INKFLOW STUDIO — FREE ULTRA LoRA
Dariusz Kwiatkowski-Osei — 5,000 steps, rank 64

WHAT THIS IS
A complete, unmodified ULTRA-trained character LoRA (.safetensors) — the exact
same file and the exact same specification we sell. Not a trial, not a cut-down
version, not watermarked. We are giving one away so you can judge the quality
yourself instead of taking our word for it.

LICENCE — free, perpetual, worldwide, royalty-free
You may use this model commercially. Pay nothing, ever. No per-use fee, no
revenue share, no attribution required, no expiry. Use it in client work, in
films you sell, in anything you like.

You may not resell or redistribute the .safetensors file itself as a product.
Use the outputs freely; do not repackage the model.

WHO THIS IS
An ORIGINAL SYNTHETIC PERSON. Not a real individual, not a celebrity likeness,
not trained on any real person's photographs. That is why it is safe for
commercial work — most character LoRAs on the open marketplaces are not.

VERIFY THE SPEC YOURSELF
Do not trust our claim. Read the rank straight out of the file:

    import json, struct
    with open('dariusz-kwiatkowski-osei_ULTRA_5000steps_rank64.safetensors','rb') as f:
        n = struct.unpack('<Q', f.read(8))[0]
        hdr = json.loads(f.read(n))
    print(min(v['shape'][0] for k, v in hdr.items()
              if 'lora_A' in k or 'lora_down' in k))     # -> 64

If it does not print 64, you have caught us.

HOW TO USE IT
Drop it in your LoRA folder (ComfyUI: models/loras). Trigger token: dariuszkwiatsp
Base model: Krea 2 turbo. Suggested strength 1.0.

MORE
inkflowstudio.org — character and set LoRAs, persistent worlds, and books.
Questions, or want a character trained to your own brief? inkflowmail@gmail.com