Takamichi Style (Pony + SD1.5)

LOCON
Original


Updated:

No showcase images available, the model won't be visible to others.

https://www.buymeacoffee.com/pettankoutei

If you like my content consider supporting me! 😊

1 Cookie = Early access to future releases! 🍪❤️

Locon trained on the style of Takamichi for Pony Diffusion v6 XL based models.

Pony version

All sample images uses a weight of 0.8 and no trigger word.

This one turned out a bit easier to train, due to more consistent dataset. This one uses the same training parameters, but I intend to experiment a bit more, so if you have any tips or more advanced knowledge for better training with Pony let me know through Chat/DM.

SD 1.5 version

All sample images uses a weight of 0.6 and no trigger word. They were generated at 640x960 with 2x hires. fix.

This time I tested a different set of paramaters for regular LoRA (instead of LyCORIS/LoCon). Also, the training was longer with, with different optmizer, a slower learning rate and more steps. I definitely overshoot the number of steps and over trained a bit. I'm not 100% satisfied with this training, but it still works well.

A bit difficult to use, but at lower weights it works fine. Depending on your checkpoint you'll need to play with it and adjust the blocks as you see fit.

At resolutions like 512x768 you will often need Adetailer or hires.fix to repair faces. You may try higher resolutions like 640x960, but it depends a lot on the checkpoint since its SD1.5, so it may be a bit of anatomy gacha. Samples are generated on 640x960 and hires. fix.

In terms of usage, there are no trigger words and best weight is 0.6-0.8, but you should adjust accordingly.

Disclamer

All the models that I have developed and published on this platform are for entertainment purposes only. They are derived from the base Stable Diffusion model, which means it also includes its inherent issues, limitations and unfiltered capabilities. With that in mind, I am in no way responsible for any content that the user creates with malicious intent while using any of these models. The user assumes complete responsibility for the misuse of these tools and any unethical content created with them. This disclaimer is subject to the laws of the United States of America and the state of California. I reserve the right to remove or report any content that violates the terms of service, ethical standards, or applicable laws.

License conditions

Takamichi Style for Pony Diffusion v6 XL and Stable Diffusion 1.5 © 2024 by Pettan Koutei is licensed under Attribution-NonCommercial-ShareAlike 4.0 International

The model deployment is abnormal, please re-upload/contact customer service.

Version Detail

Pony
<p><strong>🔍 <u>Version info:</u></strong></p><p>Initial release. Trained 100 images for 2200 steps.</p><p><strong>📋 <u>Training Parameters:</u></strong></p><ul><li><p>Base Model: <a target="_blank" rel="ugc" href="https://civitai.com/models/257749?modelVersionId=290640">Pony Diffusion V6 XL</a></p></li><li><p>Training Resolution: 1024</p></li><li><p>Training Batch Size: 1</p></li><li><p>Type: LoCon</p></li><li><p>DIM / Alpha: 8 / 4</p></li><li><p>Conv Dim / Alpha: 24 / 6</p></li><li><p>Train On: Unet + TE</p></li><li><p>Optimizer: Prodigy</p></li><li><p>Min SNR Gamma: 8</p></li><li><p>Bucket Resolution: 512 - 2048</p></li><li><p>Bucket Steps: 64</p></li><li><p>Noise Offset: Normal (0.035)</p></li><li><p>Image Count: 100</p></li><li><p>Steps: 2200</p></li></ul>

Project Permissions

    Use Permissions

  • Use in TENSOR GREEN Online

  • As a online training base model on TENSOR GREEN

  • Use without crediting me

  • Share merges of this model

  • Use different permissions on merges

    Commercial Use

  • Sell generated contents

  • Use on generation services

  • Sell this model or merges

Comments

Related Posts

No posts yet