Rem character LoRA for Anima.

Trigger: Rem (put it first in the prompt)

Recommended version: Rem V2 (camera-balanced retrain; use this for new gens).

V1 = original first publish (kept for comparison).

Rem V2 was retrained with a more balanced camera mix (eye-level / wide / low) so the prompt camera should win more often. The high-angle bias is reduced but not gone.

Suggested strength:

• Rem 0.8–1.0 for portraits / close-ups / eye-level shots

• Rem ~0.35–0.6 for full-body, wide, or low-angle shots

• Avoid 1.2+ — likeness overcooks and camera freedom drops

Stack used for samples: Anima aesthetic-v1.1 + Aesthetic Quality Modifiers @1.0 + (optional) Zoda NSFW Detailer v3_anima @0.7. Sampler dpmpp_2m / simple, CFG 4, 40 steps, 768×1344.