character generation?

[ppbrown]

Hi,
love the VN!
but.. just one problem with it.
It claims, "this is the prompt that made me", but it doesnt seem to work. At least not with the values and base file shown.
Could you please share what the actual base checkpoint file used was, and anything else different?

[comfyanonymous]

The images in the VN are untouched comfyui outputs so they contain the full workflow that was used to generate them.

https://github.com/comfyanonymous/ComfyUI_tutorial_vn/tree/master/game/images

[ppbrown]

i was hoping that was the case. but when I tried copying in every value(manually, from what I saw in the tutorial!) I did not get the same result. :(

I typed the values I saw into an editor, then cutnpasted them into comfy, with the same base checkpoint, and did not get the same output.
I actually got closer to the tutorial when I substituted "ghostmixv2.0" for v1-5-pruned-emaonly (which is what the tutorial claims it uses)
Maybe something changed in the SD code since you orignally baked the tutorial. Might you try it again and confirm?

For your cutnpasting convenience, here are my hand-written notes:

rand::  8566257.000
steps:: 20
cfg:  8
Sampler: euler
schedule: normal
denoise: 1.0

prompt::
(solo) photograph, fox ears, white hair, albino eyes, , business suit, black pants

teacher standing presenting teaching a classroom, blackboard

indoors daytime (HDR), happy

(fox tail:0.9)

neg::

(hands), text, error, cropped, (worst quality:1.2), (low quality:1.2), normal quality, (jpeg artifacts:1.3), signature, watermark, username, blurry, artist name, monochrome, sketch, censorship, censor, (copyright:1.2), extra legs, (forehead mark), depth of field) (emotionless) (penis)

[comfyanonymous]

The checkpoints used are WD-1.5 beta 2 for the first pass and AOM3A1 for the second upscale pass.

[ppbrown]

aahhhh.. much better!
Thank you very much
^.^