Text-to-Image
Diffusers
lora
flux

Amaya Hart โ€” FLUX.1-dev LoRA

A character LoRA for the original fictional model Amaya Hart, trained on FLUX.1-dev (29 images, 1000 steps, rank 16).

Trigger word: amayahart

Signature look: heterochromia โ€” green left eye / brown right eye โ€” warm golden-brown skin, freckles, long dark wavy hair with honey highlights.

Usage (diffusers)

import torch
from diffusers import FluxPipeline

pipe = FluxPipeline.from_pretrained(
    "black-forest-labs/FLUX.1-dev", torch_dtype=torch.bfloat16
).to("cuda")
pipe.load_lora_weights("pkats94/amaya-hart-lora", weight_name="amaya-hart.safetensors")

image = pipe(
    "a photo of amayahart woman, candid iPhone selfie, soft window light",
    num_inference_steps=28, guidance_scale=3.5,
).images[0]
image.save("amaya.png")

Put amayahart near the start of the prompt. Recommended LoRA scale 0.8โ€“1.0.

Downloads last month
38
Inference Providers NEW

Model tree for pkats94/amaya-hart-lora

Adapter
(42601)
this model