Skip to content

Actions: zhtmike/verl

Actions

e2e_ascend

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
e2e_ascend.yml
56 workflow runs
56 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_ascend #56: Commit ec123e6 pushed by zhtmike
2s main
main
2s
[reward] fix: reward model args and reward_kwargs bug (#5289) e2e_ascend #55: Commit 92c454a pushed by zhtmike
1s main
main
1s
[reward] fix: backward compatibility with old reward config (#5287) e2e_ascend #54: Commit 018f8dc pushed by zhtmike
1s main
main
1s
[fsdp, megatron] refactor: Refactor Fully Async Implementation via En… e2e_ascend #53: Commit aa4ca31 pushed by zhtmike
2s main
main
2s
[ci] chore: delete redundant npu ci (#5259) e2e_ascend #52: Commit c9dabc7 pushed by zhtmike
1s main
main
1s
[doc] chore: version of dapo_multi_model_optimization_practice (#5263) e2e_ascend #51: Commit a9ea5b2 pushed by zhtmike
2s main
main
2s
[algo] feat: reduce routed expert padding via NestedTensor and uint8 … e2e_ascend #50: Commit 3e0d4b1 pushed by zhtmike
1s main
main
1s
[megatron] feat: use yaml to manage mbridge args (#4584) e2e_ascend #49: Commit 0b0769e pushed by zhtmike
1s main
main
1s
[ci,doc] feat: Add ascend_ci_guide (#5163) e2e_ascend #48: Commit de9880d pushed by zhtmike
2s main
main
2s
Revert "[worker,rollout] refactor: remove set_expandable_segments cal… e2e_ascend #47: Commit d9939ad pushed by zhtmike
1s main
main
1s
[reward] fix: fix reward computation in _validate when use_reward_lo… e2e_ascend #46: Commit 7102a9a pushed by zhtmike
2s main
main
2s
[trainer] fix: resolve dataset config in agent loop (#5034) e2e_ascend #45: Commit b63fe7f pushed by zhtmike
2s main
main
2s
[veomni] feat: support model resharding between veomni and rollout en… e2e_ascend #44: Commit b53f0f1 pushed by zhtmike
2s main
main
2s
[algo, rollout, sglang] feat: Support router replay with sglang (#4840) e2e_ascend #43: Commit e1cd47b pushed by zhtmike
1s main
main
1s
[megatron] fix: Fix error in megatron workers (#4832) e2e_ascend #42: Commit b12eb3b pushed by zhtmike
1s main
main
1s
[algo] feat: add optimal token baseline and variance proxy (#4678) e2e_ascend #41: Commit 67a127b pushed by zhtmike
2s main
main
2s
[trainer] feat: VeOmniEngine supports qwen3_vl ulysses (#4806) e2e_ascend #40: Commit cf92dac pushed by zhtmike
1s main
main
1s
[tool] fix: add tools in single_turn_agent (#4798) e2e_ascend #39: Commit 246cac4 pushed by zhtmike
1s main
main
1s
[model] feat: support per sample temperature in trainer (#4787) e2e_ascend #38: Commit f9d6198 pushed by zhtmike
1s main
main
1s
[doc] feat: add v0.7 release blog (#4796) e2e_ascend #37: Commit f9c855f pushed by zhtmike
1s main
main
1s
[doc] fix: reward_loop enable flag name (#4788) e2e_ascend #36: Commit 01069ec pushed by zhtmike
1s main
main
1s
[sglang, vllm, rollout] fix: use model's max_position_embeddings for … e2e_ascend #35: Commit bb7fe68 pushed by zhtmike
1s main
main
1s
[megatron] fix: Bump Megatron-Bridge commit for PEFT recompute (#4702) e2e_ascend #34: Commit a9c7564 pushed by zhtmike
1s main
main
1s
docs: fix typos in code comments and messages (#4724) e2e_ascend #33: Commit b63fe15 pushed by zhtmike
2s main
main
2s
[trainer] fix: fallback vision tower to flash_attention_2 for Qwen2.5… e2e_ascend #32: Commit cd4072d pushed by zhtmike
1s main
main
1s