Alpha version: Caption-less dataset. 1000 Pokémon (CivitAI max, 1001-1025 missing), Official Ken Sugimori art, absolutely no variations.
Trigger: fluxfakemon
Trained on a too low a Network Dim by mistake as well.
Notes
To prompt, simply describe the Fakemon you want with the trigger word in the description.
I do have a WD-Vit-Large-V3 tagged version of the dataset, but it feels like it is lacking too much info, so I have set it aside for now.
I also have Gen 1 Pokemon captioned with short, basic captions based on what ChatGPT-4o knows of them. It does seem to break down on properly numbered Pokemon halfway through Gen 2, so continuing this way will be even more time consuming.
JoyCaption made too many mistakes in identifying the Pokemon, and I could only do 60 at a time.
If someone wants to lend a powerful PC to run captioning using some other VLM via TaggUI, please!
Based on above, Gen 1 with short captions will likely be the next version, if the interest seems here.
This can theoretically be used as a Ken Sugimori art style LoRA though I have not tried it in that way.
1. Dieses Modell dient nur Lernzwecken. Urheber- und Auslegungsrechte liegen beim Originalautor.
2. Bist du der Originalautor eines Modells, kontaktiere uns bitte zur Authentifizierung über unsere offiziellen Kanäle. Wir schützen die Rechte aller Schöpfer. Hier klicken, um es zu verifizieren.
