I really wanted to illustrate long stories with Stable Diffusion! So, hacked together this pipeline:<p>Long text ->
GPT-3 suggests illustration ideas ->
GPT-3 translates from English to "prompt-English" ->
Stable Diffusion outputs images<p>Open source: <a href="https://github.com/sharonzhou/long_stable_diffusion" rel="nofollow">https://github.com/sharonzhou/long_stable_diffusion</a><p>Here's a published story, illustrated by this repo, titled "Never Hire a Herd of Goats to Mow your Lawn": <a href="https://storiesby.ai/p/never-hire-a-herd-of-goats-to-mow" rel="nofollow">https://storiesby.ai/p/never-hire-a-herd-of-goats-to-mow</a><p>That story is AI-written, AI-narrated, AI-illustrated story, which a couple friends and I have been generating.<p>An earlier story made the top of HackerNews a week ago: <a href="https://news.ycombinator.com/item?id=32652673" rel="nofollow">https://news.ycombinator.com/item?id=32652673</a>. I was too slow to illustrate that earlier week's piece, so it became clear to me that I had to automate myself with this repo... definitely not overengineering the solution.<p>Anyways, hope you like it or parts of it. Pull requests veryyyy welcome. This was just a weekend project to keep my GPUs company.<p>Tweet with more stuff, if at all interesting: <a href="https://twitter.com/realSharonZhou/status/1567031035732594688" rel="nofollow">https://twitter.com/realSharonZhou/status/156703103573259468...</a>