Add Remote Reward Server Feature #329
Open
Conversation
This file contains 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
There was a problem hiding this comment.
Why do we have an empty file here?
There was a problem hiding this comment.
I have resolved the empty file problem
examples/trainer.experiment_name=
Outdated
There was a problem hiding this comment.
Could you remove this file?
run_ppo.sh
Outdated
There was a problem hiding this comment.
Remove this file as well
PeterSH6
requested changes
Feb 24, 2025
| @@ -0,0 +1,163 @@ | |||
| import re | |||
There was a problem hiding this comment.
Please add a license similar to other files. Thanks!
| @@ -0,0 +1,89 @@ | |||
| import asyncio | |||
There was a problem hiding this comment.
Also add a license similar to other files
|
|
|
CLA signed :) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Description
This PR introduces support for remote generative models to enhance the verification and reward assignment process in VERL. Previously, VERL only supported rule-based rewards for verification. With this update, we enable more flexible and dynamic reward mechanisms using generative models.
Key Changes
Bug Fix:
main_ppo.pywhere theval_dataloaderwas incorrectly processed in batches. This ensures proper handling of validation data during training.New Features:
workers/reward_managerto handle reward generation using remote generative models.compute_scorefunction inutils/reward_scoreto calculate scores based on generative model outputs.Issue Resolution: