You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Only problem with this is it requires us to have a live server and backend and database to store that since it's live/constantly changing data. Right now the entire site is statically built content hosted for free.
You can start by just adding your own votes (dev ones, or added via PRs). Add it to the json as a separate field.
It's still better than nothing ...
Alternatively, you could do something creative, like do a client-side lookup on number of google results for each :) I'm grasping
I would greet this feature too. This project seems very interesting, but for me as a novice to ai upsacling the sheer amount of available models is just overwhelming. A versatile rating system, maybe coupled with Github authentication would help to cluster the quality of the available models - maybe also with a comment functionality.
Think of it like printables.com which is a good example of what this platform could be.
As an idea how to make this feature suitable for a community-driven project, you could consider to switch from Github Pages to Vercel.com. Their hobby plan could suit the project and offers some serverless storage functionality in the meantime.
Weighing in to add my support for this idea too. OpenModelsDB seems like it could be such a great resource, great potential, but choosing a model right now is a bit like stabbing in the dark, imo.
But whatever feedback or rating system is implemented there should be considerations for hardening against botting or automated voting.
Activity
joeyballentine commentedon Jan 11, 2024
Only problem with this is it requires us to have a live server and backend and database to store that since it's live/constantly changing data. Right now the entire site is statically built content hosted for free.
I agree though, that would be really nice.
ipeevski commentedon Jan 11, 2024
You can start by just adding your own votes (dev ones, or added via PRs). Add it to the json as a separate field.
It's still better than nothing ...
Alternatively, you could do something creative, like do a client-side lookup on number of google results for each :) I'm grasping
calvinrbnspiess commentedon Feb 3, 2024
I would greet this feature too. This project seems very interesting, but for me as a novice to ai upsacling the sheer amount of available models is just overwhelming. A versatile rating system, maybe coupled with Github authentication would help to cluster the quality of the available models - maybe also with a comment functionality.
Think of it like printables.com which is a good example of what this platform could be.
As an idea how to make this feature suitable for a community-driven project, you could consider to switch from Github Pages to Vercel.com. Their hobby plan could suit the project and offers some serverless storage functionality in the meantime.
joeyballentine commentedon Feb 3, 2024
Wow, that hobby plan seems pretty generous. We'll definitely consider switching to that. Thanks!
n0kovo commentedon Mar 31, 2024
Came by to make this exact proposal.
metamountain commentedon Jun 28, 2024
I would love comparable pictures....Perhaps just like the ones for digital cameras...
BangDroid commentedon Aug 24, 2024
Weighing in to add my support for this idea too. OpenModelsDB seems like it could be such a great resource, great potential, but choosing a model right now is a bit like stabbing in the dark, imo.
But whatever feedback or rating system is implemented there should be considerations for hardening against botting or automated voting.
FNsi commentedon Nov 7, 2024
Could be another cvital but enhance model only