Skip to content

Actions: verl-project/verl

Actions

e2e_transferqueue

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
e2e_transferqueue.yml
2,395 workflow runs
2,395 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

[megatron] feat: enhance model offloading and loading for frozen parameters e2e_transferqueue #3134: Pull request #5412 synchronize by RobotGF
32m 26s RobotGF:fix_lora_offload
RobotGF:fix_lora_offload
32m 26s
[ckpt] feat: implement large tensor slicing in vllm rollout and CheckpointEngine for weight updating e2e_transferqueue #3133: Pull request #5378 synchronize by jianjunzhong
21m 40s jianjunzhong:feat/chunked_weight_update
jianjunzhong:feat/chunked_weight_update
21m 40s
[algo] fix: seq mean and default scale factor loss_mask.shape[-1] as in seq-mean-token-sum-norm e2e_transferqueue #3132: Pull request #5417 opened by tongyx361
33m 39s tongyx361:tyx/fix/seq-mean-in-seq-mean-token-sum-norm
tongyx361:tyx/fix/seq-mean-in-seq-mean-token-sum-norm
33m 39s
[tool] feature: scheduling analysis based on profiling data for torch profiler e2e_transferqueue #3131: Pull request #5367 synchronize by Rhetee
Action required Rhetee:main
Rhetee:main
Action required
[perf, trtllm] feat: Add Nsight support for rollout server mode (trtllm) e2e_transferqueue #3130: Pull request #5391 synchronize by davidmlw
1h 4m 58s joyang-nv:liweim/nsys
joyang-nv:liweim/nsys
1h 4m 58s
[Megatron] feat: Support routing replay on NPU with performance and compatibility enhancements e2e_transferqueue #3129: Pull request #5298 synchronize by 755651978
36m 25s 755651978:main-0212
755651978:main-0212
36m 25s
why are there multiple settings for actor_rollout_ref.model.enable_gradient_checkpointing? Is this a deliberate design choice? e2e_transferqueue #3128: Pull request #4263 synchronize by khazic
Action required khazic:main
khazic:main
Action required
[algo] feat: add DPPO with binary TV or binary KL implementation (#5397) e2e_transferqueue #3127: Commit 182383b pushed by tongyx361
1h 18m 17s main
main
1h 18m 17s
[Megatron] feat: Support routing replay on NPU with performance and compatibility enhancements e2e_transferqueue #3126: Pull request #5298 synchronize by 755651978
47m 19s 755651978:main-0212
755651978:main-0212
47m 19s
[Megatron] feat: Support routing replay on NPU with performance and compatibility enhancements e2e_transferqueue #3124: Pull request #5298 synchronize by 755651978
12m 6s 755651978:main-0212
755651978:main-0212
12m 6s
why are there multiple settings for actor_rollout_ref.model.enable_gradient_checkpointing? Is this a deliberate design choice? e2e_transferqueue #3123: Pull request #4263 synchronize by khazic
Action required khazic:main
khazic:main
Action required
[fsdp,algo] feat: Support QAT (NVFP4) in FSDPEngine for the unified engine_workers architecture e2e_transferqueue #3122: Pull request #5411 synchronize by zhangyimi
Action required zhangyimi:qat-core-v2
zhangyimi:qat-core-v2
Action required
[fsdp,algo] feat: Support QAT (NVFP4) in FSDPEngine for the unified engine_workers architecture e2e_transferqueue #3121: Pull request #5411 synchronize by zhangyimi
Action required zhangyimi:qat-core-v2
zhangyimi:qat-core-v2
Action required
[fsdp,algo] feat: Support QAT (NVFP4) in FSDPEngine for the unified engine_workers architecture e2e_transferqueue #3120: Pull request #5411 opened by zhangyimi
Action required zhangyimi:qat-core-v2
zhangyimi:qat-core-v2
Action required
[misc,trainer,rollout] feat: add Prometheus metrics logging to experiment tracking e2e_transferqueue #3118: Pull request #5291 synchronize by guillemgt
Action required guillemgt:guillem.tarrach/upstream-prometheus-metrics
guillemgt:guillem.tarrach/upstream-prometheus-metrics
Action required
[misc,trainer,rollout] feat: add Prometheus metrics logging to experiment tracking e2e_transferqueue #3117: Pull request #5291 synchronize by guillemgt
Action required guillemgt:guillem.tarrach/upstream-prometheus-metrics
guillemgt:guillem.tarrach/upstream-prometheus-metrics
Action required
[trainer] feat: add support for the GDPO algorithm e2e_transferqueue #3116: Pull request #5409 opened by yue-zeng-yue
Action required yue-zeng-yue:feat-gdpo
yue-zeng-yue:feat-gdpo
Action required
[trainer,misc] fix: fix multiple bugs in fully async trainings e2e_transferqueue #3115: Pull request #5373 synchronize by guillemgt
Action required guillemgt:upstream/fix-fully-async-bugs
guillemgt:upstream/fix-fully-async-bugs
Action required
[trainer] feat: Add Nemo-Automodel as alternative training engine e2e_transferqueue #3114: Pull request #5407 opened by HuiyingLi
1h 42m 21s HuiyingLi:add_automodel_sft_backend
HuiyingLi:add_automodel_sft_backend
1h 42m 21s
why are there multiple settings for actor_rollout_ref.model.enable_gradient_checkpointing? Is this a deliberate design choice? e2e_transferqueue #3113: Pull request #4263 synchronize by khazic
Action required khazic:main
khazic:main
Action required
why are there multiple settings for actor_rollout_ref.model.enable_gradient_checkpointing? Is this a deliberate design choice? e2e_transferqueue #3112: Pull request #4263 synchronize by khazic
Action required khazic:main
khazic:main
Action required
[perf, trtllm] feat: Add Nsight support for rollout server mode (trtllm) e2e_transferqueue #3111: Pull request #5391 synchronize by davidmlw
41m 20s joyang-nv:liweim/nsys
joyang-nv:liweim/nsys
41m 20s
[ckpt] feat: implement large tensor slicing in vllm rollout and CheckpointEngine for weight updating e2e_transferqueue #3110: Pull request #5378 synchronize by jianjunzhong
1h 30m 43s jianjunzhong:feat/chunked_weight_update
jianjunzhong:feat/chunked_weight_update
1h 30m 43s
[megatron] fix: missing model offload to CPU for forward_only mode e2e_transferqueue #3109: Pull request #5406 opened by xhx1022
7m 18s xhx1022:xhx/offload
xhx1022:xhx/offload
7m 18s
[ckpt] feat: implement large tensor slicing in vllm rollout and CheckpointEngine for weight updating e2e_transferqueue #3108: Pull request #5378 synchronize by jianjunzhong
Action required jianjunzhong:feat/chunked_weight_update
jianjunzhong:feat/chunked_weight_update
Action required