-
Notifications
You must be signed in to change notification settings - Fork 869
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
Create experimental mt-pytorch code example #1446
Conversation
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.
As all of this is experimental, I only left a note but apart from that lgtm.
while True: | ||
get_results_req = GetResultsRequest(task_ids=create_tasks_res.task_ids) | ||
|
||
# ------------------------------------------------------------------ Driver SDK | ||
get_results_res: GetResultsResponse = driver.get_results( | ||
req=get_results_req | ||
) | ||
# ------------------------------------------------------------------ Driver SDK | ||
|
||
results: List[Result] = get_results_res.results | ||
print(f"Got {len(get_results_res.results)} results") | ||
|
||
all_results += results | ||
if len(all_results) == len(task_ids): | ||
break |
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.
@danieljanes We don't have an early stop here but as this is all experimental I am just noting my thoughts here :)
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.
Oh, also would be cool to use tqdm
here too.
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.
I agree that these things would be cool, but my goal was more to create a minimal example here, which is why I'd suggest to keep is focussed :) regarding the early stopping, my assumption was that the server would enforce a timeout on all tasks anyways, so the driver can just wait until it gets either a full result or something that tells it that a particular task did not finish in time.
* main: (21 commits) Create Flower tutorial part 3 (#1474) Update notebook tutorial (#1470) Fixed epoch_loss calculation in intro tutorial (#1472) Update notbook tutorials (#1468) Update README (#1465) Add DP docs to index (#1467) Support optional client methods in VCE (#1450) Implement Driver create_tasks and get_results (#1451) Update FED-0001 wording (#1464) Move client-side message handler into a separate package (#1462) Documentation for DP-FedAvg wrappers (#1460) Ray resources float (#1453) Slow down mt-pytorch code example, improve README (#1452) Update strategy sequence diagram (#1449) Update FedAvg documentation (#1417) Fix doc generation warnings (#1448) Create experimental mt-pytorch code example (#1446) Change case of file which was missed (#1447) Implement Driver get_clients (#1445) Implement Driver connect/disconnect (#1443) ...
No description provided.