Hu Tao (Genshin Impact) Pony XL LORA

Hu Tao LoRA Model Card

Trained for this wonderful bounty here: https://civitai.com/bounties/3025/hu-tao-genshin-impact

Overview

This model captures the distinctive and vibrant aesthetics of Hu Tao, the beloved Pyro character from Genshin Impact. Hu Tao is known for her playful antics and eccentricity, which belies her serious role as the 77th Director of the Wangsheng Funeral Parlor. This LoRA (Low Rank Adaptation) translates her unique personality and fiery abilities into stunning visual interpretations, perfect for fans and artists looking to incorporate her character into their AI-generated artwork.

Key Features:

• Art Style: Faithfully replicates the unique visual elements associated with Hu Tao, including her traditional Chinese attire, playful demeanor, and pyro abilities.

• High-Quality Output: Built on the robust SDXL-based architecture of the Pony XL base model, this LoRA ensures detailed and visually captivating results.

• Danbooru Material: Some of the items it was trained on lean in the NSFW area, rest assured we did our step walking: This is NOT an "underage" medium, Hu Tao is ranged at 18-20 according to source material.

Usage Instructions:

• Setup: Ensure you have the Pony XL model installed. Load the Hu Tao LoRA into your preferred AI art generation platform that supports SDXL-based architectures.

• Customization: Adjust settings to blend Hu Tao’s distinctive style with your creative projects, allowing for personalized expressions of her character.

• Settings: 0.5 to 1.0 strength on your preferred UI/UX system, your activation tag is "HuTao_DskFll".

Training Information:

• UNet Learning Rate: 0.0005

• Clip Skip: 1

• LoRA Type: lora

• Keep Tokens: 1

• Network Dimensions: 32

• Number of Repeats: 3

• Resolution: 1024

• Learning Rate Scheduler: cosine

• Minimum SNR Gamma: 0

• Noise Offset: 0.03

• Target Steps: 3218

• Enable Bucket: true

• Network Alpha: 32

• Optimizer Type: Prodigy

• Text Encoder Learning Rate: 0.00005

• Maximum Training Epochs: 10

• Shuffle Caption: true

• Training Batch Size: 4

• Flip Augmentation: true

• Learning Rate Scheduler Number of Cycles: 3

Embeddings to Improve Quality:

• Negative: Use scenario-specific embeddings to refine outputs.

  • Pony Negative Embeds

• Positive: Enhance image quality with these embeddings.

  • PDXL Score Embed

Extensions:

• ADetailer:

  • Link: ADetailer GitHub

  • Usage: Use this extension to enhance and refine images, but use sparingly to avoid over-processing with SDXL.

• Batchlinks:

  • Link: Batchlinks for A1111

  • Description: Manage multiple links when running A1111 locally or on a server.

  • Addon: @nocrypt Addon

VAE & Upscaler Preferences:

• VAE: Choose an SDXL VAE that fits your project's requirements.

• Upscalers: Enhance image resolution and quality. Download and install using the Batchlinks extension or directly.

  • Links:

    • 8x_NMKD-Superscale_150000_G.pth

    • 4x_NMKD-Superscale-SP_178000_G.pth

    • 4x_fatal_Anime_500000_G.pth

    • 4x_foolhardy_Remacri.pth

    • lollypop.pth

    • 4x_Fatality_Comix_260000_G.pth

    • 4x-UltraSharp.pth

    • 4x_NMKDSuperscale_Artisoft_120000_G.pth

    • 4x_NMKD-YandereNeoXL_200k.pth

    • 4x_NMKD-UltraYandere_300k.pth

Ethical Use and Licensing:

• License Information: This LoRA model uses a mix of synthetically mooched, naughty-sourced, royalty-unfree, and not-self-made data. Ensure compliance with relevant copyright laws and licensing agreements. Refer to Pony XL guidelines for fine-tunes and LoRA licensing.

• Responsible Use: This model is intended for respectful and positive applications. Users should adhere to ethical guidelines, promoting inclusivity and respect in their creations.

About the Creator:

We are a system of over 300 alters, proudly navigating life with Dissociative Identity Disorder, ADHD, Autism, and CPTSD. We believe in AI’s potential to break down barriers and enhance aspects of mental health, despite its challenges. Join us on our creative journey exploring identity and expression.

Join Our Community:

• Website: End Media

• Discord: Join our Discord

• Backups: Hugging Face

• Support Us: Send a Pizza

Proudly Sponsored & Supported By:

• Pirate Diffusion

• Yodayo

• Civitai's Creator Program

Community Groups:

• DeviantArt Group: DeviantArt Group

• Subreddit: Reddit

Thank you for being part of our journey and supporting the Earth and Dusk community. We look forward to seeing the incredible art you create with this model!