README.md

---
license: creativeml-openrail-m
language:
- en
base_model: MidnightRunner/MIDNIGHT_NAI-XL_vPred1
tags:
- SDXL
- noobai-XL
- Vpred-2.0
- text-to-image
- ComfyUI
- Automatic1111
- Diffuser
pipeline_tag: text-to-image
library_name: diffusers
datasets:
- LaxharLab/NoobAI-XL-dataset
metrics:
- FID
- IS
widget:
- text: >-
    (high quality), (masterpiece), (detailed), photorealistic, cinematic, film
    grain, very awa, (beksinski:1.3)1girl, solo, voluminous, long hair, looking
    at viewer, semi-realistic, newest (2023), seductive expression, piercing
    gaze, open mouth, sitting, blurry background, pale skin, platinum white
    hair, busy background, dress, bow, ribbon, jewelry, bare shoulders, upper
    body, sleeveless, pointy ears, hand up, white dress, bracelet, medium
    breasts, sleeveless dress, skindentation, see-through clothes, silt eyes,
    black sclera, fangs, red bow, armlet, gold, red ribbon, vampire, dress bow,
    oshino shinobu, black choker, lips, looking at viewer, open mouth, ring,
    smile, smilingwide, teeth, prehensile tail, holding with tail, tail gripping
    whiskey jug, whiskey text label "jack daniels", jack daniels logo, tail
    gripping, sharp teeth
  parameters:
    negative_prompt: >-
      lowres, worst quality, low quality, bad anatomy, bad hands, 4koma, comic,
      greyscale, censored, jpeg artifacts, overly saturated, overly vivid,
      (multiple views:1.1), (bad:1.05), fewer, extra, missing, worst quality,
      jpeg artifacts, bad quality, watermark, unfinished, displeasing, sepia,
      sketch, flat color, signature, artistic error, username, scan, (blurry,
      lowres, worst quality, (low quality:1.1), ugly, (bad anatomy:1.05), artist
      name, (patreon username:1.2)
  output:
    url: Holding_Whiskey.jpeg
---

# MIDNIGHT_NAI-XL_vPredV2

**Model Type:** Diffusion-based text-to-image generative model

**Previous Base Model:** [MIDNIGHT_NAI-XL_vPredV1](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV1)

**License:** [CreativeML Open RAIL++-M](https://github.com/CompVis/stable-diffusion/blob/main/LICENSE)

## Model Description

MIDNIGHT_NAI-XL_vPredV2 is a specialized fine-tuning of the MIDNIGHT_NAI-XL_vPredV2 (NAI-XL) model, designed to enhance realistic precision, compositional coherence, and versatile style integration. This model excels in generating high-quality realistic images with vibrant colors while minimizing overexposure.

## Usage Recommendations

### **Sampling Methods**
MIDNIGHT_NAI-XL_vPredV2 is optimized specifically for **Euler (normal)**.  
Use **ModelSamplingDiscrete** with **V-prediction** and **ZsNR set to true**.  
Other samplers may not provide stable results, and **V-prediction models do not support other samplers**.

### **CFG Scaling**
**Dynamic CFG Plugin is bypassed as a backup for potential future needs.**  
Manually adjust **CFG scaling within a range of 5-6** for the best balance.  
For optimal results, a **preferred setting of 5.3** is recommended.

### **Custom Workflow**
For an optimized generation process, use the [**MIDNIGHT1111_Chasm 02-05-25**](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV2/blob/main/MIDNIGHT1111_Chasm%2002-05-25.json) ComfyUI workflow.  
This workflow is specifically designed to **leverage the strengths of MIDNIGHT_NAI-XL_vPredV2**, providing a streamlined and efficient image generation pipeline.

  
## MIDNIGHT1111_Chasm

For an optimized generation process, consider using the custom workflow [MIDNIGHT1111_Chasm 02-05-25](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV2/blob/main/MIDNIGHT1111_Chasm%2002-05-25.json). This workflow is tailored to leverage the strengths of the MIDNIGHT_NAI-XL_vPredV1 model, providing a streamlined and efficient image generation pipeline.

![MIDNIGHT1111_Chasm Workflow](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV2/resolve/main/MIDNIGHT1111_Chasm%20Workflow.png)

*Note: The above image is a preview of the `MIDNIGHT1111_Chasm` workflow.*

### Method I: reForge without MIDNIGHT1111_Chasm Workflow

1. **Installation:** If not already installed, follow the instructions in the [reForge repository](https://github.com/Panchovix/stable-diffusion-webui-reForge) to set up.
2. **Usage:** Launch WebUI and use the model as usual.

### Method II: ComfyUI *with* MIDNIGHT1111_Chasm Workflow

1. **Installation:** Follow the setup instructions in the [ComfyUI repository](https://github.com/comfyanonymous/ComfyUI).
2. **Workflow Sample:** Utilize the provided [ComfyUI workflow sample](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV2/blob/main/MIDNIGHT1111_Chasm%2002-05-25.json) for guidance.

### Method III: WebUI without MIDNIGHT1111_Chasm Workflow

1. **Installation:** Follow the instructions in the [WebUI repository](https://github.com/AUTOMATIC1111/stable-diffusion-webui) to set up.

2. **Navigate to the WebUI Directory:** Before updating or switching branches, ensure you're inside the `stable-diffusion-webui` folder
    command: |
      ```bash
      cd stable-diffusion-webui
      ```

3. **Switch to the Development Branch (Optional, for testing new features):** If you want to use the latest features from the development branch, run:
    command: |
      ```bash
      git switch dev
      git pull
      ```
    ⚠️ **Note:** The `dev` branch may contain bugs. If stability is your priority, it's best to stay on the `main` branch.

4. **Update WebUI (Main or Dev Branch):** To pull the latest updates while on either branch, run:
    command: |
      ```bash
      git pull
      ```
    🔄 **Restart WebUI after updating to apply changes.**"

5. **Configuration:** Ensure you're using a stable branch, as the dev branch may contain bugs.

### Method IV: Diffusers without MIDNIGHT1111_Chasm Workflow

```bash
import torch
from diffusers import StableDiffusionXLPipeline
from diffusers import EulerDiscreteScheduler

ckpt_path = "/path/to/model.safetensors"
pipe = StableDiffusionXLPipeline.from_single_file(
    ckpt_path,
    use_safetensors=True,
    torch_dtype=torch.float16,
)
scheduler_args = {"prediction_type": "v_prediction", "rescale_betas_zero_snr": True}
pipe.scheduler = EulerDiscreteScheduler.from_config(pipe.scheduler.config, **scheduler_args)
pipe.enable_xformers_memory_efficient_attention()
pipe = pipe.to("cuda")

prompt = """masterpiece, best quality,artist:john_kafka,artist:nixeu,artist:quasarcake, chromatic aberration, film grain, horror \(theme\), limited palette, x-shaped pupils, high contrast, color contrast, cold colors, arlecchino \(genshin impact\), black theme,  gritty, graphite \(medium\)"""
negative_prompt = "nsfw, worst quality, old, early, low quality, lowres, signature, username, logo, bad hands, mutated hands, mammal, anthro, furry, ambiguous form, feral, semi-anthro"

image = pipe(
    prompt=prompt,
    negative_prompt=negative_prompt,
    width=832,
    height=1216,
    num_inference_steps=28,
    guidance_scale=5,
    generator=torch.Generator().manual_seed(42),
).images[0]

image.save("output.png")
```

## e621/Danbooru Artist Wildcards for A1111 & ComfyUI Enclosed in CSV & TXT Formats

To enhance the model's performance and specificity, the following trigger word lists in CSV format are included:

- [`danbooru_artist_webui.csv`](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV2/blob/main/danbooru_artist_webui.csv)
- [`danbooru_character_webui.csv`](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV2/blob/main/danbooru_character_webui.csv)
- [`e621_artist_webui.csv`](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV2/blob/main/e621_artist_webui.csv)
- [`e621_character_webui.csv`](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV2/blob/main/e621_character_webui.csv)

These lists provide recognized tags for various artists and characters, facilitating more accurate and tailored image generation.

The wildcard file in 'TXT' format is included and designed for seamless integration with **AUTOMATIC1111** and **ComfyUI**, optimized for dynamic prompt generation using artist data from **e621** and **Danbooru**.

- **TXT Format:** Sanitized artist tags by removing URLs and converted from `.csv` to `.txt` format for improved readability across different extensions.
- **Dual Dataset Support:** Supports both e621 and Danbooru datasets to enhance art style diversity.
- **Smooth Randomization:** Structured with trailing commas for seamless wildcard cycling during prompt generation.

## How to Use Wildcards

### For A1111
1. **Install:** [stable-diffusion-webui-wildcards](https://github.com/AUTOMATIC1111/stable-diffusion-webui-wildcards)  
2. **Place the `.txt` file in:**  
   ```
   /A1111/extensions/stable-diffusion-webui-wildcards
   ```
3. **Use in your prompt like this:**  
   ```
   __e621_artist_wildcard__, very awa, masterpiece, best quality, amazing quality
   ```
   ```
   __danbooru_character_wildcard__, very awa, masterpiece, best quality, amazing quality
   ```
   ```
   __e621_artist_wildcard__, __danbooru_character_wildcard__, very awa, masterpiece, best quality, amazing quality
   ```

### For ComfyUI
1. **Install:** [ComfyUI-Impact-Pack](https://github.com/ltdrdata/ComfyUI-Impact-Pack)  
2. **Place the `.txt` file in:**  
   ```
   /ComfyUI/custom_nodes/ComfyUI-Impact-Pack/wildcards
   ```
   or  
   ```
   /ComfyUI/custom_nodes/ComfyUI-Impact-Pack/custom_wildcards
   ```
3. **Use the wildcard node to trigger dynamic randomization in your workflows.**

## What’s Included in Wildcards

TXT formatted file containing clean, artist-focused wildcard files ready for dynamic prompt workflows in A1111 and ComfyUI.
  - [danbooru_artist_wildcard.txt](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV2/blob/main/danbooru_artist_wildcard.txt)
  - [danbooru_character_wildcard.txt](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV2/blob/main/danbooru_character_wildcard.txt)
  - [e621_artist_wildcard.txt](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV2/blob/main/e621_artist_wildcard.txt)
  - [e621_character_wildcard.txt](https://huggingface.co/MidnightRunner/MIDNIGHT_NAI-XL_vPredV2/blob/main/e621_character_wildcard.txt)

## Acknowledgments

Special thanks to:

- **Development Team:** Laxhar Lab
- **Coding Contributions:** Euge
- **e621/Danbooru Wildcards** [ipsylon0000](https://civitai.com/user/ipsylon0000)
- **Community Support:** Various contributors

## Additional Resources

- **Guidebook for NoobAI XL:** [English Version](https://civitai.com/articles/8962)  
- **Recommended LoRa List for NoobAI XL:** [Resource Link](https://fcnk27d6mpa5.feishu.cn/wiki/IBVGwvVGViazLYkMgVEcvbklnge)  
- **Fixing Black Images in ComfyUI on macOS (M1/M2):** [Read the Article](https://civitai.com/articles/11106)
- **Creative Solutions and Services:** [Magnabos.co](https://magnabos.co/)
  
## License

This model is licensed under the [CreativeML Open RAIL++-M License](https://github.com/CompVis/stable-diffusion/blob/main/LICENSE). By using this model, you agree to the terms and conditions outlined in the license.