Skip to content
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

Add gradient buffers to CollaborativeOptimizer #220

Merged
merged 9 commits into from
Apr 14, 2021
Merged

Conversation

justheuristic
Copy link
Member

@justheuristic justheuristic commented Apr 14, 2021

  • added an option to accumulate gradients locally (default = True, can be disabled to save some memory, not critical in tutorials)
  • fixed a bug in CollaborativeOptimizer that caused it to implicitly scale all updates by self.local_steps_accumulated
  • fixed typos in description

@justheuristic justheuristic requested review from mryab and leshanbog and removed request for mryab April 14, 2021 11:22
hivemind/client/optim/collaborative.py Outdated Show resolved Hide resolved
hivemind/client/optim/collaborative.py Outdated Show resolved Hide resolved
hivemind/client/optim/collaborative.py Show resolved Hide resolved
hivemind/client/optim/collaborative.py Outdated Show resolved Hide resolved
@mryab mryab changed the title Gradient buffers in CollaborativeOptimizer Add gradient buffers to CollaborativeOptimizer Apr 14, 2021
@leshanbog leshanbog merged commit 2359906 into master Apr 14, 2021
@leshanbog leshanbog deleted the optim_use_buffers branch April 14, 2021 12:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants