[feat] Add RLHF rollout integration support (verl)#549

Open

sijyang wants to merge 20 commits into

mainfrom

sijyang/verl_dev

Commits on May 26, 2026

[verl] feat: add trust_remote_code arg and compilation_config dict support

sijyang
authored and
ZhangLirong-amd
committed
[verl] feat: add logprobs and request_id support across sampling pipeline

sijyang
authored and
ZhangLirong-amd
committed
[verl] feat: weight sync, memory lifecycle and DP isolation for verl integration (TP+DP)

sijyang
authored and
ZhangLirong-amd
committed
[verl] feat: utility command dispatch and broadcast communication

sijyang
authored and
ZhangLirong-amd
committed
[verl] feat: basic integration with verl - load_weights, sleep/wake_up API

sijyang
authored and
ZhangLirong-amd
committed
[atom] fix: rope parameters handling, remove CLI trust_remote_code, and minor fixes

sijyang
authored and
ZhangLirong-amd
committed
[atom] feat: implement packed weight handling in ModelRunner for FP8 parameters

sijyang
authored and
ZhangLirong-amd
committed
[verl] refactor: decouple RLHF rollout logic from inference engine into atom/rollout/

sijyang
authored and
ZhangLirong-amd
committed
[verl] feat: extend tokenIDProcessor for logprobs support and enhance ModelRunner with DP isolation handling

sijyang
authored and
ZhangLirong-amd
committed
fix: patch NCCL device binding for DP-isolated ModelRunner

sijyang
authored and
ZhangLirong-amd
committed
refactor: minimize diff against main by reverting non-functional changes

sijyang
authored and
ZhangLirong-amd
committed
refactor: improve code readability by formatting and organizing function parameters and comments across multiple files

sijyang
authored and
ZhangLirong-amd
committed
refactor: extract sleep logic from engine_core busy_loop into helper methods

sijyang
authored and
ZhangLirong-amd
committed
[verl] refactor: merge logprobs and DP isolation into base ModelRunner, simplify RLHFModelRunner

sijyang
authored and
ZhangLirong-amd
committed
refactor: rename sleep state variables and update related logic for RL training in EngineCore and ModelRunner

sijyang
authored and
ZhangLirong-amd
committed
fix: restore mark_trace profiler around cudagraph capture

sijyang
authored and
ZhangLirong-amd
committed
docs: add veRL + Megatron + ATOM environment setup guide for ROCm

sijyang
authored and
ZhangLirong-amd
committed
[verl] feat: add logprobs and request_id support across sampling pipeline

sijyang
authored and
ZhangLirong-amd
committed
[verl] refactor: unify load_weights API with auto mode selection

sijyang
authored and
ZhangLirong-amd
committed
fix: batch token ID processing in tokenIDProcessor

sijyang
authored and
ZhangLirong-amd
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[feat] Add RLHF rollout integration support (verl)#549

[feat] Add RLHF rollout integration support (verl)#549
sijyang wants to merge 20 commits into
mainfrom
sijyang/verl_dev

Commits on May 26, 2026

[verl] feat: add trust_remote_code arg and compilation_config dict support

[verl] feat: add logprobs and request_id support across sampling pipeline

[verl] feat: weight sync, memory lifecycle and DP isolation for verl integration (TP+DP)

[verl] feat: utility command dispatch and broadcast communication

[verl] feat: basic integration with verl - load_weights, sleep/wake_up API

[atom] fix: rope parameters handling, remove CLI trust_remote_code, and minor fixes

[atom] feat: implement packed weight handling in ModelRunner for FP8 parameters

[verl] refactor: decouple RLHF rollout logic from inference engine into atom/rollout/

[verl] feat: extend tokenIDProcessor for logprobs support and enhance ModelRunner with DP isolation handling

fix: patch NCCL device binding for DP-isolated ModelRunner

refactor: minimize diff against main by reverting non-functional changes

refactor: improve code readability by formatting and organizing function parameters and comments across multiple files

refactor: extract sleep logic from engine_core busy_loop into helper methods

[verl] refactor: merge logprobs and DP isolation into base ModelRunner, simplify RLHFModelRunner

refactor: rename sleep state variables and update related logic for RL training in EngineCore and ModelRunner

fix: restore mark_trace profiler around cudagraph capture

docs: add veRL + Megatron + ATOM environment setup guide for ROCm

[verl] feat: add logprobs and request_id support across sampling pipeline

[verl] refactor: unify load_weights API with auto mode selection

fix: batch token ID processing in tokenIDProcessor