Skip to content

Actions: yellowbee686/verl

Actions

e2e_ascend

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
e2e_ascend.yml
116 workflow runs
116 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

[data] feat: passing tool_config to data (#3950) e2e_ascend #116: Commit 11f536f pushed by yellowbee686
1s main
main
1s
[doc] feat: add a doc for vllm+megatron training (#3974) e2e_ascend #115: Commit 2769986 pushed by yellowbee686
1s main
main
1s
[ci, recipe] fix: add the missing key compute_prox_log_prob to fully_… e2e_ascend #114: Commit 0bd0345 pushed by yellowbee686
1s main
main
1s
[trainer, recipe] feat: Fully Async Policy add Rollout Importance Sa… e2e_ascend #113: Commit e3b77a6 pushed by yellowbee686
1s main
main
1s
[ci] chore: Rename dockerfile and update e2e_ascend (#3966) e2e_ascend #112: Commit c9910c2 pushed by yellowbee686
1s main
main
1s
[data] fix: Pass video metadata to vLLM and support change image_patc… e2e_ascend #111: Commit dabbec2 pushed by yellowbee686
1s main
main
1s
[rollout, vllm] fix: handle lora request when base_sync_done is false… e2e_ascend #110: Commit 4576678 pushed by yellowbee686
1s main
main
1s
[algo] refactor: Rollout Importance Sampling - Separate IS Weights f… e2e_ascend #109: Commit 4424616 pushed by yellowbee686
1s main
main
1s
[recipe] fix: DAPO using KL in reward (#3916) e2e_ascend #108: Commit ede8cdd pushed by yellowbee686
1s main
main
1s
[recipe] fix: fix bug of tranfer queue runtime env (#3904) e2e_ascend #107: Commit ecdaa8d pushed by yellowbee686
1s main
main
1s
[megatron] fix: VLMs using fused kernels (#3849) e2e_ascend #106: Commit ab676dc pushed by yellowbee686
1s main
main
1s
[perf, data] feat: DP workload balance (#3605) e2e_ascend #105: Commit 62221fa pushed by yellowbee686
1s main
main
1s
[fsdp] fix : fix moe model run on full-async error (#3874) e2e_ascend #104: Commit 827f0f5 pushed by yellowbee686
1s main
main
1s
[rollout, vllm] fix: name change for compilation level (#3848) e2e_ascend #103: Commit 1c1d53d pushed by yellowbee686
1s main
main
1s
[misc] feat: add megatron script for open math reasoning (#3844) e2e_ascend #102: Commit 981d781 pushed by yellowbee686
2s main
main
2s
[worker] fix: to create a new event loop if none exists when building… e2e_ascend #101: Commit f6a1c55 pushed by yellowbee686
1s main
main
1s
[algo] fix: remove torch.quantile-based percentile metrics to resolve… e2e_ascend #100: Commit 4f1c489 pushed by yellowbee686
1s main
main
1s
[ci] fix: Install mlflow dependency (#3817) e2e_ascend #99: Commit f209c6f pushed by yellowbee686
1s main
main
1s
[misc] fix: Sanitize MLFlow metric names (#3736) e2e_ascend #98: Commit 4da0d3d pushed by yellowbee686
2s main
main
2s
[megatron] feat: script of qwen3vl 235b (#3799) e2e_ascend #97: Commit dd8864f pushed by yellowbee686
2s main
main
2s
[trainer] feat: ReMax support using reward model for baseline (#3780) e2e_ascend #96: Commit ae5d850 pushed by yellowbee686
1s main
main
1s
[trainer] fix: batch size mismatch with n>1 when gen_max for ReMax (#… e2e_ascend #95: Commit a80ed95 pushed by yellowbee686
Skipped main
main
Skipped
[doc] fix: actor_rollout_ref.critic is not correct (#3778) e2e_ascend #94: Commit acfcf98 pushed by yellowbee686
1s main
main
1s
[recipe] feat: Add example for gpt-oss training using agent loop (#3774) e2e_ascend #93: Commit 0615352 pushed by yellowbee686
Skipped main
main
Skipped
[trainer] feat: set interleave to False in dapo trainer (#3760) e2e_ascend #92: Commit 67f9a21 pushed by yellowbee686
1s main
main
1s