- 
                Notifications
    You must be signed in to change notification settings 
- Fork 203
Implement CenteredClip in averager #379
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
base: master
Are you sure you want to change the base?
Conversation
07d8f5f    to
    6621bb5      
    Compare
  
    | Codecov Report
 
 @@            Coverage Diff             @@
##           master     #379      +/-   ##
==========================================
- Coverage   83.98%   83.68%   -0.31%     
==========================================
  Files          70       71       +1     
  Lines        6383     6443      +60     
==========================================
+ Hits         5361     5392      +31     
- Misses       1022     1051      +29     
 | 
6621bb5    to
    fee6195      
    Compare
  
    | weights: torch.Tensor, | ||
| tau: float = 1.0, | ||
| n_iters: int = 20, | ||
| stop_delta: Optional[float] = None, | 
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
preference: let's default to some reasonable delta and a very large n steps
| Random thoughts: 
 | 
| prev_diff = result | ||
|  | ||
| # We only need to update `diff` (not `result`) between iterations | ||
| diff.addmm_(-coeffs.repeat(n_peers, 1), diff) | 
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
| diff.addmm_(-coeffs.repeat(n_peers, 1), diff) | |
| diff -= coeffs @ diff | 
It seems like addmm_() doesn't work correctly if the destination is equal to one of the operands.
MeanReducer,CenteredClipReducer(btw, why not*Aggregator?)Factoryto smth else,make_reducer: Callable[...]