sd-parseq VS deforum-for-automatic1111-webui

This is how it was done https://github.com/rewbs/sd-parseq/discussions/106

rewbs/sd-parseq (github.com)

sd-parseq for parameter control / keyframing.

I need to modify Parseq to add a way for you to be able to specify absolute values and then render deltas/slopes for use in loopback-based animations. Otherwise it's too hard to say e.g. "I want a 180 degree rotation over 16 beats": you have to figure out what increments will add up to 180 manually.

Cool! Correct me if I'm wrong, but isn't that kind of animation essentially loopback with incremental transforms on each frame before feeding back into SD? If so, that's why I'm so excited about the colour improvements too (I'm the author of sd-parseq, a tool for SD animations: https://github.com/rewbs/sd-parseq)! :)

As described in this post, I've been working on a script for automatic1111 UI with its own separate companion UI for "sequencing" parameter changes over multiple generations, resulting in interesting videos that you can control precisely. This video gives a quick & dirty demo to give you a better idea of what it does. The output described in that video looks like this. If you're feeling adventurous you can check out the param flows directly in the parseq UI here (NB: only tested in Chrome so far, uses massive URLs which some browsers don't like). The code is here: https://github.com/rewbs/sd-parseq

I've been playing with the idea of a custom script + companion UI to give fine grained control over various parameters when generating videos. It's still very rough but I think it's just about ready to share: https://github.com/rewbs/sd-parseq

Deforum https://github.com/deforum-art/deforum-for-automatic1111-webui

might be deforum (project) (webui extension) or img2img loopback wave (webui extension). it's definitely something with img2img though. that's clear from how it's warping.

Thanks for the workflow. To simplify the process I created a PR that hopefully will make it a bit faster, specially if you're working in rented cloud GPU's.

The sequence after the train with the flashing background is the only exception (1:03 to 1:24), that part was made using the Deforum notebook: https://deforum.github.io/

ive gotten better results applying some color correction to my loops using the same way as they do in deforum https://github.com/deforum-art/deforum-for-automatic1111-webui/blob/automatic1111-webui/scripts/deforum_helpers/colors.py

🔸 Deforum extension for Automatic1111 (Local Install): https://github.com/deforum-art/deforum-for-automatic1111-webui

Deforum plugin settings files: render, coloration. The black and white canny edges + scribble-mode mid-stage rendered video. You can check the prompts keyframing file to see if I got the characters names right. Despite my best efforts, some of the characters appearances were off, I think it was either the character names were incorrectly spelled (I was referring to the Touhou wiki) or they had been underrepresented in the dataset or there was overlap between their naming and some other franchises. There is also the issue of random unrelated characters appearing when the prompt contains no characters and instead is purely abstract, I guess it relates to this specific model retraining.