hodgy/summer_krea2

Summer — Krea 2 LoRA

An identity/character LoRA for krea/Krea-2-Turbo, provided in the original fal/PEFT layout and a ComfyUI-compatible layout.

Suggested trigger word

Start prompts with summer. This word is inferred from the repository and filenames; the training captions are not included, so the exact trigger token could not be independently verified.

Example prompt:

Photorealistic portrait of summer, natural daylight, realistic skin texture, detailed eyes, soft background blur.

Repository image

Files

• summer_krea2_lora.safetensors — original fal/PEFT Krea 2 LoRA
• summer_krea2_lora_v2_comfyui.safetensors — ComfyUI-style tensor names
• bAaSF4LQ2AniLWXpIRo21_dCaQnXW2.png — repository image/example asset

Both weight files contain 528 LoRA tensors. Embedded metadata reports:

Parameter	Value
Base model	fal/krea-2
Rank / alpha	32 / 32
Weight dtype	bfloat16
Trainer	krea-2-trainer@2026-06-25-sched-warmup

Recommended settings

• Use only with a compatible Krea 2 workflow.
• Start around LoRA strength 0.8 and adjust within approximately 0.6–1.0.
• Use the ComfyUI file for workflows expecting diffusion_model.*.lora_down/lora_up keys.
• Describe framing, pose, clothing, lighting, and environment explicitly.

Limitations

The exact training steps, learning rate, dataset composition, and caption token are not included in the repository metadata. Identity consistency can decrease with extreme angles, distant framing, occlusion, or strong stylization. This adapter inherits the base model's limitations and license obligations.

Responsible use

Use only with appropriate rights and consent. Do not use the adapter for impersonation, deception, harassment, exploitation, or privacy violations. Disclose AI-generated imagery when it might otherwise be mistaken for reality.