whew runpod is $$$
whew runpod is $$$
Fye it is paid
T25 dollars in one day and trained maybe 15 models
Tnot worth it imo
Tbut thanks for the tutorial thats just info that should be included
BCan you do a tutorial on deforum
Fi plan to
Fcurrently working on voice clone and a new animation technique
Fvideo to animation
UHow easy is it to use? Ive never used any Collab virtual VMware type stuff. Had been hoping maybe something exists where it's plug in play like an actual video card so I wouldn't have to upload my client or smthg. I built a new comp but the riser cable is bad. So gotta wait. Sucks not being able to use SD
UAnd I just mean easy to set up and go. The literal idea of a plug and play virtual video card I mean be neat for it to just show up like a printer on a network or smthg
Frunpod is as easy as use on your computer once you watch my tutorials :d
UHaha it's a brand new interest. Hadn't even gone anywhere with it yet, but yes sir, that will be the first place I look. Trying to get my old comp running again but I'm afraid I broke the installation
Fi am recording deep voice cloning tutorial
CDoes anyone know how the controlnet for tile actually works? I know you can use it with the ultimate sd upscale script and i understand that it modifies the prompt based on what is in each tile. But what does it do in cases where you arent using ultimatr sd upscale? Ive seen demos where people set denoise to 1 and used it to regenerate an image that maintained composition but i cant find anyone who bothers to explain why that actually works. How does it split the image into tiles? What does the tile resamplr preprocessor actually do? Is there a way for me to specify the tiles or slice the image myself?
Factually i was also gonna look for this
Fhavent found yet
SOooOOooO. I might see if I can get this working on runpod.
Awhen i use dreambooth to train on runpod i get this
Acan someone help?
Awhen i use the same settings locally everything works fine but when i use runpod it no work
SProbably a python library issue. Try this https://discordapp.com/channels/772774097734074388/1042802773365891173/1104460615252180992
Abruh thats t he thing i use
SWhat source checkpoint are you using?
Asays i ran out of gpu memory some how
Av2.1 ema
A768
SSee if it works with v1.5 as a test. I've seen that error with other source checkpoints as well. I haven't figured out why.
Aaight ill test
SOh wait, are you doing Lora?
Ano
SHmm. I guess try v1.5 and see what happens.
Ayuh
Aoh its working now
Ai think i had the wrong path put in
Aor not
Alemme try 2.1 again
Ahuh ok the 2.1 model works as well
Athe only other things i changed were the capitalization of my folders thats my images were stored in and the path of it
Aah for some reason i can only do 2.1 if i first start doing the 1.5 then canceling it then doing the 2.1
Athats weird
B@Furkan Gözükara SECourses my new model works even worse
B
TI trainend on pictures of my son who wants to have a GTA5 Style Poster 
This really close. What would some of you suggest how to go further?
How to fix the eyes? I dont know how to use Inpaint properly...
And are threre options to refine this image...?
This really close. What would some of you suggest how to go further?
How to fix the eyes? I dont know how to use Inpaint properly...
And are threre options to refine this image...?
ANothing worked
Fi am using gui trainer. that gui trainer actually starting a cmd command
Fso only that part would be different
Fwhat are your settings xformers version and such
Fi can help you privately too if you are our patreon supporter. i can connect your pc and help to setup
Fwhat difference your new model had?
