Home Creators Posts Import Register Favorites Logout
haven't archived this post yet. have a subscription? use the importer!

Content

Back to basics, with new technology. 😊

Yesterday I posted the Dev version on CivitAI, but you have the priority of the schnell version, which requires more processing and time to train.

🖼️ Gallery DEV version

🖼️ Gallery Schnell version

⬇️ Download Link

Same password of JK Bosom V03. I will send the password to new active patrons by direct message.

The Workflow I'm using on ComfyUI


Gen Parameters:

  • Best Settings DEV version:

    • Max Shift: 1.5

    • Base Shift: 1

    • 25-30 steps

    • Euler sampler

    • Simple Scheduler

    • Guidance 5 to 8

  • Best Settings Schnell version:

    • Max Shift: 1.5

    • Base Shift: 1

    • 4-6 steps

    • Euler sampler

    • Simple Scheduler

    • Guidance 5 to 8

Limitations:

This LoRA has problems with the flexibility of faces, I recommend using adetailer in ComfyUI.

Files

Previews only

Comments

julip_plum

Thanks for this! It's still pretty hard to control but early days for the model. Excited for your fine-tuning as you bring all that you did with v1.5 to Flux!