Enjoying my character LoRAs? Support new releases through my <a target="_blank" rel="ugc" href="https://ko-fi.com/tigerfan69/tiers">$5/month Ko-fi membership</a>, follow my <a target="_blank" rel="ugc" href="https://civarchive.com/user/TigerFan69">Civitai releases</a>, and find my models on <a target="_blank" rel="ugc" href="https://huggingface.co/TigerFan69">Hugging Face</a>. Some characters launch with paid early access before becoming free—see each release for details. Thank you for supporting my work!Cam is an original fictional adult woman character LoRA for Krea 2. It is intended for creating character portraits and scenes with a consistent character identity and does not depict an actual person. This is a LoRA adapter, not a complete base model.Compatibility and installationThe creator has verified this adapter with Krea 2 and Krea 2 Turbo. Use a compatible Krea 2 base model and LoRA workflow; obtain the base model separately under its own terms. In ComfyUI, place the .safetensors file in models/loras/, refresh the model list, and select the adapter in your compatible LoRA loader.Trigger and usageVersion 2 trigger word is camwomanVersion 1 - Include Camv1 in your prompt. Start with your working Krea 2 workflow and adjust the LoRA strength, prompt and seed to suit your scene. A simple starting prompt is: Camv1, a fictional adult woman in a casual jacket, portrait photograph, natural daylight. Appearance varies with the base model, workflow and generation settings; no single prompt guarantees identical results.Training and releaseThe adapter metadata identifies Krea 2 and AI Toolkit training at 2,000 steps. Version v1.0 is an fp16 SafeTensor adapter. The gallery contains the creator’s sample outputs.License and accessCam is free to download and does not require a paid membership. The Krea 2 Community License Agreement and its Acceptable Use Policy apply. Review the license and this listing’s permissions before use or redistribution; free download does not remove those restrictions.

Krea 2 Cam — Fictional Character LoRA

I originally started with 12 photos and used AI to composite them into a single identity. I started with the same 12 photos but used a different technique to create training images, then used the new training images for v2.0.

Files with SHA256: c45a07c1ca8cbe8913fe23e6a162f8827f047e1f1a23883329e2a4a79212ff76