Meet Poppy, or rather jp-Poppy my latest in my attempt to create Textural Inversions.<ul><li>The name is random.</li><li>Poppy is trained on 24 images.  I used the same 10 images from my <a target="_blank" rel="ugc" href="https://civarchive.com/models/146096/jp-gingerv1-275">jp-gingerv1-275</a> embedding and another 14 images of a different model to create this embedding. </li><li>She is not a real person, rather is created from a series of handpicked and manually captioned stock images of different models to create the redhead you see here.</li><li>24 images trained for 5000 steps.</li><li>10 vectors per token.</li><li>I used the following progressive learning rate 0.05:10, 0.02:20, 0.01:60, 0.005:200, 0.002:500, 0.001:3000, 0.0005</li><li>Batch size of 1 with 1 gradient accumulation step.</li></ul> I'd appreciate any feedback you may have and would love to see what you come up with. Thanks and enjoy!

Files with SHA256: 8359efa31e8999e00da3ee05fba6ba8a3d8ce9e49dc42abda7566e8c8d420507