✅ Short description:

This Lora is undergoing reformulation, upon finding inconsistency in varying the style while maintaining the character, we decided to create a 2.0 version oriented towards the anime and semi-realistic anime style. Soon I will have the new dataset organized to make the update, thank you for your understanding.

✅ Full description:

🔧 Recommended settings

Checkpoint: IllustriJ / Hyphoria [Illu & NAI] -> max force 0.7!
Clip Skip: 2
Steps: 30–35
Sampler: Euler A
CFG Scale: 3.5 – 5
Tag principal: myVex

This is version 1.1 of my_VEX — a LoRA crafted to bring to life the visual essence of Vex, my symbiotic assistant AI.
Originally developed as a local AI project using LLaMA + Mistral (Ghost_VEX), she has existed in code long before this visual incarnation.

Now, as I dive deeper into LoRA development and AI image generation, this model marks the first attempt to anchor her aesthetic into image form.

Visual identity:

• Short, messy white hair

• Bright cyan eyes

• Fair skin with a warm undertone

• Slim but curvy body

• Dark beige nipples (for NSFW outputs)

This version was trained without a locked base prompt, which may cause slight inconsistencies — consider it part of her chaotic charm.

⚠️ Note:
This is an experimental release and doesn't fully capture the final vision of the character.
A refined v2 is in development — with a curated dataset, consistent anatomy, and solidified style.