yes I am doing 512x512
yes I am doing 512x512
DTry 25 images
DOr 20
Hyep I'm trying 10, just to confirm reducing images helps the problem
DWhen they give you caching time
DCatching must be finish with in 30 minutes
DI have did 1024,1024 with 15 images too
D13 is best for 1024,1024
DMaybe with 512 more imgaes will also do the job for around 20
Hwill kaggle premium give more caching time? or is it a bug which will probably appear there too?
DI have never used premium
Athe cache time issue will be resolved if you just start running again right after the 30 min
VFurkan
AThat is what @Dr. Furkan Gözükara told me and it worked for me:
Furkan Gözükara SECourses — 11/30/2023 12:16 PM
ok easy to fix
click start training again
it will cache remaining files
Furkan Gözükara SECourses — 11/30/2023 12:16 PM
ok easy to fix
click start training again
it will cache remaining files
Vseeming huggingface changed, i tried to upload
Vbut yousee all models come to this place
Fload it from hugging face
Vit seems uploading to branch, i do not know how to download it from branch
Fi think it is about meta data. you message me about this from patreon?
Vyes i guess so, let me send it to you from patreon
Fsadly i am not that expert of comfyui
Vno does not matter with comfyui
Vonly matters with uploading to huggingface
Fyes this happens when caching takes too long
Freduce number repeat
For start process again
Fit will continue caching
Fit is working just wait
Vi need it to be seen on this page
Vnow all here
Vin the past it is okay, now changed
Vyou see, checkpoint is always about 7 g, without uploading to huggingface, the training is useless
L@Dr. Furkan Gözükara have you had any good results with magic animate yet? all I've tried was quite disappointing haha
LI've tried very slow movement too and roughly same stature of characters etc
Lit's nowhere near as good as the repo implies it is
Flooks like another fake research paper
Lmaybe animate anything will be slightly better, because it doesnt use this weird area approach and uses actual posenet
Muse the special tokens you used in training. For example main prompt could be: ohwx man standing next to bbuk woman on a street corner. In the adetailer prompt (using face yolov8n) use: ohwx man <ohwx man lora> [SEP] bbuk woman <bbuk woman lora>. If it swaps faces then reverse this command. Make sure to set detection K of 2 so it does not swap other faces if there are more than 2. If you had a ckpt trained on both then you wouldn't have to use the <lora> portion. Either way works
AIsn't using this loading it from huggingface?
stablediffusionapi/juggernaut-xl-v7
stablediffusionapi/juggernaut-xl-v7
Dok will try. its safetensor , not ckpt 
.
I still have 1024 pictures of both of us. is it possioble to put both of our photos and train ? If so what would be the regularization pictures ? it should be about couples right ?
.I still have 1024 pictures of both of us. is it possioble to put both of our photos and train ? If so what would be the regularization pictures ? it should be about couples right ?
D@Dr. Furkan Gözükara some guidence would be great.
Msorry I meant the same thing: safetensor is a version of a ckpt. i use safetensor on everything
Fit should work
Fi can try later
Mtrain using the methods described and in video by @Dr. Furkan Gözükara but add another folder for the other person. for one person the training folder looks like this C:\koyha\training\img\40_ohwx man Simply add a second folder with 40_bbuk woman and put 10-15 pics at 1024x1024 of each person in the folders. The training will create a safetensor chkpt with both ohwx and bbuk people in it.
Aok i'll try again, thanks
Fyes it will train both
Faccurate
A@Dr. Furkan Gözükara how would I get it to not use VAE in addition to Juggernaut for training? Juggernaut already has vae baked in