Update README.md (FIX) [10/17/2025]
#4
by
qpqpqpqpqpqp
- opened
README.md
CHANGED
|
@@ -2,8 +2,6 @@
|
|
| 2 |
base_model:
|
| 3 |
- Laxhar/noobai-XL-Vpred-1.0
|
| 4 |
pipeline_tag: text-to-image
|
| 5 |
-
tags:
|
| 6 |
-
- not-for-all-audiences
|
| 7 |
---
|
| 8 |
|
| 9 |
Wahtastic Merge is a high-quality Stable Diffusion XL (SDXL) model designed to generate stunning images with improved aesthetics and excellent prompt adherence. This model is built upon the robust `noobai-XL-Vpred-1.0` base and has been further refined through the strategic merging of various other models and extensive additional training.
|
|
@@ -22,24 +20,24 @@ Up to V7, the model was entirely merged. V8 and above has additional fine-tuning
|
|
| 22 |
- In short, e6 tags have implications; `hyper_*` implies `huge_*`, and `huge_*` implies `big_*`
|
| 23 |
- Because of this, the model leans to associate big with huge, and huge with hyper, causing `big_*` to cause disproportionately large body parts at times.
|
| 24 |
- Natural language captioning
|
| 25 |
-
- Yes, CLIP sucks
|
| 26 |
-
- Using
|
| 27 |
-
- This is inspired by EasyFluff
|
| 28 |
- Superior style knowledge
|
| 29 |
- ~20k e6 artists with > 500 < 20 posts
|
| 30 |
-
- Potentially
|
| 31 |
|
| 32 |
*(Previously known as Pando Merge)*
|
| 33 |
|
| 34 |
-
Compute is expensive, and while plenty has been granted to me by kind acquaintances, a fair bit of money has been poured into the training process
|
| 35 |
If you like the model, or would like to help me offset the sunken cost of this, please consider donating:
|
| 36 |
|
| 37 |
**ETH Wallet Address for Donations:** `0x645BebF82373865eC520d8AC2527524BfB174FF8`
|
| 38 |
-
If you prefer PayPal
|
| 39 |
|
| 40 |
## How to Use
|
| 41 |
|
| 42 |
-
This model can be used with any standard SDXL-compatible interface or library (e.g., Diffusers,
|
| 43 |
|
| 44 |
### Recommended Settings
|
| 45 |
|
|
@@ -60,7 +58,7 @@ from diffusers import AutoPipelineForText2Image
|
|
| 60 |
import torch
|
| 61 |
|
| 62 |
pipeline = AutoPipelineForText2Image.from_pretrained(
|
| 63 |
-
"
|
| 64 |
torch_dtype=torch.float16,
|
| 65 |
variant="fp16",
|
| 66 |
use_safetensors=True
|
|
@@ -74,7 +72,7 @@ prompt=prompt,
|
|
| 74 |
negative_prompt=negative_prompt,
|
| 75 |
num_inference_steps=20,
|
| 76 |
guidance_scale=5,
|
| 77 |
-
height=1200,
|
| 78 |
width=832
|
| 79 |
).images[0]
|
| 80 |
|
|
@@ -98,6 +96,6 @@ If you find Wahtastic Merge useful and would like to support its continued devel
|
|
| 98 |
|
| 99 |
## Feedback and Issues
|
| 100 |
|
| 101 |
-
We welcome your feedback\! If you encounter any issues or have suggestions for improvement, please open
|
| 102 |
|
| 103 |
-----
|
|
|
|
| 2 |
base_model:
|
| 3 |
- Laxhar/noobai-XL-Vpred-1.0
|
| 4 |
pipeline_tag: text-to-image
|
|
|
|
|
|
|
| 5 |
---
|
| 6 |
|
| 7 |
Wahtastic Merge is a high-quality Stable Diffusion XL (SDXL) model designed to generate stunning images with improved aesthetics and excellent prompt adherence. This model is built upon the robust `noobai-XL-Vpred-1.0` base and has been further refined through the strategic merging of various other models and extensive additional training.
|
|
|
|
| 20 |
- In short, e6 tags have implications; `hyper_*` implies `huge_*`, and `huge_*` implies `big_*`
|
| 21 |
- Because of this, the model leans to associate big with huge, and huge with hyper, causing `big_*` to cause disproportionately large body parts at times.
|
| 22 |
- Natural language captioning
|
| 23 |
+
- Yes, CLIP sucks
|
| 24 |
+
- Using lodestones' natural-language captions, ideally some amount of natural language understanding can be brought back
|
| 25 |
+
- This is inspired by EasyFluff XL
|
| 26 |
- Superior style knowledge
|
| 27 |
- ~20k e6 artists with > 500 < 20 posts
|
| 28 |
+
- Potentially Danbooru artists too
|
| 29 |
|
| 30 |
*(Previously known as Pando Merge)*
|
| 31 |
|
| 32 |
+
Compute is expensive, and while plenty has been granted to me by kind acquaintances, a fair bit of money has been poured into the training process.
|
| 33 |
If you like the model, or would like to help me offset the sunken cost of this, please consider donating:
|
| 34 |
|
| 35 |
**ETH Wallet Address for Donations:** `0x645BebF82373865eC520d8AC2527524BfB174FF8`
|
| 36 |
+
If you prefer PayPal or Stripe, please contact me on Discord @velvet.toroyashi
|
| 37 |
|
| 38 |
## How to Use
|
| 39 |
|
| 40 |
+
This model can be used with any standard SDXL-compatible interface or library (e.g., Diffusers, Stable Diffusion WEBUI, ComfyUI).
|
| 41 |
|
| 42 |
### Recommended Settings
|
| 43 |
|
|
|
|
| 58 |
import torch
|
| 59 |
|
| 60 |
pipeline = AutoPipelineForText2Image.from_pretrained(
|
| 61 |
+
"VelvetToroyashi/WahtasticMerge",
|
| 62 |
torch_dtype=torch.float16,
|
| 63 |
variant="fp16",
|
| 64 |
use_safetensors=True
|
|
|
|
| 72 |
negative_prompt=negative_prompt,
|
| 73 |
num_inference_steps=20,
|
| 74 |
guidance_scale=5,
|
| 75 |
+
height=1200,
|
| 76 |
width=832
|
| 77 |
).images[0]
|
| 78 |
|
|
|
|
| 96 |
|
| 97 |
## Feedback and Issues
|
| 98 |
|
| 99 |
+
We welcome your feedback\! If you encounter any issues or have suggestions for improvement, please open a discussion on the Hugging Face repository.
|
| 100 |
|
| 101 |
-----
|