Art style based on えぶるぃ (@everyntrge) (<a target="_blank" rel="ugc" href="https://ci-en.dlsite.com/creator/2935">https://ci-en.dlsite.com/creator/2935</a>)v2 Release Changes:<ul><li>Removed some old training images for consistent art style</li><li>v2_AnimagineXLV31 is trained on Animagine XL V3.1 (<a target="_blank" rel="ugc" href="https://civarchive.com/models/260267">https://civarchive.com/models/260267</a>)</li><li>v2_PonyDiffusionV6XL is trained on Pony Diffusion V6 XL (<a target="_blank" rel="ugc" href="https://civarchive.com/models/257749">https://civarchive.com/models/257749</a>)</li><li>v2_AnythingXL is trained on Anything XL (<a target="_blank" rel="ugc" href="https://civarchive.com/models/9409/or-anything-xl?modelVersionId=384264">https://civarchive.com/models/9409/or-anything-xl?modelVersionId=384264</a>)</li></ul>v2 Training Parameters:<ul><li>213 images with Repeats of 8</li></ul><hr />v1 Release:<ul><li>v1_AnimagineXLV31 is trained on Animagine XL V3.1 (<a target="_blank" rel="ugc" href="https://civarchive.com/models/260267">https://civarchive.com/models/260267</a>)</li><li>v1_PonyDiffusionV6XL is trained on Pony Diffusion V6 XL (<a target="_blank" rel="ugc" href="https://civarchive.com/models/257749">https://civarchive.com/models/257749</a>)</li></ul>v1 Training Parameters:<ul><li>389 images with Repeats of 4</li><li>Batch size of 5, Epoch of 5</li><li>Learning rate of 0.0003</li><li>Network Rank of 32, Network Alpha of 16</li><li>Clip skip of 2</li><li>No regularisation</li></ul>

えぶるぃ (@everyntrge) Art Style LoRA

hobbyanon_everyntrge_PonyDiffusionV6XL.safetensors

Files with SHA256: 7cd2577bed8b4434604c5d50bfb55f167de12672ea87f0032be54ab75b9ca274