why are there multiple settings for actor_rollout_ref.model.enable_gradient_checkpointing? Is this a deliberate design choice?
#5940
This workflow is awaiting approval from a maintainer in #4263
Triggered via pull request
February 26, 2026 10:31
Status
Action required
Total duration
–
Artifacts
–
This workflow is awaiting approval from a maintainer in #4263
e2e_one_step_off_policy.yml
on: pull_request
setup
e2e_one_step_off_policy_fsdp2
e2e_one_step_off_policy_megatron
cleanup