CivArchive
    Mandi Collins - v0.2
    NSFW
    Preview 1
    Preview 2
    Preview 3
    Preview 4

    This Stable Diffusion LoRA profile is based on the Photomodel named Mandi Collins.

    Place it in "stable-diffusion-webui\extensions\sd-webui-additional-networks\models\lora" folder in Automatic1111 WebUI root folder.

    This has been trained using kohya-ss so make sure to have that extension installed!

    In txt2img, under "Additional Networks", select "Enable" and under Model 1 select "LoRA_MandiCollins_SD1.5_v0.2" and assign a weight 0.65. In the prompt text window use "MandiCollins" within your prompt to generate images.

    This has been trained on Stable Diffusion 1.5 model.

    Max resolution: 768,768

    Based on 281 photographic images.

    Description

    • 200 training iterations per image, instead of 100 in v0.1

    • Network Rank (Dimension) of 160 (instead of 128 in v0.1)

    • 281 training images (238 in v0.1)

    FAQ