so in your opinion what settings give the very best quality in low , medium , and highest quality
so in your opinion what settings give the very best quality in low , medium , and highest quality
![Dazzastrous [4090]](https://cdn.discordapp.com/avatars/871389071686135838/c4a5f4d81d9eeb18d856295d692e2c07.webp?size=16)
Fi ranked them already here :d
FPatreon
Get more from SECourses: Tutorials, Guides, Resources, Training, FLUX, MidJourney, Voice Clone, TTS, ChatGPT, GPT, LLM, Scripts by Furkan Gözü on Patreon
F
![Dazzastrous [4090]](https://cdn.discordapp.com/avatars/871389071686135838/c4a5f4d81d9eeb18d856295d692e2c07.webp?size=40)
Dahh
Dbetter
Dso rank for 4090
D3?
Dis the magic number
Fye 3 best
Ftesting different batch size speeds for flux
Fto add new configs
Fthe batch size speed gain with flux is minimal. almost no gain. testing on RTX A6000
NI've wondered about the avr_loss and tensorboard graphs available during training... anything to look for there that indicates how the training is progressing?
FPatreon
Get more from SECourses: Tutorials, Guides, Resources, Training, FLUX, MidJourney, Voice Clone, TTS, ChatGPT, GPT, LLM, Scripts by Furkan Gözü on Patreon
Fyou can literally get 1.4375 second / it with 4x gpu with minimal loss of quality
Fso 3000 steps takes around 72 minutes 
D
Fhahahaha xcellent how do you do that?
Fhahahaha xcellent how do you do that?
Fvery likely trained my face 
Dnah just prompted in ideogram 2
in ideogram 2Dits the best for text i find
�Midjourney is in shambles!
F
Hseems the runpod .sh dosent work on regular linux ubuntu
Hi tried so many times and it fails because it dosent recordnize what i want but its okay i may just rent a runpod server
Fyou need to have python c++ tools and cuda installed
Frunpod comes with pytorch docker template
Fif you can load it it would work
Hi have cuda but not sure about c++
Fbut it is hard to install those
Falso runpod doesnt have sudo command
Fyou need to append them
Hill have to look into it tomorrrow
DOn the new flux lora test updates, i see speeds in s/it for each config json. Are those the speeds we shoukd be getting? Im running config 5 which is the one recommended for 12gb vram cards like my 3060. It was 12.2 s/it but im getting 25.5 s/it. Not sure what im doing wrong if those are the expected rates
Fno those are for A6000 GPU
Frelative speed
Falso on linux
F25 second is looking normal
Fyou have to reduce resolution for speed up but reduces quality
Fyou can do 768x768
DThanks. Lol i thought i was way off. 25 seems fine
M@Dr. Furkan Gözükara doesn't seem like comfy or forge accept loras trained with text encoder yet
Ftext encoder training didnt added to kohya yet
Fso i dont know :d
Mi got it to work in sample images its fine, but they don't work on comfy or forge yet
in sample images its fine, but they don't work on comfy or forge yetFwell i am waiting kohya
Fhe said he found today gonna add :d
DFor multi gpu training do the gpus need to be equal in vram? Like i have an 8 gb card im not using and a 12 gb card that is installed. Would the 8 gb card work for multi gpu training with the 12 gb one?
