Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Small script for AUTOMATIC1111/stable-diffusion-webui to create images between two seeds
| Date | Stars |
|---|---|
| 2026-07-24 | 311 |
| 2026-07-25 | 311 |
| 2026-07-28 | 311 |
| 2026-07-30 | 311 |
| 2026-07-31 | 311 |
| 2026-08-06 | 311 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# Seed Travel
Small script for AUTOMATIC1111/stable-diffusion-webui to create images that exists between seeds.
Will let you pick two seeds (or more) and then create a number of images while it morph between the two points of noise generated.
Samplers that work well are: Euler, LMS, Heun, DPM2, DPM++ 2M, LMS Karras, DPM2 Karras, DPM++ 2M Karras, DDIM and PLMS
[Look here for more information about issues with some samplers.](SAMPLER_TEST.md)
Batch Count and Batch Size will be ignored and set to 1. Trying to generate images with Euler a will fail and print an error. If you really want to use it there is a checkbox to allow it.
[Examples of what you can do](USER_EXAMPLES.md).
# Installation
The recommended way to install it is to:
1. Find "seed travel" in the list of available extensions in the webui, and then click Install.
If you can't find it in the list:
1. Go to the "Extensions" tab in the webui
2. Click on the "Install from URL" tab
3. Paste https://github.com/yownas/seed_travel.git into "URL for extension's git repository" and click install
4. ("Optional". You will need to restart the webui for dependensies to be installed or you won't be able to generate video files.)
Manual install (not recommmended):
1. Place the files from this repo in a folder in the extensions folder.
2. Restart. Pray. It might work properly. Maybe.
# What is "seed traveling"?
To understand what "seed traveling" is I first have to give a very very simplified explaination of how Stable Diffusion generates pictures.
It starts with noise and then tweak and poke it until it gets an image that looks like something that match the prompt you gave. And to get all this initial noise you use a seed. The seed is used to generate a bunch of random numbers which will be your inital noise. A different seed generates a different set of random numbers.
The problem is that the seed is a 32 bit value, around 4 billion different possible values. Which is a lot, but it is infinitesimal compared to the possible initial noises.
If we imagine all the possible noises as a map, a seed could be seen as one single point on that map. A huge universe with only a few possible places to stand and look at it. What seed traveling lets you do is to pick two "points" on this map and then travel between them, each step a new variation of noise that no one could ever reach by only using a seed.
So, interpolating between the noise from two seeds will not only look cool as a video, you will also get images you might never have seen.
# Usage
`Destination seeds`: Seeds to travel to from `Seed`. If initial seed is empty, the first destination seed will be chosen as start seed. Seeds placed between parentheses will be ignored. (Might be useful while testing.)
`Use random seeds`: Let you set `Number of random seeds` instead of typing them manually. `Seed` will be ignored unless comparing paths (see below).
`Steps`: Number of images to generate between each seed.
`Loop back to Initial seed`: When reaching the end, generate images to get back to the first seed.
`FPS`: The Frames Per Second of the video. It has a hidden feature where if you set this to a negative value, it will be used as the length (in seconds) of the resulting video(s). Setting this to 0 will skip generating a video.
`Lead in/out`: Amount of frames to be padded with a static image at the start and ending of the video. So you'll get a short pause before the video start/ends.
`SSIM threshold`: If this is set to something other than 0, the script will first generate the steps you've specified above, but then take a second pass and fill in the gaps between images that differ too much according to Structual Similarity Index Metric [(pdf)](https://www.cns.nyu.edu/pub/eero/wang03-reprint.pdf). A good value depends a lot on which model and prompt you use, but 0.7 to 0.8 should be a good starting value. More than 0.95 will probably not improve much. If you want a very smooth video you should enable RIFE below.
`SSExcerpt of 9,601 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:ff994e74cfcc5a82, topic:stable-diffusion, desc:stable diffusion, readme:stable diffusion