BlackRealFishnetsP for FLUX_dev – Thin mesh black fishnet pantyhose

The Lora model is specifically trained to render Black Fishnet pantyhose (thin mesh) for FLUX.

There are more Loras related to pantyhose theme:

-          for regular glossy pantyhose for FLUX , you can download GlossyPantyhose for FLUX

-          for nylon feet for FLUX  you can download Nylonfeet for FLUX

-          for opentoe heels with nylon for FLUX  you can download Opentoe heels with legwear for FLUX

Moreover many hosiery themed Lora for FLUX, SDXL, Pony are avalilable in the Hosiery collection

Concerning this Lora a native 1024x1024 BlackRealFishnetsP set was selected to train an FLUX Lora, with reliable thin mesh fishnet layout.

If you like this Lora, please REVIEW IT, use it and SHARE PICTURES.

For a great output please read the following instructions.

Rendering the thin mesh pantyhose pattern details is a very challenging issue, due to the maximum allowed dimensions of the training set (1024x1024) that is barely enough to train a model for a very thin pattern. Results are definitely better than base image without Lora, but some tricks for best results are explained in the following.

Triggering prompt keyword:

BlackRealFishnetsP 

Lora weight:

A value of 0.8 or 100 is highly recommended. Higher values can trigger artifacts.

So a good activation keyword for lora is:

<lora:BlackRealFishnetsP_FL:0.9>

Sampling steps:

About 30-40 sampling steps, with Euler for first generation.

WARNING: Fine details of this Lora require in most of cases to use the inpaint algorithm only on legs, to improve the fishnet pattern.

Image dimensions

Better result are obtained with a base image of about 1200 px dimension on the long image side with an image width/length ratio of  1.5. After the base generation upscaling is a good practice.

USE DEMO IMAGES PROMPTS AS REFERENCE

Final statements:

1. You are solely responsible for any creative work using this model.

2. You cannot use this model to intentionally produce or share illegal images or content.

3. Furthermore can't use this model for malicious, hurtful, defamatory, fraudulent or political purposes.