Anima vs. Illustrious XL: Which Base Model for a Flexible Anime Style LoRA?

Hy4ne · reddit · 2026-09-24

A user who previously trained an anime style LoRA on Illustrious XL (30-49 images of one artist, good style but weak cross-character consistency) is weighing Anima as a base model and asking for Kohya SS advice on dataset size, resolution, captions, learning rate, and epochs.

Original post →

More from Multimodal

• 4D reconstruction demo goes viral across platforms, attracting unlikely fans — bilawalsidhu · 2026-09-24
• Fable releases The Council, an AI-produced four-part mini series, episode one now live — TomLikesRobots · 2026-09-24
• Walkthrough: Blender MCP plus FLORA for precise product turntable animations — round · 2026-09-24
• NVIDIA and Nebius benchmark Cosmos 3 Super video serving: latency vs throughput on HGX B200/H200 — NVIDIA Developer · 2026-09-24
• ElevenLabs voice generation comes to Meta's personal AI agent Muse — lukeharries · 2026-09-24
• ElevenLabs' Big Week: AI Receptionist, Studio 4.0, and Scribe v2 Medical Cut Errors 35% — lukeharries · 2026-09-24

Multimodal channel →