A request of @Inmel. Hope you like it, mate!
Veronika Zemanová is a retired Czech porn actress. For her TI I have used a dataset of 30 images, taken from various sites. The method I'm using as of lately tends to produce pretty accurate results already before reaching the 1000 steps, and it seems to be the case again with this embedding, which, as usual, might be a tiny bit overcooked already. However, and as the example pictures prove, you can easily build a pretty good prompt without even touching the embedding weight. I've basically replicated the training process I used for Maria Guardiola and Camila Morrone (have a look at the "About this version" section of the latter if you want more info on my process).
Description
FAQ
Comments (10)
Damn! That was fast!
Also, maybe I try this one out after all... For science...
Have you made any tutorials? Written or video? I would pay just to watch you go through a training session.
....and now unretired in the metaverse.
Holy shit that is awesome!
Retired, but her boobs will be relevant forever!!
Really great work here by everyone to date. Getting very close to being a true likeness of Veronika :-)
The last cowboy one had the best likeness but all are great.
I did note that on one of the posted images AI decided to give her a 6th finger on her hand 😆
Hi guys ... I am actually a Veronika super fan (run a fan tribute site dedicated to her). I have to admit though that I am a total newbie when it comes to Stable Diffusion but have been reading how you can provide a series of images to improve the rendering. I am an expert with Photoshop so can easily provide any post processing support. I saw on one comment thread here a link to a series of images you used as a basis for the model. I of course have many many high resolution images of virtually every angle and position of her body and face. I would love to help out to improve the model by either providing an additional series of images or learning how to use those to create what I think is the .PT file.
I will try and read the Camilla Monroe "About section"to see if I can glean any useful information.
Let me know ... am willing to help.
Update . . I did install Stable Diffusion AUTOMATIC1111 on my computer and managed to create a few samples using the provided vz3m4n0v4.PT embedding file. I assume that is the file based on the taught images. Hopefully I can figure out how to create another one based on higher resolution images of Veronika ;-)
Update Apr 12: I have figured out how to create the custom embeddings now ... I plan to create one for each of the "Veronika Vintages" ... natural, first implants and second implants. We'll see how that works out in the end :-)
I absolutely love your stuff, but I seem to be having a major issue with one, and it's kinda a real mood killer... lol. Any thoughts on this one? Sorry it popped up as such a long message. I just copied everything related to the error from the console.
"*** Error verifying pickled file from C:\stable-diffusion-webui\embeddings\Veronika Zemanová\vz3m4n0v4.pt
*** -----> !!!! The file is most likely corrupted !!!! <-----
*** You can skip this check with --disable-safe-unpickle commandline argument, but that is not going to help you.
***
Traceback (most recent call last):
File "C:\stable-diffusion-webui\modules\safe.py", line 92, in check_pt
with z.open(data_pkl_filenames[0]) as file:
File "C:\Users\Ken\AppData\Local\Programs\Python\Python310\lib\zipfile.py", line 1535, in open
raise BadZipFile("Bad magic number for file header")
zipfile.BadZipFile: Bad magic number for file header
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "C:\stable-diffusion-webui\modules\safe.py", line 137, in load_with_extra
check_pt(filename, extra_handler)
File "C:\stable-diffusion-webui\modules\safe.py", line 104, in check_pt
unpickler.load()
_pickle.UnpicklingError: invalid load key, 'E'.
---
*** Error loading embedding vz3m4n0v4.pt
Traceback (most recent call last):
File "C:\stable-diffusion-webui\modules\textual_inversion\textual_inversion.py", line 219, in load_from_dir
self.load_from_file(fullfn, fn)
File "C:\stable-diffusion-webui\modules\textual_inversion\textual_inversion.py", line 178, in load_from_file
if 'string_to_param' in data:
TypeError: argument of type 'NoneType' is not iterable"
So glad that this TI is in the community, but an update would be great. It's really good but there's something about the chin/jaw that isn't quite right IMHO...
Ah, I See You're a Man of Culture As Well.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.