Years ago when gpu power was basically free through google colab I used that to my advantage by racking up sheepit credits rendering other people's projects on google's dime. I still have about 10 million credits right now and I don't think I'll be able to use them all.
SheepIt Render Farm
11–20 of 23 posts
Re: SheepIt Render Farm
#12Does a system exist where we can train open source LLM models in this way? Maybe distribute it across a cluster of computers with credit points that can be later used for inference.
Not for training, but execution in a SheepIt-similar way can be done with AI Horde, which not only does LLM inference but also Stable Diffusion. https://stablehorde.net/
Re: SheepIt Render Farm
#13I haven’t done render farm stuff since the 3DSmax 5 era in high school. Do they let you do stuff like ask to see a frame, kind of like when someone walks up to a production line to spot check the product?
You can just send a job with the specific frame you want to preview, get the results back and once you're happy, send the entire render job.
Re: SheepIt Render Farm
#14Earlier quoted context omitted.
Then make it a feature so that the person receiving the renders can accept/reject them. If they are rejected, then the renderer is not credited. If the person receiving racks up too many rejects, then they cannot submit new renders. The reject process would be hairy in that you'd need to explain why. It has potential for getting nasty, but any 2-party system can.
Then I'll be a dick and modify the client to take the frames anyway and still send a reject. It's going to be abused.
Oh. You're why we can't have nice things.
Seriously though, any platform is going to have its share of liars, cheaters, grifters, con artists, and outright scams. What the world relies on is that enough people aren't. The trick is attracting enough honest people while driving away enough of the dishonest people. it's a constant game of cat and mouse.
Re: SheepIt Render Farm
#15Earlier quoted context omitted.
Then make it a feature so that the person receiving the renders can accept/reject them. If they are rejected, then the renderer is not credited. If the person receiving racks up too many rejects, then they cannot submit new renders. The reject process would be hairy in that you'd need to explain why. It has potential for getting nasty, but any 2-party system can.
Then I'll be a dick and modify the client to take the frames anyway and still send a reject. It's going to be abused.
The website exists since 2013, and it seems like they managed just fine.
Re: SheepIt Render Farm
#16Earlier quoted context omitted.
Then make it a feature so that the person receiving the renders can accept/reject them. If they are rejected, then the renderer is not credited. If the person receiving racks up too many rejects, then they cannot submit new renders. The reject process would be hairy in that you'd need to explain why. It has potential for getting nasty, but any 2-party system can.
Then I'll be a dick and modify the client to take the frames anyway and still send a reject. It's going to be abused.
Have the reject frames re-rendered by someone else to find out the truth. Ban abusers.
Better yet, make all abusers render each other frames.
Re: SheepIt Render Farm
#17Years ago when gpu power was basically free through google colab I used that to my advantage by racking up sheepit credits rendering other people's projects on google's dime. I still have about 10 million credits right now and I don't think I'll be able to use them all.
This exudes chaotic good energy
Re: SheepIt Render Farm
#18Years ago when gpu power was basically free through google colab I used that to my advantage by racking up sheepit credits rendering other people's projects on google's dime. I still have about 10 million credits right now and I don't think I'll be able to use them all.
Are they not transferable?
Re: SheepIt Render Farm
#19Years ago when gpu power was basically free through google colab I used that to my advantage by racking up sheepit credits rendering other people's projects on google's dime. I still have about 10 million credits right now and I don't think I'll be able to use them all.
Are they not transferable?
From memory they specifically discourage a points economy.
It’s more just basic fairness and score keeping
Re: SheepIt Render Farm
#20Does a system exist where we can train open source LLM models in this way? Maybe distribute it across a cluster of computers with credit points that can be later used for inference.
Not for training, but execution in a SheepIt-similar way can be done with AI Horde, which not only does LLM inference but also Stable Diffusion. https://stablehorde.net/
Found it to be a bit of a mess given how many different possible configurations and models there are