Skip to content

Actions: verl-project/verl

Actions

e2e_sft_llm_ascend

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
e2e_sft_llm_ascend.yml
858 workflow runs
858 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

[ci] chore: bump the version of vllm-ascend to v0.11.0 in the ascend … e2e_sft_llm_ascend #923: Commit a0c3333 pushed by wuxibin89
In progress main
main
In progress
[rollout] feat: support auto resume on abort in FullyAsyncLLMServerManager e2e_sft_llm_ascend #922: Pull request #5430 synchronize by wuxibin89
In progress wuxibin89:wuxibin/resume_client
wuxibin89:wuxibin/resume_client
In progress
[rollout] feat: support auto resume on abort in FullyAsyncLLMServerManager e2e_sft_llm_ascend #921: Pull request #5430 synchronize by wuxibin89
26m 15s wuxibin89:wuxibin/resume_client
wuxibin89:wuxibin/resume_client
26m 15s
[model] feat: add qwen3.5 fsdp grpo training support with npu optimization e2e_sft_llm_ascend #920: Pull request #5434 opened by Kirrito-k423
Action required Kirrito-k423:qwen3.5_fsdp_npu
Kirrito-k423:qwen3.5_fsdp_npu
Action required
[BREAKING][rollout,cfg] refactor: get rid of actor_rollout_ref config… e2e_sft_llm_ascend #919: Commit de6a1bc pushed by ArronHZG
1h 13m 55s main
main
1h 13m 55s
[megatron, vllm] feat: NVFP4 (W4A16) QAT training support via ModelOpt e2e_sft_llm_ascend #918: Pull request #5254 synchronize by jQizhang
Action required jQizhang:nvfp4_cleanup
jQizhang:nvfp4_cleanup
Action required
[WIP][algo] Migrate and implement the GDPO algorithm into the existing framework. e2e_sft_llm_ascend #917: Pull request #5422 synchronize by Rhetee
Action required Rhetee:GDPO
Rhetee:GDPO
Action required
[WIP][algo] Migrate and implement the GDPO algorithm into the existing framework. e2e_sft_llm_ascend #916: Pull request #5422 synchronize by Rhetee
Action required Rhetee:GDPO
Rhetee:GDPO
Action required
[WIP][algo] Migrate and implement the GDPO algorithm into the existing framework. e2e_sft_llm_ascend #915: Pull request #5422 synchronize by Rhetee
Action required Rhetee:GDPO
Rhetee:GDPO
Action required
[rollout] refactor: bucketed transfer utils e2e_sft_llm_ascend #914: Pull request #5309 synchronize by pengwu22
33m 10s pengwu22:pw/test-wt-utils
pengwu22:pw/test-wt-utils
33m 10s
[rollout] refactor: bucketed transfer utils e2e_sft_llm_ascend #913: Pull request #5309 synchronize by pengwu22
53m 6s pengwu22:pw/test-wt-utils
pengwu22:pw/test-wt-utils
53m 6s
[rollout] feat: support auto resume on abort in FullyAsyncLLMServerManager e2e_sft_llm_ascend #912: Pull request #5430 opened by wuxibin89
32m 49s wuxibin89:wuxibin/resume_client
wuxibin89:wuxibin/resume_client
32m 49s
[doc] add Dr. MAS to awesome work (#5427) e2e_sft_llm_ascend #911: Commit cab2422 pushed by vermouth1992
1h 3m 24s main
main
1h 3m 24s
[misc] feat: Add code for data grouping in no-padding scenario (#5424) e2e_sft_llm_ascend #910: Commit c3fc222 pushed by vermouth1992
32m 18s main
main
32m 18s
[ci] feat: add profiling tests to vLLM ci e2e_sft_llm_ascend #909: Pull request #5215 synchronize by Gary-cjy
Action required Gary-cjy:main
Gary-cjy:main
Action required
[BREAKING][rollout,cfg] refactor: get rid of actor_rollout_ref config from rollout e2e_sft_llm_ascend #908: Pull request #5418 synchronize by wuxibin89
1h 31m 36s wuxibin89:wuxibin/refactor_config
wuxibin89:wuxibin/refactor_config
1h 31m 36s
[tool] refactor: remove tool schema plumbing from SingleTurnAgentLoop… e2e_sft_llm_ascend #907: Commit 48b367b pushed by wuxibin89
1h 10m 58s main
main
1h 10m 58s
[megatron, vllm] feat: NVFP4 (W4A16) QAT training support via ModelOpt e2e_sft_llm_ascend #906: Pull request #5254 synchronize by jQizhang
Action required jQizhang:nvfp4_cleanup
jQizhang:nvfp4_cleanup
Action required
[perf, trtllm] feat: Add Nsight support for rollout server mode (trtllm) e2e_sft_llm_ascend #905: Pull request #5391 synchronize by davidmlw
1h 29m 34s joyang-nv:liweim/nsys
joyang-nv:liweim/nsys
1h 29m 34s
[trainer] feat: Add Nemo-Automodel as alternative training engine e2e_sft_llm_ascend #904: Pull request #5407 synchronize by HuiyingLi
In progress HuiyingLi:add_automodel_sft_backend
HuiyingLi:add_automodel_sft_backend
In progress
[tool, rollout] refactor: cache tool list initialization with MCP timeout e2e_sft_llm_ascend #903: Pull request #5426 opened by denismegerle
Action required denismegerle:public/tool-list-caching
denismegerle:public/tool-list-caching
Action required
[trainer] feat: Add Nemo-Automodel as alternative training engine e2e_sft_llm_ascend #902: Pull request #5407 synchronize by HuiyingLi
Action required HuiyingLi:add_automodel_sft_backend
HuiyingLi:add_automodel_sft_backend
Action required
[trainer] feat: Add Nemo-Automodel as alternative training engine e2e_sft_llm_ascend #901: Pull request #5407 synchronize by HuiyingLi
Action required HuiyingLi:add_automodel_sft_backend
HuiyingLi:add_automodel_sft_backend
Action required
[WIP][algo] Migrate and implement the GDPO algorithm into the existing framework. e2e_sft_llm_ascend #900: Pull request #5422 synchronize by Rhetee
Action required Rhetee:GDPO
Rhetee:GDPO
Action required
[tool] refactor: remove tool schema plumbing from SingleTurnAgentLoop e2e_sft_llm_ascend #899: Pull request #5425 opened by denismegerle
46m 59s denismegerle:public/single-turn-disable-tools
denismegerle:public/single-turn-disable-tools
46m 59s