Open Model Radar · Dossier

Olesya Poryadko (This person does not exist)

Character/person LoRA by rakamik, base model Qwen 2.1, no fixed trigger words.

LoRAQwen-Image 2.1

Highlights

  • No trigger words needed; just write a description
  • 401 total downloads · 29 likes · 1 version
  • File size approx. 152.1 MB
  • Platform-labeled as general content

Scope

  • Suited for: character/person direction, used with the Qwen-Image 2.1 base model
  • Suited for: stacking with other LoRAs on the same base model (reduce each weight by 0.1–0.2)
  • Not suited for: direct cross-base-model use—feeding a Qwen 2.1 LoRA to another base model will noticeably break it

Tips

  • Start with weight 0.8; too high will overfit
  • If it looks grayish or overly stylized, lower the weight first; don't rush to switch samplers
  • The base model must be Qwen-Image 2.1; choosing the wrong base model is the most common cause of failure

Facts

Type
LoRA
Base model
Qwen-Image 2.1
Org
rakamik
Downloads
401

Links