Type | |
Stats | 84 2,266 |
Reviews | (7) |
Published | May 1, 2024 |
Base Model | |
Training | Steps: 2,840 Epochs: 8 |
Usage Tips | Clip Skip: 1 Strength: 1 |
Trigger Words | grahatiaffxivdsk hair ornament, miqo'te, slit pupils, braided ponytail, red eyes, animal ears, x hair ornament, red hair, cat ears, braid, neck tattoo, swept bangs |
Hash | AutoV2 891343B092 |
Fixed version is here: https://civitai.com/models/429295?modelVersionId=478298
Overview: The G'raha Tia LoRA model is designed to bring the beloved character from Final Fantasy XIV into the realm of AI-generated art. Focused on G'raha Tia's appearance and characteristics from Patch 5.3 onwards, this model captures his essence post-Shadowbringers, including elements from the Endwalker expansion and even the upcoming Dawntrail content. This LoRA is built on the Pony XL base model, leveraging the power of SDXL-based architecture for high-quality outputs.
Key Features:
Character Authenticity:
Faithfully represents G'raha Tia's look and feel from Final Fantasy XIV, ensuring accurate and detailed renditions.
Includes elements from his journey post-Shadowbringers, Endwalker, and even Dawntrail, providing a comprehensive portrayal of his evolution.
Spoiler Alert:
This model contains spoilers for players who have not completed Endwalker or seen the Dawntrail trailer. Users should be aware of potential plot reveals when using this model.
Fun Additions:
Features unique and fun elements such as G'raha Tia enjoying tacos, inspired by the Dawntrail trailer, adding a playful touch to the character representation.
About Me:
We are a system of over 300 alters, proudly navigating life with Dissociative Identity Disorder, ADHD, Autism, and CPTSD. We believe in the potential of AI to break down barriers and enhance aspects of mental health, even as it presents challenges. Our creative journey is an ongoing exploration of identity and expression, and we invite you to join us in this adventure.
Join Our Community:
We're always looking for more content creators to collaborate with and support. Join us on our various platforms:
Website: End Media
Discord: Join our Discord
Backups: Hugging Face
Support Us: Send a Pizza
Proudly Sponsored & Supported By:
Join Our Community Groups:
Technical Details:
Data Sources:
The model is trained on a mix of game snapshots, including scenes from Endwalker and Dawntrail, ensuring a rich and diverse dataset.
Specific attention to maintaining low noise levels (.03 or below) and optimal gamma settings (kept at 0) to ensure quality output.
Integration:
Designed as a LoRA for the Pony XL base model, making it compatible with SDXL-based architecture for seamless integration and use.
Usage Instructions:
Setup:
Ensure you have the Pony XL model installed, as this serves as the foundational architecture for the G'raha Tia LoRA.
Integration:
Load the G'raha Tia LoRA into your preferred AI art generation platform that supports SDXL-based architectures.
Customization:
Adjust settings to fine-tune the representation of G'raha Tia, blending elements from different patches and expansions to achieve your desired look.
Ethical Use and Licensing:
License Information:
Users must ensure their use complies with relevant copyright laws and licensing agreements with Square Enix, the creators of Final Fantasy XIV.
Responsible Use:
This model is intended for creative and respectful applications. Users should adhere to ethical guidelines in their creations, promoting inclusivity and respect.
Disclaimer:
The G'raha Tia LoRA model aims to provide a tool for fans and creators to explore and celebrate the character of G'raha Tia from Final Fantasy XIV. Users are encouraged to use this model in ways that are supportive, respectful, and inclusive. Any inappropriate or harmful use of the model is strongly discouraged, in line with community standards and ethical practices.
Notes:
Current Issues:
The model has some gamma issues that require a retrain to correct. Users might experience suboptimal performance with noise levels above .03 and gamma settings not kept at 0.
Extensions:
ADetailer:
Link: ADetailer GitHub
Usage: This extension assists with detailing and refining images. It’s a powerful tool for enhancing the precision of your artwork, though used sparingly with SDXL to avoid over-processing.
ControlNet:
Usage: Occasionally used for more precise control over image generation. A tutorial is not yet available due to time constraints.
Batchlinks:
Link: Batchlinks for A1111
Description: Useful for managing multiple links when running A1111 locally or on a server.
Addon: @nocrypt Addon
VAE & Upscaler Preferences:
VAE:
Recommendation: Select an SDXL VAE that fits your project's requirements.
Upscalers:
Instructions: These upscalers can be downloaded and installed using the Batchlinks extension or directly to your drive. They significantly enhance image resolution and quality.
Links: