going to try taking more photos if i get done working early today
going to try taking more photos if i get done working early today
MSame seed, and epoch 7 and 10
Mgenerated on RealisticVision 5.1, DPM++ SDE Karras steps 30, Hires 2, steps 15, 0,4
MThere is minimal difference between epoch 7 and 10, so epoch 7 is better if I am looking for flexibility.
MHowever, there is a striking difference between SD and RealisticVision, and between 128/128 and 128/64.
MIf you are short of photos, I read an interesting trick on reddit (updated!):
- generate images using the IP-Adapter in ControlNet, the ip-adapter-plus-face model (Starting Control Step: 0,5), and the ReActor (Roop based face changer) together on RealisticVision model.
- Put the better photos next to your few photos.
- make a good lora with lots of photos.
Snice, i won't be hard tho, i take photos for a living just not of myself
Si'll just put the camera on a tripod and go at it.
S
Saccuracy is aslo probably off because the photos were i used were shot over the course of 6 years ahaha
S
Si think results will be better if my dataset is a closer match to each other
MIf you're training with male class images, you might not mind a bigger difference, but yes, it's worth using similar images for age.
Si train with 0 class images
Ldid you train on kohya?
Sdreambooth
Lcan i know your setting pls
LIm trying to train my model on dreambooth
L@Sandstormleader is reactor an extension?
S
MYes, ReActor is a roop-based extension for auto1111
Snever used it
MGitHub
Fast and Simple Face Swap Extension for StableDiffusion WebUI (A1111, SD.Next, Cagliostro) - GitHub - Gourieff/sd-webui-reactor: Fast and Simple Face Swap Extension for StableDiffusion WebUI (A1111...
KHi @here
Lhow much training steps did you use?
S150
Lhow much images is there in your dataset?
Sthis one above about 26
Sbut a lot 10 were probably the same image cropped closer to face
Li will try with 20 images first
Sso i think i'm just going to set up a white backdrop for myself this afternoon
Si also do not downscale my images
Sso that first image was 2581x3872
S@Dr. Furkan Gözükara have you tested results of data photos having a beard vs no beard?
Swondering if i need to shave lol
RHey @Dr. Furkan Gözükara got amazing results for 12 image character LoRA using AdamW8bit 8000 steps--testing different batch sizes quality vs speed of training. Did you test LoRA Adafactor settings or only Full Finetune? I tried LoRA Adafactor with your settings with LR 0.0004 and it learned very little, color discoloration etc.. Can you point in right direction for LoRA--Adafactor?
RI see you are checking Adafactor SDXL TE problem on Issue #855. Does this mean Adafactor is broken at the moment, in light of TE? I thought it was beneficial to exclude SDXL Text Encoder from learning? Is Dadaptation a better option?
Rcan someone explain why google colab gives me this error:
ImportError Traceback (most recent call last)
<ipython-input-7-fdf29cbd2fa1> in <cell line: 22>()
20 import io
21 import math
---> 22 import timm
23 from IPython import display
24 import lpips
11 frames
/usr/local/lib/python3.10/dist-packages/PIL/ImageFont.py in <module>
32
33 from . import Image
---> 34 from ._util import isDirectory, isPath
35
36 LAYOUT_BASIC = 0
ImportError: cannot import name 'isDirectory' from 'PIL._util' (/usr/local/lib/python3.10/dist-packages/PIL/_util.py)
---------------------------------------------------------------------------
NOTE: If your import is failing due to a missing package, you can
manually install dependencies using either !pip or !apt.
To view examples of installing some common dependencies, click the
"Open Examples" button below.
ImportError Traceback (most recent call last)
<ipython-input-7-fdf29cbd2fa1> in <cell line: 22>()
20 import io
21 import math
---> 22 import timm
23 from IPython import display
24 import lpips
11 frames
/usr/local/lib/python3.10/dist-packages/PIL/ImageFont.py in <module>
32
33 from . import Image
---> 34 from ._util import isDirectory, isPath
35
36 LAYOUT_BASIC = 0
ImportError: cannot import name 'isDirectory' from 'PIL._util' (/usr/local/lib/python3.10/dist-packages/PIL/_util.py)
---------------------------------------------------------------------------
NOTE: If your import is failing due to a missing package, you can
manually install dependencies using either !pip or !apt.
To view examples of installing some common dependencies, click the
"Open Examples" button below.
JHello guy's, How can I run 2 instances of SD ( a1111 ) at the same time ?
EHey @mikemenders , try RV5 (not 5.1) with the 8400 VAE. I personally had better luck. And yes, SD1.5 vs. RV models are worlds apart.
ERegardless though your pics look great!
Sso the --share command thing is broken i believe
SCould not create share link. Please check your internet connection or our status page: https://status.gradio.app
Startup time: 13.1s (import torch: 1.2s, import gradio: 0.5s, import ldm: 0.3s, other imports: 0.8s, load scripts: 4.6s, create ui: 0.3s, gradio launch: 5.4s).
Startup time: 13.1s (import torch: 1.2s, import gradio: 0.5s, import ldm: 0.3s, other imports: 0.8s, load scripts: 4.6s, create ui: 0.3s, gradio launch: 5.4s).
EHey @mikemenders , this is interesting. Do you know the difference between this vs. adetailer, or have you tested them against each other?
TI also want to know this. Do I need different SD folders?
TIt will only work if SD is on local. You can't use that command on RunPod or Kaggle
SI only use local
SDo I need an extension or something as well?
PQuestion for any of you: I have trained LORAs for my wife, myself, and my two kids. Whats the best way to compose a family portrait or group photo? Any good tutorials out there? I haven't found much.
Si thought about doing something lie that but my approach was going to train multiple concepts within the same model with dreambooth
E@xPIXELPAPIx you'll need to use the Loras in a single prompt. So for example it'll be something like "husband:lora" sitting next to "wife:lora" with "son:lora" chasing "daughter:lora" in a living room. Something along those lines. I'm not too experienced with complicated Loras myself but it should be something along those lines and there are a few youtube videos out there of people using multiple loras.