Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,46 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: creativeml-openrail-m
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: creativeml-openrail-m
|
| 3 |
+
tags:
|
| 4 |
+
- stable-diffusion
|
| 5 |
+
- sd-xl
|
| 6 |
+
- lora
|
| 7 |
+
- anime
|
| 8 |
+
- hatsune miku
|
| 9 |
+
- character-style
|
| 10 |
+
- text-to-image
|
| 11 |
+
base_model: stabilityai/stable-diffusion-xl-base-1.0
|
| 12 |
+
instance_prompt: MikuPix2
|
| 13 |
+
library_name: diffusers
|
| 14 |
+
inference: true
|
| 15 |
+
widget:
|
| 16 |
+
- output:
|
| 17 |
+
url: images/from-PixAI-1973228487944483186-1.png
|
| 18 |
+
text: >-
|
| 19 |
+
MikuPix2, Hatsune Miku doing chest press yoga pose on balcony mat at
|
| 20 |
+
sunrise, loose sports bra pulled down under large breasts covered by hair
|
| 21 |
+
and ribbon tie, back arched pressing chest forward, cute focused moan blush
|
| 22 |
+
+ wink at viewer, twin-tails flowing in breeze, thigh-high leggings, golden
|
| 23 |
+
light on skin, ultra-sexy ultra-flexible anime, 8k
|
| 24 |
+
pipeline_tag: text-to-image
|
| 25 |
+
---
|
| 26 |
+
|
| 27 |
+
# MikuPix2: The Evolution of MikuPixAI (SDXL)
|
| 28 |
+
|
| 29 |
+
<Gallery />
|
| 30 |
+
|
| 31 |
+
**MikuPix2** is the official successor to the archived **MikuPixAI** model. While the original model served us well, this version is rebuilt from the ground up on the **SDXL 1.0** architecture to provide superior anatomy, better lighting, and more accurate character details.
|
| 32 |
+
|
| 33 |
+
### 🔄 What's New (Transition from MikuPixAI)
|
| 34 |
+
* **Architecture:** Upgraded from SD 1.5/Ancient tech to **SDXL 1.0**.
|
| 35 |
+
* **Prompting:** Use the new trigger word `MikuPix2` for optimal results (Legacy prompts from MikuPixAI are still compatible but work best with the new trigger).
|
| 36 |
+
* **Stability:** Fixed the "hair-artifacting" issues found in the previous archived version.
|
| 37 |
+
|
| 38 |
+
### 🚀 Usage & Recommendations
|
| 39 |
+
* **Trigger Word:** `MikuPix2`
|
| 40 |
+
* **LoRA Weight:** `0.75`
|
| 41 |
+
* **Base Model:** `SDXL 1.0 Base`
|
| 42 |
+
|
| 43 |
+
---
|
| 44 |
+
**Note:** The previous model (MikuPixAI) is now considered a **Legacy Archive**. It remains functional for older projects, but for all new high-fidelity generations, **MikuPix2** is the recommended standard.
|
| 45 |
+
|
| 46 |
+
**Developed by:** [M5 Dev]
|