FLUX.1-dev-lora-adaml

Flux.1-Dev LoRA Adapter Trained on Me

LoRA Adapter for black-forest-labs/FLUX.1-dev trained on pictures of me with ai-toolkit

Model Details

Some Amusing Examples <img src="https://cdn-uploads.huggingface.co/production/uploads/65ba68a15d2ef0a4b2c892b4/ZG8SCJngcopb_eapEIWpf.png" width=400> <img src="https://cdn-uploads.huggingface.co/production/uploads/65ba68a15d2ef0a4b2c892b4/9EJxi0-wVpRmhndEOtqet.png" width=400> <img src="https://cdn-uploads.huggingface.co/production/uploads/65ba68a15d2ef0a4b2c892b4/yzM48yFaHmHgYFAQJdATw.png" width=400> <img src="https://cdn-uploads.huggingface.co/production/uploads/65ba68a15d2ef0a4b2c892b4/Ye1-e96hf1g1BPB29CyUE.png" width=400> <img src="https://cdn-uploads.huggingface.co/production/uploads/65ba68a15d2ef0a4b2c892b4/4P7ctpWuHNzLO8aVYJ8q8.png" width=400>

LoRA was trained with the trigger phrase 4d4m luc3k

Full training config available at train_lora_flux_me.yaml

Usage

With diffusers package
Note: FLUX uses ~70GBvram when loaded directly with diffusers
Note: Recommended to load at ~70% scale for best results

from diffusers import DiffusionPipeline

pipeline = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-dev")
pipeline.load_lora_weights("AdamLucek/FLUX.1-dev-lora-adaml", weight_name="flux_lora_Adam.safetensors.safetensors")
pipeline.to("cuda")

prompt = "4d4m luc3k as a DJ in a fancy nightclub wearing sunglasses."

out = pipeline(
    prompt=prompt,
    guidance_scale=3.5,
    num_inference_steps=50,
    cross_attention_kwargs={"scale": 0.7}
).images[0]

out.save("adam_flux_lora.png")

ComfyUI Setup

<img src="https://cdn-uploads.huggingface.co/production/uploads/65ba68a15d2ef0a4b2c892b4/7c_8aakMNzY4OPjSn5q-5.png" width=800>

File available at workflow_adamlora.json

Additional Details

Please see base model page black-forest-labs/FLUX.1-dev for all details on appropriate usage, licensing, and more.