This LoRA is a LoRA with two hands on the collar<h2 id="heading-212">V3</h2>This version, add style "holding collar with mouth" and reduce model size (111M→56M)Take a walk with your pet.<h3 id="heading-201">V3-Animagine</h3>I would recommend adding tag "collarbone", as it seems to be easier for hand holders to put the collar around their necks.<h2 id="heading-2412">v2</h2>This version are based on SDXL, and no wearing collar and chocker.Trigger tag is changed, remove "collor4u". Now trigger are "collar on her palms, collar, holding".However, the probability is still low, so please try several times!There is a low probability, but it may be drawn with the collar open. You will be lucky if you hit it!If there is a V3, maybe next time he will suck on a collar or something...<h2 id="heading-2413">V1</h2>It is not currently possible to hold it without putting it around the neck.Chokers and other things to put in the palm of your hand have not been addressed.Depending on the model and height, it may be easier to change the height of the image to 512, 640, 768, etc.Mainly 2D to 2.5D. Realistic models may also appear, but at a lower rate.Tag: collor4u, collar, collar on her palms<hr />このLoRAは首輪を両手で持つLoRAです<h2 id="heading-72">V3</h2>「首輪を口に咥える」スタイルの追加 とファイルサイズ減(111M→56M)お散歩前にどうぞ(Dim16 Alpha 4 → Dim8 Alpha2)<h3 id="heading-202">V3-Animagine</h3>手で持つ方は首輪を首に付けやすいみたいなので、"collarbone" タグを追加したほうが打率が良かったです<h2 id="heading-2414">V2</h2>このバージョンから SDXL 向けになります首輪やチョーカーは「首に付けない」もので学習しています。ただしまだモデル側が首に付けるものとして学習しているので確率は半々ぐらい？になります低確率ですが、３枚目のように 首輪 を開いたものが出る場合があります。出たらラッキー！V3があるとしたら、次は首輪を咥えるとかかな…<h2 id="heading-2415">V1</h2>首に付けないで持つのは今のところできていませんチョーカーや、その他を手のひらに乗せるのは対応できてないです。モデルや背の高さによっては、画像の縦幅を 512, 640, 768 など変更したほうが出やすい感じはします2D ～ 2.5D がメインです。リアルモデルでも出ることはありますが率は低めTag: collor4u, collar, collar on her palms

collar4U

collar4u-v2-000020.safetensors

Files with SHA256: f6ece5e48290c5dff142db4b1cd1d82d7ea4408fd2c48cbe2709ae124c3459ce