Style: Drawn / Anime | Character-Focused | Strong PosingHard Mix LoRA is an experimental build designed for anime-style character generation. It produces a clean drawn look, handles faces and bodies with consistency, and gives reliable posing and positioning across prompts. Different than the checkpoint, trained with similar data but more focused and on a smaller data set.Im actively working on the next refined version with sharper eyes and improved facial detailing — this experimental release is a preview of what’s to come.<ul><li>✅ Strong drawn/anime illustration style</li><li>✅ Faces and bodies render with good accuracy</li><li>✅ Posing &amp; positioning work naturally</li><li>✅ Pairs well with Illustrious LoRAs &amp; Illustrious Base</li><li>✅ Best results between 0.4 – 1.0 strength (tested at 0.6 &amp; 0.8)</li></ul>⚡ My other models, along with ongoing updates and exclusive builds below. V6 is here!: <a target="_blank" rel="ugc" href="https://linktr.ee/XRart">https://linktr.ee/XRart</a>.

Hard Mix V1 (Experimental) | Illustrious

Small update, New version is on patreon

Files with SHA256: 17f5ba999ac1544a0a3daf567bcdaf0b82bb3f1ee2a47c0f08e0aeaf82555812