The relevant code is linked below and is a mess, but the idea is:
1. Generate a base image
2. Use inpainting to expand the left-hand side of the root image. You can do this by submitting an image whose left side is transparent and whose right side is the left side of the root image
3. Ditto for right-hand side
4. Stitch the three separate images together
https://github.com/charlesjlee/twitter_dalle2_bot/blob/main/...
I used @minimaxir's aitextgen library to create a Twitter bot powered by GPT-2.
This bot takes Joe Biden's tweets as input and tries to respond in the first-person voice of Donald Trump. I didn't have any success with fine-tuning on a dataset of Trump's tweets, so I "included" all the information in the prompt to GPT-2 via [magic sauce](https://github.com/charlesjlee/trump_gpt2_bot/blob/main/twee...). It took a lot of fiddling, but I'm satisfied with the final output.
The relevant code is linked below and is a mess, but the idea is: 1. Generate a base image 2. Use inpainting to expand the left-hand side of the root image. You can do this by submitting an image whose left side is transparent and whose right side is the left side of the root image 3. Ditto for right-hand side 4. Stitch the three separate images together https://github.com/charlesjlee/twitter_dalle2_bot/blob/main/...