Add ray-based distributed reward manager #1206
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
|
rawsh
force-pushed
the
ray_verify
branch
2 times, most recently
from
0254cc8 to
9c68bc5
Compare
rawsh
changed the title
rawsh
changed the title
rawsh
force-pushed
the
ray_verify
branch
3 times, most recently
from
227ac7a to
261b591
Compare
Author
|
apologies for the update noise, cleaned it up a good bit |
|
Hello. I've achieved something similar to what you have. Well, I've found that when the Ray Cluster is started, only the master node uses all cpu cores for calculation, while other worker nodes don't. I'd like to ask, should we transform this function into the form of a working group? |
Author
|
I think this approach is cleaner #1693 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Add ray reward manager to support cross-node parallelization for verification
Tested with math-verify, dropped verification time to ~3s consistently for 128 x 16 = 2048 generations and ~30s for 256 x 64 = 16384 generations, with naive reward manager taking 200+ seconds