Learned on <a target="_blank" rel="ugc" href="https://civarchive.com/models/4823/deliberate">Deliberate v.2</a>. Also tested on: <a target="_blank" rel="ugc" href="https://civarchive.com/models/10961/lazymix-real-amateur-nudes">LazyMix+</a> , <a target="_blank" rel="ugc" href="https://civarchive.com/models/6424/chilloutmix">ChilloutMix</a> , Pretty 2.5D V2 This LoRA perfactly works for 3D (like a game screenshot), Arts, Real Photos. Also, it's perfectly works with Triss body (Breasts size, body proportions and age), Emotions, Accessories, clothes, underwear. The LoRA was trained on Triss in clothes, nude body and only face. PS - This LoRA not trained to work with Triss dress. Only casual clothes and underwear. To my surprise, this LoRA works with very large absolutely different models. Learning on Deliberate - creates a miracle. Triss_v1 uses 58 pictures. Training on original game screenshots/renders and a little bit sfm models.Recommended promt: &lt;lora:Triss-v1:0.7&gt;,". The coefficient can be increased. All pics genned with --xformers. Not matching with seed - measurement error.

Files with SHA256: 7861ed8d591697e6d50e2db290d8983dc9514844959dd297957e56ffeffb22d5