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
What if you could allow people to contribute to decentralized training by simply duplicating a HuggingFace space?
In other words, enable "1-click to contribute compute".
This would be a really powerful tool, because it would greatly lower the activation barrier for people who are excited about a project to contribute compute. Those people are already on HF, and by duplicating the space they can simply choose whatever hardware they can afford to contribute, for as long as they want to contribute. Then they stop contributing by spinning down the space.
I am not an expert in training and definitely not in distributed training...
...but my sense is that if I were going to attempt this, prime would be the best tool to use.
What do you think? Any ideas for how to implement this?
The text was updated successfully, but these errors were encountered:
This morning I had an idea:
What if you could allow people to contribute to decentralized training by simply duplicating a HuggingFace space?
In other words, enable "1-click to contribute compute".
This would be a really powerful tool, because it would greatly lower the activation barrier for people who are excited about a project to contribute compute. Those people are already on HF, and by duplicating the space they can simply choose whatever hardware they can afford to contribute, for as long as they want to contribute. Then they stop contributing by spinning down the space.
I am not an expert in training and definitely not in distributed training...
...but my sense is that if I were going to attempt this, prime would be the best tool to use.
What do you think? Any ideas for how to implement this?
The text was updated successfully, but these errors were encountered: