Get early access to build and test build, be able to try all epochs and test them by yourself on Patreon or contact me for support on Discord

Training settings:

35 images, 25 steps, 10 epochs, you find many epochs (with test images using plot xy script) on my Patreon while supporting me.

Recommended settings:

• Make sure you are aware on the usage instructions of LORA

• Sampler: DPM++ SDE Karras (Recommended for best quality, you may try other samplers)

• CFG Scale : 5 to 9

• LORA weight for txt2img: reccomended 0.4-0.7

• Keys: wearing belly dance dress, dancing

• Adviced Image Size: 512x768

This model has been trained on images without background, it was my first time doing this and i discovered it is a bit problematic to generate the background you want.

My advice is to start your prompt with "a woman wearing belly dance dress" with a weight of the model of atleast 0.4, adviced 0.5 but it may give better results in the dress design till 0.8.

Then put a big weight to the kind of background you want, atleast 1.3. or as you will see in many of the images i posted with the model, you will get an empty background.

After this experience i strongly advice to always use images with background for LORA model trainings.

If you'd like to support my efforts and have access to early and bleeding edge builds, please think about joining my Patreon: https://www.patreon.com/betterCuteAsians