Skip to content

Actions: zhtmike/verl

Actions

e2e_one_step_off_policy

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
e2e_one_step_off_policy.yml
48 workflow runs
48 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

[megatron] feat: Support LoRA training with FP16 using Megatron-Bridg… e2e_one_step_off_policy #48: Commit ec123e6 pushed by zhtmike
8s main
main
8s
[reward] fix: reward model args and reward_kwargs bug (#5289) e2e_one_step_off_policy #47: Commit 92c454a pushed by zhtmike
7s main
main
7s
[reward] fix: backward compatibility with old reward config (#5287) e2e_one_step_off_policy #46: Commit 018f8dc pushed by zhtmike
13s main
main
13s
[fsdp, megatron] refactor: Refactor Fully Async Implementation via En… e2e_one_step_off_policy #45: Commit aa4ca31 pushed by zhtmike
7s main
main
7s
[ci] chore: delete redundant npu ci (#5259) e2e_one_step_off_policy #44: Commit c9dabc7 pushed by zhtmike
8s main
main
8s
[doc] chore: version of dapo_multi_model_optimization_practice (#5263) e2e_one_step_off_policy #43: Commit a9ea5b2 pushed by zhtmike
9s main
main
9s
[algo] feat: reduce routed expert padding via NestedTensor and uint8 … e2e_one_step_off_policy #42: Commit 3e0d4b1 pushed by zhtmike
1m 35s main
main
1m 35s
[megatron] feat: use yaml to manage mbridge args (#4584) e2e_one_step_off_policy #41: Commit 0b0769e pushed by zhtmike
10s main
main
10s
[algo, rollout, sglang] feat: Support router replay with sglang (#4840) e2e_one_step_off_policy #40: Commit e1cd47b pushed by zhtmike
7s main
main
7s
[megatron] fix: Fix error in megatron workers (#4832) e2e_one_step_off_policy #39: Commit b12eb3b pushed by zhtmike
8s main
main
8s
[algo] feat: add optimal token baseline and variance proxy (#4678) e2e_one_step_off_policy #38: Commit 67a127b pushed by zhtmike
9s main
main
9s
[trainer] feat: VeOmniEngine supports qwen3_vl ulysses (#4806) e2e_one_step_off_policy #37: Commit cf92dac pushed by zhtmike
8s main
main
8s
[tool] fix: add tools in single_turn_agent (#4798) e2e_one_step_off_policy #36: Commit 246cac4 pushed by zhtmike
7s main
main
7s
[model] feat: support per sample temperature in trainer (#4787) e2e_one_step_off_policy #35: Commit f9d6198 pushed by zhtmike
7s main
main
7s
[sglang, vllm, rollout] fix: use model's max_position_embeddings for … e2e_one_step_off_policy #34: Commit bb7fe68 pushed by zhtmike
7s main
main
7s
[megatron] fix: Bump Megatron-Bridge commit for PEFT recompute (#4702) e2e_one_step_off_policy #33: Commit a9c7564 pushed by zhtmike
8s main
main
8s
docs: fix typos in code comments and messages (#4724) e2e_one_step_off_policy #32: Commit b63fe15 pushed by zhtmike
10s main
main
10s
[trainer] fix: fallback vision tower to flash_attention_2 for Qwen2.5… e2e_one_step_off_policy #31: Commit cd4072d pushed by zhtmike
11s main
main
11s
[hardware] feat: Auto set device_name to npu for Ascend NPU (#4489) e2e_one_step_off_policy #30: Commit 392791b pushed by zhtmike
9s main
main
9s
[megatron] fix: correct typo in modeling_qwen2_megatron.py (#4486) e2e_one_step_off_policy #29: Commit 01ab536 pushed by zhtmike
9s main
main
9s
[fsdp] fix: reward model also reads override config attn_implementati… e2e_one_step_off_policy #28: Commit d66120d pushed by zhtmike
15s main
main
15s
[trainer] feat: add reward loop config to default config (#4452) e2e_one_step_off_policy #27: Commit 0102d04 pushed by zhtmike
7s main
main
7s
[megatron] fix: respect use_distributed_optimizer in config (#4392) e2e_one_step_off_policy #26: Commit 6d4fd9a pushed by zhtmike
9s main
main
9s
[doc] chore: Update ascend quickstart and docker build guidance doc (… e2e_one_step_off_policy #25: Commit 51f3190 pushed by zhtmike
10s main
main
10s
[algo] feat: support router replay (#4101) e2e_one_step_off_policy #24: Commit cb23607 pushed by zhtmike
14s main
main
14s