Very sick of the continuous concept blend of SD 1.5, I've created this embedding model to obtain exactly the clothing I wanted, a now I want to share my work with the community. This is the second of the NStyles series.

Description:

Embedding of 32 vectors, trained on a dataset of 57 images, to generate high-heels, under the knee, platform, leather boots. This embedding born from the necessity of creating this particular boot style: trying to prompt them gives with every checkpoint totally random results, so I selected the output images and built a dataset with the specified boots. As any other embedding, it can be personalized with positive and negative prompt, and regulating its weight. Black pantyhose may be included sometimes, if you don't want them, simply put "pantyhose" in negative prompt.

Activation prompt: "boots"