-
Notifications
You must be signed in to change notification settings - Fork 20
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Older commit
data can reappear on the bittensor network
#68
Comments
So in my view this is not necessarily a problem. If you look at it this way its a feature:
This has happened to me a few times and I found it good that I could actually keep my original block for the model I submitted. With the evaluation being so slow right now I did have to pay an extra 2 tao but I have an option to not loose my submission to the cloners. |
It appears that this is a desired behavior, but will keep this open for further discussion down the road |
Update: for the specific case of modifying submissions after the
|
Given the following:
commit
callA miner can register again on the same subnet and their older
commit
data may still be retained. It is unclear how long the period for retaining thiscommit
call on the network would last, but this may lead to inconsistency issues if left unchecked.The above behavior can also be exploited when updating the huggingface repository details to effectively "submit" a new model under the guise of an older one.
Given that there are many implications for this current behavior, open discussion is encouraged here to decide on an optimal path forward
The text was updated successfully, but these errors were encountered: