Skip to content

feat(cs): add async multi-turn agentic rollout #361

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.

Sign up for GitHub

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

Draft
Yunnglin wants to merge 17 commits into
base: dev-swift-v5
Choose a base branch
from
Draft

feat(cs): add async multi-turn agentic rollout #361

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
17 commits
Select commit Hold shift + click to select a range
514ae2f
fix(cs): initialize score heads and correct training metric reporting
Yunnglin Oct 30, 2026
82a3ee4
feat(cs): add async multi-turn agentic rollout
Yunnglin Oct 31, 2026
5e7c47a
Merge remote-tracking branch 'origin/dev-swift-v5' into cs-async-rl-f…
Yunnglin Nov 2, 2026
fe59fd6
fix(cs): drain interrupted rollouts and harden agentic validation
Yunnglin Nov 2, 2026
aeb73da
docs(client): consolidate async GRPO usage in user guides
Yunnglin Nov 2, 2026
6f6a480
test(agentic): add scoped OpenEnv functional acceptance
Yunnglin Nov 2, 2026
16a1bc4
fix(model): select only the exact LoRA slot for optimizer groups
Yunnglin Nov 3, 2026
38952e9
fix(agentic): bind complete environment tool schemas
Yunnglin Nov 3, 2026
89138e2
test(agentic): clarify virtual workspace feedback and guard parity ro…
Yunnglin Nov 3, 2026
4d1a8c6
test(agentic): accept confirmed OpenEnv interpreter failures
Yunnglin Nov 3, 2026
88836c9
test(agentic): probe full rollout groups on data parallel models
Yunnglin Nov 4, 2026
447daf2
fix(agentic): preserve valid multi-turn training rollouts
Yunnglin Nov 5, 2026
b9514cf
Merge remote-tracking branch 'origin/dev-swift-v5' into cs-async-rl-f…
Yunnglin Nov 5, 2026
d822a94
fix(processor): preserve absent multimodal fields
Yunnglin Nov 5, 2026
d8521d5
test(agentic): record rollout supervision evidence
Yunnglin Nov 5, 2026
9d918c0
fix(model): synchronize gradients on completed accumulation steps
Yunnglin Nov 5, 2026
17e05cc
docs: move async multi-lora GRPO usage into guides
Yunnglin Nov 6, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
156 changes: 0 additions & 156 deletions cookbook/client/async_rl/README.md
View file
Open in desktop

This file was deleted.

Loading
Loading