Stable Diffusion - From Gras to Grace

Created using a preconfigured Stable Diffusion Server with Automatic Web Gui on AWS: https://aws.amazon.com/marketplace/pp/prodview-v2qrtpqachrtc

We used the loopback script in the image 2 image tab to automatically create a sequence of 20-25 images with variations in denoising and adapting prompt after each sequence.

More background on the server used: https://www.ni-sp.com/how-to-run-stable-diffusion-on-your-own-cloud-gpu-server/. Stable Diffusion was running on a nVidia T4 GPU with 16 GB memory.

First prompt: A place in wales, tucked out of view magic happens, only seen by a few. for just one day, for only just one hour. The last summer’s day break at Gelli aur. there you must follow a winding trout stream. search all the oaks with a tiny light beam
Steps: 30, Sampler: Euler, CFG scale: 3.5, Seed: 400479045, Size: 1024x512, Denoising strength: 0.48, Denoising strength change factor: 1

Last prompt: A highly realistic, true to life portrait of a young woman looking similar to grace kelly, by karol bak, james jean, tom bagshaw, rococo, sharp focus, trending on artstation, cinematic lighting, hyper realism, octane render, 8 k, hyper detailed
Steps: 30, Sampler: Euler, CFG scale: 10, Seed: 3518801996, Size: 1024x512, Denoising strength: 0.64908963, Denoising strength change factor: 1.01

The 510 images where then put together into a mpg with this command:
ffmpeg -y -f image2 -framerate 7 -pattern_type glob -i "*.png" -c:v libx264 -r 30 -pix_fmt yuv420p -s 1024x512 output-7fps.mp4