Software Engineering Courses (SECourses)•2y ago

I have tried the latest pack for amssedcompute update auto1111, controlnet, reactos, facefusion... B

I have tried the latest pack for amssedcompute update auto1111, controlnet, reactos, facefusion... But when starting auto1111 all of them work except facefusion which gives me this error [FACEFUSION.CORE] FFMpeg is not installed

TTheToday I have tried the latest pack for amssedcompute update auto1111, controlnet, reac...

Furkan Gözükara SECourses•5/20/24, 4:46 PM

easy to solve

Furkan Gözükara SECourses•5/20/24, 4:46 PM

i will update script

TTheToday I have tried the latest pack for amssedcompute update auto1111, controlnet, reac...

Furkan Gözükara SECourses•5/20/24, 4:50 PM

open a new terminal and execute below command and restart

Furkan Gözükara SECourses•5/20/24, 4:50 PM

web ui

Furkan Gözükara SECourses•5/20/24, 4:50 PM

Furkan Gözükara SECourses•5/20/24, 4:52 PM

scripts also updated

TheTodayOP•5/20/24, 4:53 PM

works! thanks

rafraf•5/20/24, 4:55 PM

Hey @Dr. Furkan Gözükara how’s it going with your text encoder sdxl training tests?

rafraf•5/20/24, 4:57 PM

Btw, kohya finally merged passing args to wand PR. It’s now available on the dev branch. It’s working now on kohya_ss by passing command in cli with “—log_config”

Rrafraf Hey @Dr. Furkan Gözükara how’s it going with your text encoder sdxl training tes...

Furkan Gözükara SECourses•5/20/24, 4:57 PM

running tests right now

Furkan Gözükara SECourses•5/20/24, 4:57 PM

rafraf•5/20/24, 4:58 PM

Great looking forward to it. I’m testing style model but want to go banck to testing subject models with Adafactor vs adamW8bit vs prodigy

Original message was deleted

rafraf•5/20/24, 5:26 PM

nice but?

virginia•5/20/24, 5:45 PM

@Dr. Furkan Gözükara Hey Furkan which is the best one to upscale an image

hmm•5/20/24, 6:11 PM

FFurkan Gözükara SECourses ip adapter

AIGambino•5/20/24, 9:15 PM

How can I fix it? It seemed to work last week, no idea what happened since then?

AIGambino•5/20/24, 9:40 PM

I tried this way through the unified loader faceid but still no luck

AIGambino•5/20/24, 10:12 PM

Furkan Gözükara SECourses•5/20/24, 10:16 PM

testing x/y/z

dxqb•5/20/24, 10:34 PM

next time you are running a OneTrainer comparison, @Dr. Furkan Gözükara , would you mind adding 1 run with a change that I have made to OneTrainer? it's just a couple of lines of code, I'd send you a patch file.
it allows you to set TE1 and TE2 (!) to full learning rate, same as unet, without overcooking it.

I think it improves quality. but it's subtle and I'd like a second opinion

dxqb•5/20/24, 10:38 PM

or kohya, I have it for kohya, too. just didn't use it now because of the learning rate issue.

Ddxqb next time you are running a OneTrainer comparison, @Dr. Furkan Gözükara , would ...

Furkan Gözükara SECourses•5/20/24, 10:43 PM

what do you mean by that?

Furkan Gözükara SECourses•5/20/24, 10:43 PM

can you tell me more details

dxqb•5/20/24, 10:45 PM

sure. in my opinion, the TEs are overcooking because we let them just run in one direction, uncontrolled. When you train the unet, there is loss function - which is how good it is denoising an image. when you enable TE training, it simply also optimizes the TE parameters, but the loss function is still only how good the denoising of unet is. so you allow the TE to optimize towards your training data set and it looses all flexibility.

dxqb•5/20/24, 10:45 PM

so I have added a loss function that also controls the TE.

dxqb•5/20/24, 10:46 PM

I hope that was kinda clear? difficult to explain in two sentences

Ddxqb I hope that was kinda clear? difficult to explain in two sentences

Furkan Gözükara SECourses•5/20/24, 10:47 PM

did you make a PR for this?

Furkan Gözükara SECourses•5/20/24, 10:47 PM

only for one trainer?

dxqb•5/20/24, 10:48 PM

no, I've talked to the OT developer and he is not interested. he thinks it is too specific for my/our use case. he also thinks the same can be done with additional embeddings - but I've tried that. it's worse.

AIGambino•5/20/24, 10:48 PM

The workflow works on this clip but other clips just dont load at all and give that error

dxqb•5/20/24, 10:48 PM

so whenever you have the chance, it would be great to get your opinion.

Ddxqb no, I've talked to the OT developer and he is not interested. he thinks it is to...

Furkan Gözükara SECourses•5/20/24, 10:50 PM

sure i would like to test

Furkan Gözükara SECourses•5/20/24, 10:50 PM

currently i was using kohya

Furkan Gözükara SECourses•5/20/24, 10:50 PM

do you have same for kohya?

Furkan Gözükara SECourses•5/20/24, 10:50 PM

that would be a nice comparison but i can test on OT too

dxqb•5/20/24, 10:51 PM

I had it for kohya, as I was using kohya until we noticed the quality isn't good anymore. so I think testing this for kohya only makes sense after you have found the perfect hyperparameters again. otherwise we'd compare against something that is bad for other reasons

Ddxqb I *had* it for kohya, as I was using kohya until we noticed the quality isn't go...

Furkan Gözükara SECourses•5/20/24, 10:54 PM

ye i found them

Furkan Gözükara SECourses•5/20/24, 10:54 PM

going to update now

dxqb•5/20/24, 11:04 PM

I DMed you a patch. The class it controls against is hardcoded to 'man', so if you are training a woman or something else, you need to change that

dxqb•5/20/24, 11:05 PM

parameters: leave everything as-is, but turn on TE2, too, and set both TE1 and TE2 learning rates to the main learning rate 1e-05

dxqb•5/20/24, 11:05 PM

if you want it for kohya, I'd have to first test it again tomorrow, as it was a while ago and I don't know if I have left it in a working state

dxqb•5/20/24, 11:07 PM