Stacking 10+ LoRAs on Krea 2 Raw FP8: chasing max photorealism in ComfyUI
DragonfruitNo9822 · reddit · 2026-10-30
A ComfyUI user is chasing maximum photorealism with a trained model: Krea 2 Raw FP8 checkpoint plus Qwen3-VL CLIP, stacking a dozen LoRAs including Dora Identity V2 at 1.50, skin details at 2.00, and skin tone at 1.85. Results are decent but skin still looks plastic and anatomy/lighting need work, all while preserving subject identity.
They ask the community which LoRAs to keep, drop, or reweight; optimal sampler, steps, CFG and resolution; and how to test changes incrementally without sacrificing identity consistency. A help post, but also a snapshot of a typical photoreal portrait LoRA stack.
More from Multimodal
- Liquid AI launches LFM3 model family; 1.7B hits 185 tok/s on-device — JosephJacks_ · 2026-10-30
- Open-source ComfyUI film studio turns one prompt into an editable, refinable movie — Goat_bless · 2026-10-30
- Abiza turns a voicenote into a realistic AI influencer UGC video in seconds — adrianscottcom · 2026-10-30
- One-person AI music video: Suno + Hermes Agent + ComfyUI on a single RTX 3090 — Ok-Wolverine-5020 · 2026-10-30
- Same girl every scene vs. a new face each run: what a character LoRA actually changes — wuraang9 · 2026-10-30
- Projection mapper open-sources ComfyUI node pack that drafts full light shows locally on 12GB laptops — Puzzled_Parking2556 · 2026-10-30