch/dadaptation https://github.com/kohya-ss/sd-...

Furkan Gözükara SECourses•6/1/23, 8:18 PM

with d adaptation you dont chose learning rate right?

Furkan Gözükara SECourses•6/1/23, 8:18 PM

i had heard this

Furkan Gözükara SECourses•6/1/23, 8:18 PM

definitely should test

flowwolfOP•6/1/23, 8:18 PM

flowwolfOP•6/1/23, 8:18 PM

yeah you set the learn rate to 1 and it adjusts it all on the fly

flowwolfOP•6/1/23, 8:18 PM

if i understand it right, the lr acts more of a weight towards the on the fly estimates

Fflowwolf Click to see attachment

Furkan Gözükara SECourses•6/1/23, 8:18 PM

that lion is the best one?

flowwolfOP•6/1/23, 8:19 PM

lion is what i've been using but it's all i've tried. using an automatic algorithm to tame it seems suiting

Furkan Gözükara SECourses•6/1/23, 8:19 PM

actually i am plannin to compare all learning optimizers haha

flowwolfOP•6/1/23, 8:21 PM

That's good! i've stuck only to adam8bit so far until jumping to lion. i've not enough exepreince to say what's better than others yet. Don't even know how to effectively test these models other than throwing spaghetti at the wall

flowwolfOP•6/1/23, 8:24 PM

i've only recently started using the kohya setup. i think itsa much better workflow than what the webui-extension provides

Furkan Gözükara SECourses•6/1/23, 8:26 PM

now lates roop is supported . updated github file : https://github.com/FurkanGozukara/Stable-Diffusion/blob/main/Tutorials/1-Click-DeepFake-Tutorial.md

GitHub

Stable-Diffusion/1-Click-DeepFake-Tutorial.md at main · FurkanGozuk...

Best Stable Diffusion and AI Tutorials, Guides, News, Tips and Tricks - Stable-Diffusion/1-Click-DeepFake-Tutorial.md at main · FurkanGozukara/Stable-Diffusion

Fflowwolf That's good! i've stuck only to adam8bit so far until jumping to lion. i've not...

Furkan Gözükara SECourses•6/1/23, 8:26 PM

what network rank you are using? 128?

flowwolfOP•6/1/23, 8:27 PM

that terran lora was rank 8 alpha 1. wanted to see how low i could take it

flowwolfOP•6/1/23, 8:29 PM

accelerate launch --num_cpu_threads_per_process=2 "train_network.py" --enable_bucket --pretrained_model_name_or_path="runwayml/stable-diffusion-v1-5" --train_data_dir="J:\ImageSets\terran\set3" --resolution=640,640 --output_dir="J:\loras\terrantechv4" --logging_dir="J:\loras\terrantechv4\logs" --network_alpha="1" --save_model_as=safetensors --network_module=networks.lora --text_encoder_lr=1.0 --unet_lr=1.0 --network_dim=8 --output_name="terrantechv4" --lr_scheduler_num_cycles="3" --learning_rate="1.0" --lr_scheduler="cosine_with_restarts" --lr_warmup_steps="280" --train_batch_size="4" --max_train_steps="2800" --save_every_n_epochs="5" --mixed_precision="bf16" --save_precision="bf16" --caption_extension=".txt" --optimizer_type="DAdaptLion" --max_data_loader_n_workers="0" --bucket_reso_steps=64 --mem_eff_attn --color_aug --flip_aug --shuffle_caption --gradient_checkpointing --xformers --bucket_no_upscale --random_crop --multires_noise_iterations="7" --multires_noise_discount="0.1" --sample_sampler=k_dpm_2_a --sample_prompts="J:\loras\terrantechv4\sample\prompt.txt" --sample_every_n_epochs="1"

Ffelixsanz there is no 2.10.1

Rclemente27•6/1/23, 9:05 PM

Thanks! It worked out. But now I don't now how to solve this one:
ERROR: triton-2.0.0-cp310-cp310-win_amd64.whl is not a supported wheel on this platform.
I tried looking for a differnt version but the one I've tried apparently doesn't satisdies the riquirement triton

▲▲Click to see attachment

flowwolfOP•6/1/23, 9:21 PM

okay so i had simeilar problems. i think it's cause preexisting stuff exists on the system outside of venvs. like when i was installing things for other ui's but forgot to activate a venv

flowwolfOP•6/1/23, 9:22 PM

terminal > make sure no venv is active >

cd \kohya_ss\  [whereever it is]

cd \kohya_ss\  [whereever it is]

pip freeze > uninstall.txt

pip freeze > uninstall.txt

pip uninstall -r uninstall.txt

pip uninstall -r uninstall.txt

flowwolfOP•6/1/23, 9:23 PM

then i deleted the whole kohya folder and started over. worked

flowwolfOP•6/1/23, 9:24 PM

i dont know if that's what i did to fix it, but it worked when i reinstalled and that's the only step i tried before

flowwolfOP•6/1/23, 9:25 PM

got the info here https://github.com/bmaltais/kohya_ss#filenotfounderror

Furkan Gözükara SECourses•6/1/23, 9:53 PM

looks like latest web ui commit breaking dreambooth

Furkan Gözükara SECourses•6/1/23, 9:53 PM

testing your template @ashleyk right now

RRclemente27 Thanks! It worked out. But now I don't now how to solve this one: *ERROR: triton...

Furkan Gözükara SECourses•6/1/23, 9:53 PM

at which step did you get this error

FFurkan Gözükara SECourses at which step did you get this error

Rclemente27•6/1/23, 9:54 PM

At the beggining. I got the 3 error at the same time

Rclemente27•6/1/23, 10:03 PM

When I run .\setup.bat

RRclemente27 At the beggining. I got the 3 error at the same time

Furkan Gözükara SECourses•6/1/23, 10:10 PM

which tutorial you were following can i get link again

FFurkan Gözükara SECourses which tutorial you were following can i get link again

Rclemente27•6/1/23, 10:11 PM

https://www.youtube.com/watch?v=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

https://bit.ly/SECoursesDiscord

If I have been of assis...

Furkan Gözükara SECourses•6/1/23, 10:13 PM

https://github.com/FurkanGozukara/Stable-Diffusion/blob/main/Tutorials/Generate-Studio-Quality-Realistic-Photos-By-Kohya-LoRA-Stable-Diffusion-Training-Full-Tutorial.md

GitHub

Stable-Diffusion/Generate-Studio-Quality-Realistic-Photos-By-Kohya-...

Best Stable Diffusion and AI Tutorials, Guides, News, Tips and Tricks - Stable-Diffusion/Generate-Studio-Quality-Realistic-Photos-By-Kohya-LoRA-Stable-Diffusion-Training-Full-Tutorial.md at main · ...

Furkan Gözükara SECourses•6/1/23, 10:13 PM

ok let me test

RRclemente27 https://www.youtube.com/watch?v=TpuDOsuKIBo

Furkan Gözükara SECourses•6/1/23, 10:14 PM

you select like this right?

Furkan Gözükara SECourses•6/1/23, 10:14 PM

i am doing a fresh install

Furkan Gözükara SECourses•6/1/23, 10:14 PM

also what is your python version

FFurkan Gözükara SECourses also what is your python version

Rclemente27•6/1/23, 10:15 PM

22.3.1

FFurkan Gözükara SECourses you select like this right?

Rclemente27•6/1/23, 10:17 PM

I have done it with the 4 options but in all of them I get similar or the same errors

flowwolfOP•6/1/23, 10:18 PM

you'll want python 3.10.x

flowwolfOP•6/1/23, 10:18 PM

22.3.1 looks like an amd driver version

Rclemente27•6/1/23, 10:19 PM

my bad it's Python 3.11.3

flowwolfOP•6/1/23, 10:19 PM

you'll want 3.10

flowwolfOP•6/1/23, 10:19 PM

if 22.3.1 is an amd driver, that's an old one, like from when i used amd drivers

felixsanz•6/1/23, 10:22 PM

@Dr. Furkan Gözükara i don't understand your video! how is controlnet inpainting different to normal inpainting?

code•6/1/23, 10:25 PM

Hey lads, trying to install Dreambooth, manually installed, and removed my virtual env, but still getting errors on start up, looked through the chat and found no answers, hoping someone would know what to do:

ERROR: Could not install packages due to an OSError: [WinError 5] Access is denied:

ERROR: Could not install packages due to an OSError: [WinError 5] Access is denied:

first error is the: https://github.com/d8ahazard/sd_dreambooth_extension/issues/813

GitHub

Could not install packages due to an OSError: [WinError 5] Access i...

Kindly read the entire form below and fill it out with the requested information. Please find the following lines in the console and paste them below. If you do not provide this information, your i...

Furkan Gözükara SECourses•6/1/23, 10:26 PM

ok dev branch is now working on runpod with torch 2

Furkan Gözükara SECourses•6/1/23, 10:26 PM

great

i made those images with dreamshaper 6 using the terran lora i trained. haven't had great success o

Similar Threads