A LoRA trained on the artwork from the artist Tortuga. It was trained on Pony Diffusion V6 XL and WAI-ANI-NSFW-PONYXL.
For the record: I did not test the base version on any other Pony checkpoint merges.

For the sake of transparency, showcase images for v1.1 (I don't remember about v1.0) were generated without using ADetailer or inpainting. However, they were upscaled 2x using R-ESRGAN 4x+ Anime6B with 0.3-0.4 denoising strength.

PonyV7 announcement: As you might know, PonyV7 has been released, but from what I've seen / read - at this moment it doesn't look good. So I'm not migrating these to V7, at least not until I have more info.

(I am thinking about switching to WAI-Illustrious anyway).

Keywords:

v1.1

No keyword required.

v1.0

PonyXLV6: tortugaponystyle

WAI-ANI-NSFW-PONYXL: waitortugastyle

Seems to work well with the default weight of 1.

This LoRA was trained with scale weight norms set to 0, so I am not sure if it will work well with other LoRAs. So far I had no problems mixing it with characters LoRAs.

Feedback or just a nice comment would be appreciated.

Changelog:

09.10.2025 - Version 1.1. Refined dataset and tags. Removed trigger word.

05.10.2025 - Uploaded version for WAI-Pony merge. Same training settings.

20.08.2025 - Initial release for PonyXLV6