Skip to content

Conversation

wuxibin89
Copy link
Collaborator

No description provided.

Copy link
Collaborator

@PeterSH6 PeterSH6 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The multi-turn and tool-calling examples LGTM.
However, for the single-turn generation, I think we could simplify the workflow without using the ChatScheduler.
But we can merge this PR in advance.

I believe we should set up a meeting to discuss a better "ChatScheduler" to generalize single-turn/multi-turn/tool calling scenarios:

  1. General for all kinds of rollout backends
  2. A suitable scheduler for single-turn/multi-turn and can handle per-request callback
  3. Can be called by any environments

@PeterSH6 PeterSH6 merged commit 1e47e41 into main May 4, 2025
25 checks passed
@PeterSH6 PeterSH6 deleted the wuxibin/async_streaming_generator branch May 4, 2025 11:07
ScottCTD pushed a commit to ScottCTD/verl that referenced this pull request May 5, 2025
…l paper (volcengine#1297)

Co-authored-by: shengguangming <shengguangming@bytedance.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants