rtx 4080 it should work
rtx 4080 it should work
Fyes i used the one linked on SECourses and a few others here and there. kohya is very widely supported . most tutorials i've noticed are for it, rather than the webui
Fyes
Fi also managed to get pytorch 2 working so i trained a style using lion. got some nice results but i might've set the lr way to low
Fi use 1e-7 for lion
Ftorch 2 works on windows
Fbut on unix i couldnt yet
Ysimple question : i've trained a dreambooth that work fine as long i don't put " oil painting in the prompt : in this case the face is destroyed
EI figured epoch is more important, but then how would I decid how many repeats? :/ ugh I feel that dream booth extension is simpler, but it always gives us trouble.
Fthat means either overtrained or undertrained
Fye it depends on used script
Fnormally 1 epoch means 1 time all training items are trained
Yi have several steps ckpt of this model i will try them all
Fdreambooth learning rate 0.00001 vs 0.000002 anyone?
Fi prefer 1e-6 for adam, 1e-7 for lion
Yi use 0.000002
Fbut you tried 0.00001 ?
Ynot so far
Mif we’re trying to train concepts (again I’m doing fantasy races), does it make sense to include a lot more training images?
Fquality over quantity i lean towards. add more but only if they meet a bar
Fthe concept in a myriad of ways too . not more images that are similar to whats there already. unless they're high quality haha
Edo i need to upgrade xformers for 1111 dreambooth extension to work properly, or are these settings alright?
Fi would update to latest
EWill do! You have a video on it? I’ll watch it. On how to update it
Pwouldn't you just need to force pip to upgrade the xformers package
Fi got bf16 working. i'm not sure why training would just hang up the other day. came back to it with fresh eyes a day later and it just works this time. cool. style training is going well. i'm doing lion with cosine 3 restarts. te lr of 1e-6 and unet lr 2e-5. i guess i'm supposed to use less but the cosine schedulling keeps dipping it towards 1e-6
FYouTubeSECourses
Discord : https://bit.ly/SECoursesDiscord. Torch 2 / Pytorch 2 is now supported along with new DreamBooth Automatic1111 Web UI extension. If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on 
https://www.patreon.com/SECourses
In this video, I am explaining how you can man...
https://www.patreon.com/SECoursesIn this video, I am explaining how you can man...
Fi am trying to find best settings for speech training 
Fi have a famous voice in my mind to train
F496.wav1.88MB
Fi am going to sleep hopefully see you later
Tanyone had massive it/s slowdown on the new 1.3.0 a1111 commit? reverting commits didn't give me back my old speed either and i'm trying to figure out what might've caused this
AI also couldn't get the latest version of torch working on Linux
AIt shouldn't make any difference to things like that. It's usually your GPU that affects things like that.
KHi, need your guys help i tried using stable diffusion on runpod and have downloaded the control net extension but somehow its not working. i use the control net extension and have downloaded the relevant files
@Furkan Gözükara SECourses
@Furkan Gözükara SECourses
Twell there's quite a few software versions and settings that can massively affect speed in a1111. i didn't change any of that besides doing a git pull though, and i didn't do any operating system updates on my linux system which is why i'm confused
Tdowngrading to my old webui commit didn't help either ; /
生when i run the code in this vedio .cmd show this to me. is that means i dont run on the GPU?
生YouTubeSECourses
Databricks’ #Dolly v2 is a free, open source, commercially useable ChatGPT-style #AI model. Dolly 2.0 could spark a new wave of fully open source LLMs similar to #ChatGPT . Open source community working hardest to bring up a model that can compete with GPT4. Our discord: https://bit.ly/SECoursesDiscord
If I have been of assistance to you and yo...
If I have been of assistance to you and yo...
生
生i look my Task Manager my ram is increasing so high but gpu not increase
Fdo you need regularization images for training a photograph style lora from movie references?
or is there a good tutorial specifically for this?
its unclear when you need to tag the images, (or remove stuff from tags) when you need regularization images, when it comes to training a photographic style vs a character vs a pose/action vs details applied to any of the above, is there good info about this anywhere? (info seems to change every 2 weeks lol)
or a loha/lycoris if they are better
never tried to train one
or is there a good tutorial specifically for this?
its unclear when you need to tag the images, (or remove stuff from tags) when you need regularization images, when it comes to training a photographic style vs a character vs a pose/action vs details applied to any of the above, is there good info about this anywhere? (info seems to change every 2 weeks lol)
or a loha/lycoris if they are better
never tried to train one
Flike a table that says:
training photoreal faces? dont need regularisation images, here's good starter settings
training a photographic style, dont do this, these settings are good, etc, would be amazing
training photoreal faces? dont need regularisation images, here's good starter settings
training a photographic style, dont do this, these settings are good, etc, would be amazing
Fthat could be kept updated
Θare these good settings for dreambooth
Θalso for training people is it best to use LoRa or Dreambooth training?
Θso for the best resultst possible
Θsize doesnt matter for me
ADr. Furkan Gözükara told me that Dreambooth is the best for training people. I haven't actually tried LoRA yet, but Textual Inversion produced terrible results. Dreambooth produces great results.
ΘOk so good choice i chose dreambooth
ΘThank you very much for your repsonse