Skip to content

Actions: meituan-search/verl

Actions

e2e_ascend

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
e2e_ascend.yml
248 workflow runs
248 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

[vllm] fix: npu disable flash_attn for RotaryEmbedding (#5640) e2e_ascend #248: Commit b2c489c pushed by sl-1314
1s main
main
1s
[rollout] fix: fix some compatibility issue with qwen vl seris suppo… e2e_ascend #247: Commit efc48d5 pushed by ArronHZG
1s main
main
1s
[ci] chore: change machine for npu ci (#5578) e2e_ascend #246: Commit 5e2f2b2 pushed by ArronHZG
1s main
main
1s
[megatron, model] feat: qwen3.5 example (#5381) e2e_ascend #245: Commit ef072ac pushed by ArronHZG
1s main
main
1s
[doc] chore: add note for kimi ckpt engine (#5546) e2e_ascend #244: Commit 3bbdece pushed by ArronHZG
2s main
main
2s
[misc] feat: support dynamic bsz using group size (#5438) e2e_ascend #243: Commit 80eb57e pushed by sl-1314
2s main
main
2s
[Megatron] feat: Support routing replay on NPU with performance and c… e2e_ascend #242: Commit 257d4fa pushed by ArronHZG
2s main
main
2s
[doc] chore: update FP8 guide with E2E training section and reorganiz… e2e_ascend #241: Commit e25992b pushed by sl-1314
1s main
main
1s
[doc] feat: add mtp spec log (#5491) e2e_ascend #240: Commit d4399a6 pushed by ArronHZG
1s main
main
1s
[rollout] feat: support auto resume on abort in FullyAsyncLLMServerMa… e2e_ascend #239: Commit c4593e3 pushed by ArronHZG
1s main
main
1s
[rollout] feat: update trtllm docker (#5386) e2e_ascend #238: Commit ac4ec1b pushed by ArronHZG
1s main
main
1s
[doc] chore: fix npu docs (#5428) e2e_ascend #237: Commit c179476 pushed by AniZpZ
1s main
main
1s
[trainer] feat: add padding for tensor alignment in preprocess_thd_no… e2e_ascend #236: Commit 32705dc pushed by sl-1314
1s main
main
1s
[doc] chore: ascend add performance analysis guide and update some ve… e2e_ascend #235: Commit 712de01 pushed by sl-1314
1s main
main
1s
[misc] fix: pass torch dtype when init random model (#5370) e2e_ascend #234: Commit f56c893 pushed by sl-1314
1s main
main
1s
[BREAKING][reward] refactor: deprecate batch reward manager (#5237) e2e_ascend #233: Commit 3df21a6 pushed by AniZpZ
1s main
main
1s
[perf] fix: fix npu profiling scripts (#5226) e2e_ascend #232: Commit 7cf31bf pushed by AniZpZ
1s main
main
1s
refactor params async e2e_ascend #231: Pull request #45 synchronize by ArronHZG
1s separation_update_weights
separation_update_weights
1s
refactor params async e2e_ascend #230: Pull request #45 synchronize by ArronHZG
1s separation_update_weights
separation_update_weights
1s
refactor params async e2e_ascend #229: Pull request #45 synchronize by ArronHZG
2s separation_update_weights
separation_update_weights
2s
[recipe] refactor: refactor ray trainer for separate recipe use. (ful… e2e_ascend #228: Commit 32f2d3d pushed by ArronHZG
1s main
main
1s
feat: support export fp8 weights e2e_ascend #227: Pull request #42 synchronize by eternally-z
2s feature/export_fp8_weights
feature/export_fp8_weights
2s
feat: support export fp8 weights e2e_ascend #226: Pull request #42 synchronize by eternally-z
1s feature/export_fp8_weights
feature/export_fp8_weights
1s
[megatron] fix: add protections for logits_processor_args.pop("loss_m… e2e_ascend #225: Commit ad404cb pushed by ArronHZG
1s main
main
1s
[worker, training_utils] fix: Metric Aggregation Across DP Ranks (#5203) e2e_ascend #224: Commit 2320603 pushed by AniZpZ
Skipped main
main
Skipped