This repo has tools to interpolate between prompts for Stable Diffusion models. It can generate pictures like the ones below as well as videos.
Interpolating between two fixed prompts (leftmost and rightmost image) in multiple steps:
Pairwise combination of prompts: each image is generated from the combination of the (prompt used for generating the) corresponding images in the first column and first row. E.g.: the image in row 3 column 2 is generated by combining the prompts used to generate image 2 and 3 and feeding them into SD.
Same as above but with different input prompts:
This repo also has tools to create (relatively) smooth videos, though this requires lots (!) of compute.