Does anybody use ONNX for LLMs? I've only encountered it for small models packaged to run on consumer devices, and never seen it show up in any workstation or datacenter scenario.
I don't understand why this is set up the way it is. Maintaining a BitTorrent tracker for models seems like a reasonably good idea, but you have to "apply to join" to be able to contribute?
Just do it like how libgen does. Maybe publish a RSS torrent feed people can subscribe to for the newer publications - knowing someone is curating the feed.
Weird to have open licensed stuff on private tracker.
I can definitely picture a piece of content that's been scrubbed from everywhere except old copies of LLMs in which it still exists in a lossy form, with people trying to extract it from those old copies with the help of newer models. Private trackers have thrived for decades and will continue to do so because they allow a community to foster around a shared purpose (pardon the pun).
Although that being said, i doubt i would go through the trouble of signing up just for this purpose, personally.
Its standard for private trackers, you don't want a malware infested hell like piratebay, plus generally you would want people to seed back what they downloaded at least.
> generally you would want people to seed back what they downloaded at least
But the goal is preservation rather than sharing.
I'm confused about how the economics of that would work. If I wanted to contribute to this, I'd have to download the models, so I'd immediately be in the red, and then I could never get out of it because who's going to download these obsolete models? There's no download demand for them; if they were, the whole project would not be necessary.
I see the merit in keeping a decentralized archive for open models. But there has to be a better way to do it vs keeping it behind auth and an application form.
it's light, not like usual trackers, it's only to make sure people seed 1:1 and to reduce load on a couple of people seeding large models to everyone. the only way to do that is to enforce it, and for that you need a place where you can possibly enforce it without making it platform run, so, that's why you have this.
Why the restriction to safetensors and GGUF? onnx is the more common format and specifically made for interchange.
removed
Does anybody use ONNX for LLMs? I've only encountered it for small models packaged to run on consumer devices, and never seen it show up in any workstation or datacenter scenario.
I don't understand why this is set up the way it is. Maintaining a BitTorrent tracker for models seems like a reasonably good idea, but you have to "apply to join" to be able to contribute?
To what end?
Accountability?
You mean for people trying to poison the swarm? Or what? I'm genuinely confused.
people not maintaining 1:1 ratio, read about private trackers
The content will be mostly obsolete LLM models. I don't see the incentive to download them in large quantities.
Just do it like how libgen does. Maybe publish a RSS torrent feed people can subscribe to for the newer publications - knowing someone is curating the feed.
Weird to have open licensed stuff on private tracker.
I can definitely picture a piece of content that's been scrubbed from everywhere except old copies of LLMs in which it still exists in a lossy form, with people trying to extract it from those old copies with the help of newer models. Private trackers have thrived for decades and will continue to do so because they allow a community to foster around a shared purpose (pardon the pun).
Although that being said, i doubt i would go through the trouble of signing up just for this purpose, personally.
> I can definitely picture a piece of content that's been scrubbed from everywhere except old copies of LLMs in which it still exists in a lossy form
Yes. So it seems to me that the purpose of the project is to maintain a common good, but then I don't see how keeping it private achieves that.
And what are you going to train the models on?
Its standard for private trackers, you don't want a malware infested hell like piratebay, plus generally you would want people to seed back what they downloaded at least.
> generally you would want people to seed back what they downloaded at least
But the goal is preservation rather than sharing.
I'm confused about how the economics of that would work. If I wanted to contribute to this, I'd have to download the models, so I'd immediately be in the red, and then I could never get out of it because who's going to download these obsolete models? There's no download demand for them; if they were, the whole project would not be necessary.
What is the difference between this and https://pirateface.co/ ? Private tracker, and a focus on old models?
There doesn't need to be a difference, as with other torrent trackers, there are many of them, and I think thats good if nothing else, for redundancy.
I see the merit in keeping a decentralized archive for open models. But there has to be a better way to do it vs keeping it behind auth and an application form.
it's light, not like usual trackers, it's only to make sure people seed 1:1 and to reduce load on a couple of people seeding large models to everyone. the only way to do that is to enforce it, and for that you need a place where you can possibly enforce it without making it platform run, so, that's why you have this.
oink/waffles.fm [0] for models. very cool
[0]: https://en.wikipedia.org/wiki/Oink%27s_Pink_Palace
RIP waffles and red.
I love this, but why do I have to login tho? Would love to seed if you remove login
it's a private tracker
Support piratebay. The powers that be wants to take away your second amendment.
https://pirateface.co/audnai/penclaw-GLM-5.3-abliterated