Software Engineering Courses (SECourses)•3y ago

hello! 🙂 its been days i can't use my kohya anymore. once i start training it freeze and dont proce

hello!

its been days i can't use my kohya anymore.
once i start training it freeze and dont proceed.
this is the only error i got in the list

▲

▲OP•6/1/23, 7:42 PM

▲▲hello! 🙂 its been days i can't use my kohya anymore. once i start training it f...

Furkan Gözükara SECourses•6/1/23, 7:58 PM

hello. did you make a new installation ?

FFurkan Gözükara SECourses hello. did you make a new installation ?

▲

▲OP•6/1/23, 8:01 PM

3 times

no luck

▲

▲OP•6/1/23, 8:01 PM

i always chose torch 2 cause i use it for auto1111 (not sure if is shared or whatever)
should i redo install choosing torch 1?

▲▲i always chose torch 2 cause i use it for auto1111 (not sure if is shared or wha...

Furkan Gözükara SECourses•6/1/23, 8:03 PM

dont do that

▲

▲OP•6/1/23, 8:03 PM

Furkan Gözükara SECourses•6/1/23, 8:03 PM

kohya makes its own venv

Furkan Gözükara SECourses•6/1/23, 8:03 PM

follow as i did

Furkan Gözükara SECourses•6/1/23, 8:03 PM

it wont affect your main install

▲

▲OP•6/1/23, 8:03 PM

ok so i reinstall and choose torch 1

Aashley Torch version 2 actually works with Dreambooth on Linux

Furkan Gözükara SECourses•6/1/23, 8:04 PM

but on runpod dont i wonder the difference

RRobboRobbo @Dr. Furkan Gözükara Hi, I have followed your tutorial "Easiest Way to Install &...

Furkan Gözükara SECourses•6/1/23, 8:05 PM

i suggest you to do manual install. explained in the end of this video . it is so easy. also watching this video help you tremendously https://youtu.be/B5U7LJOvH6g

YouTubeSECourses

How To Install Python, Setup Virtual Environment VENV, Set Default ...

How to install Python, have multiple Python installations, set system wide default Python version. How to compose venv for any Python installation, change Python default path, and install SD web UI properly. Discord: https://bit.ly/SECoursesDiscord.

If I have been of assistance to you and you would like to show your support for my work, please...

▲

▲OP•6/1/23, 8:05 PM

thnaks

RRikk @Dr. Furkan Gözükara the new Adobe Firefly is truly remarkable. These are some ...

Furkan Gözükara SECourses•6/1/23, 8:05 PM

yes it is amazing and controlnet is getting this feature . they already made inpainting. working on outpainting

Ffelixsanz @Dr. Furkan Gözükara do you really think the model has chinesse spyware?

Furkan Gözükara SECourses•6/1/23, 8:06 PM

i dont think has spyware. otherwise the author wouldnt remove it :d

Aashley New RunPod template available for Stable Diffusion Web UI: https://runpod.io/gs...

Furkan Gözükara SECourses•6/1/23, 8:06 PM

nice ty

Original message was deleted

Furkan Gözükara SECourses•6/1/23, 8:07 PM

not yet what does it do?

Fflowwolf here's a terran style i trained with it last night. not the best work. but it s...

Furkan Gözükara SECourses•6/1/23, 8:07 PM

very nice pictures. which model is this?

FFurkan Gözükara SECourses very nice pictures. which model is this?

flowwolf•6/1/23, 8:08 PM

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

flowwolf•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

flowwolf•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

flowwolf•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

flowwolf•6/1/23, 8:18 PM

flowwolf•6/1/23, 8:18 PM

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

flowwolf•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?

flowwolf•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

flowwolf•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

flowwolf•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?

flowwolf•6/1/23, 8:27 PM

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

flowwolf•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

flowwolf•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

flowwolf•6/1/23, 9:22 PM

terminal > make sure no venv is active >