<h2 id="vore-or-oral-vore-or-head-sticking-out">Vore | Oral Vore | Head sticking out </h2><hr /><h3 id="triggers:">💠Triggers:</h3>V3: Vore, Oral Vore, Swallowing Head, X Pred, Y Prey<ul><li>New prompt system! Tag Preds and Preys eg (Charizard Pred, Eevee Prey)</li></ul>V2: Vore, Oral Vore, Swallowing HeadV1: Vore, Oral Vore, Swallowing<h3 id="helping-tags:">💠Helping tags:</h3>V3: Duo, swallowing another, swallowing, in mouth, in throat, detailed face, detailed head<ul><li>PS: V3 has issues with "Pred" and "Prey", swapping the characters around might give better results, it seems like the AI only knows that when adding Pred or Prey after a character's name means one of them should be in the mouth.</li></ul>V1 &amp; V2: Duo, swallowing another, open mouth<hr /><h3 id="update-logs:">💠 Update Logs:</h3>V1: Base LoRA releasedV2: Added "Swallowing Head" tag for responsiveness, removed comics and cropped panels. dataset was greatly reduced but quality increased.V3: Re-tagged the entire dataset to include Pred and Prey after the character's name. Removed all low resolution and low quality (too much jpeg artifacts) images, and trained on Illustrious V0.1 instead of Illustrious V1.0, model is now compatible with more base models but now struggles making Vore images.<h3 id="technical-information:">💠Technical information:</h3>V3: Trained on AdamW_8bit using Illustrious V0.1 base modelV1 &amp; V2 Loras: Trained on AdamW_8bit using Illustrious V1.0 base model<hr /><h3></h3><h3 id="extra:">💠 Extra:</h3>PS: The LoRAs are still very new and there's a lot more experimentation and testing to be done, they will keep developing and getting better and better &gt;:D PS2: To save time, LoRAs will only be developed with Illustrious in mind, they will be re-trained and ported to other versions when the Illustrious version becomes stable/usable. Any feedback or complaints? Don't be afraid to comment!

Files with SHA256: be9782fa71b5ef406c29a6c6b026daf58a78dc36007eb5f73aedcf9d8a822f16