sd-webui-bayesian-merger
digraph
sd-webui-bayesian-merger | digraph | |
---|---|---|
4 | 6 | |
118 | 48 | |
- | - | |
8.3 | 9.2 | |
4 months ago | 2 months ago | |
Python | Rust | |
MIT License | MIT License |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
sd-webui-bayesian-merger
-
Questions regarding SDXL architecture
As context, for a while now I've been looking into improving current automated model merging methods(and some other stuff, gonna present something nice in a few months when I get to finishing it). It's practically a 26 dimensional stochastic black box optimization problem and that was kind of fine so far (shoutout to https://github.com/s1dlx/sd-webui-bayesian-merger and all the work people put there, especially the author).
-
Ask HN: Most interesting tech you built for just yourself?
I’ve implemented a Bayesian optimizer for stable diffusion model merging [0]. This is because I do not have patience and/or time to try all different block combinations by hand. It started as a personal thing but now multiple people are working on it and a small community was born.
[0] https://github.com/s1dlx/sd-webui-bayesian-merger
- Automatic Bayesian Block Merger for SD-webui
digraph
-
Ask HN: Most interesting tech you built for just yourself?
My own purpose in using it is to be able to get back to any link that I've read or have potentially wanted to read at a later point in time.
You scan see screenshots here: https://github.com/emwalker/digraph.
-
My Bad Habit of Hoarding Information
I have the same habit and wrote a web app to catalog the links I come across:
https://digraph.app/
-
Google Search Is Dying
This was kind of the idea behind a side project I started a few years ago:
https://digraph.app/
https://blog.digraph.app/2020-06-13-democratization-of-searc...
I definitely think crowd-sourcing and a well-conceived reputation management system that can influence results are good next areas for exploration.
-
Paul Graham's Twitter thread on Search engines and SEO spam
> I think building search vertical that are hand-curated would be very interesting to see.
That was my inspiration behind a side project I made a few years ago — a decentralized, hand curated "search engine" [0]. Never got beyond the side project stage. But I see promise in this in the future. Eventually we'll figure out that human and moderated curation is better than the best machine learning.
[0] https://github.com/emwalker/digraph
What are some alternatives?
FordACP-AUX - Ford CD changer emulator with AUX playback control using Arduino UNO
endoflife.date - Informative site with EoL dates of everything
iptv-filter
mblaze - Unix utilities to deal with Maildir
uBlock-Origin-dev-filter - Filters to block and remove copycat-websites from DuckDuckGo, Google and other search engines. Specific to dev websites like StackOverflow or GitHub.
hacker-scripts - Based on a true story
loda-identify-similar-programs - Measure how similar LODA programs are
ghidra - Ghidra is a software reverse engineering (SRE) framework
adm-zip - A Javascript implementation of zip for nodejs. Allows user to create or extract zip files both in memory or to/from disk
malten - Anonymous ephemeral messaging
notebooks - Just various notebooks I sometimes write to help me, no unifying theme