--- tags: - text-to-image - stable-diffusion - lora - diffusers - template:sd-lora widget: - text: spacephoto of a satellite output: url: images/2024-09-18_04-49-20_9568.png - text: spacephoto of a space station with solar panels output: url: images/2024-09-18_04-50-38_5877.png - text: spacephoto of a satellite with giant solar panels output: url: images/2024-09-18_04-58-37_2523.png - text: >- spacephoto of a soviet spacecraft with rockets, blurry polaroid, 1980s colors output: url: images/2024-09-18_08-13-04_3640.png - text: >- spacephoto of a soviet spacecraft with rockets, blurry polaroid, 1980s colors output: url: images/2024-09-18_08-13-34_9011.png - text: spacephoto of a circular NASA space station, blurry polaroid, 1980s colors output: url: images/2024-09-18_08-14-56_1187.png - text: >- spacephoto of a torus like NASA space station, earth in the background, blurry polaroid, 1980s colors output: url: images/2024-09-18_08-16-11_2724.png base_model: stabilityai/stable-diffusion-xl-base-1.0 instance_prompt: spacephoto --- # Spacephoto LoRA XL ## Model description This is a compact LoRA to create more realistic space stations, satellites and spaceships in Stable Diffusion XL based models. It also emulates the look and feel of early NASA photos. The LoRA was trained with 55 images with 1,500 steps. The base model is Stable Diffusion XL Base. It's a bit of a gamble whether a well-constructed space station, spaceship or satellite is generated, or just a chaotic array of modules, antennas and solar panels. Simply enter “spacephoto” as the trigger word and then specify whether you want a space station, a satellite or a spaceship and which attachments - such as solar panels, antennas, modules or rocket engines. The LoRA works best with models such as RealVisXL 5.0. ## Trigger words You should use `spacephoto` to trigger the image generation. ## Download model Weights for this model are available in Safetensors format. [Download](/IcelosAI/Spacephoto_LORA_XL_Base/tree/main) them in the Files & versions tab.