This is just a proof-of-concept lora, if you will. It is not very editable, and it meant to produce imagery only of that style anime girls.It is made on the base of extra lazy tagging way, which involves only barebones editing of auto-tagged images. This dataset consists of 140 images 100% SFW images, that are mostly anime girl portraits made in v5.1 MJ, of specific style.Generally, this is not supposed to be used for anythin complex, because it was not tagged in a way to be, but you still can try and show me your results :3 It will have a humongous background and ornament bleed, as they were complex and abstract, which is impossible to describe with the use of current auto-taggers.If you want it to be editable and still have some style effect, try weight from 0.5 to 0.75. That should make it editable and still be looking good :DFuture lazy styles will be more editable and diverse, and probably will be bigger, featuring higher ranks(&gt;=64) with dynamic normalization of images in dataset(This is already utilized in unreleased v4.2 of S12).

GPTS1 | Lazy Style LoRA | Extended LoRA

Base version. Lazy tagging, averagely low ranks. Nothing unusual(relatively to my current way of training). NAI-trained, therefore compatible with most anime models.

GPTS1 lite.safetensors

GPTS1 Lite v1.0

&lt;p&gt;This is just a proof-of-concept lora, if you will. It is not very editable, and it meant to produce imagery only of that style anime girls.&lt;/p&gt;&lt;p&gt;It is made on the base of extra lazy tagging way, which involves only barebones editing of auto-tagged images. This dataset consists of 140 images 100% SFW images, that are mostly anime girl portraits made in v5.1 MJ, of specific style.&lt;/p&gt;&lt;p&gt;Generally, this is not supposed to be used for anythin complex, because it was not tagged in a way to be, but you still can try and show me your results :3&lt;br /&gt;&lt;br /&gt;It will have a humongous background and ornament bleed, as they were complex and abstract, which is impossible to describe with the use of current auto-taggers.&lt;/p&gt;&lt;p&gt;If you want it to be editable and still have some style effect, try weight from 0.5 to 0.75. That should make it editable and still be looking good :D&lt;/p&gt;&lt;p&gt;Future lazy styles will be more editable and diverse, and probably will be bigger, featuring higher ranks(&amp;gt;=64) with dynamic normalization of images in dataset(This is already utilized in unreleased v4.2 of S12).&lt;/p&gt;

&lt;p&gt;Base version. Lazy tagging, averagely low ranks. Nothing unusual(relatively to my current way of training). NAI-trained, therefore compatible with most anime models.&lt;/p&gt;

Files with SHA256: 0ea360985c2b33abf5a3e4ff2938b7207e5940bbe3f5549975842d2bdcb43b0b