Skip to content

Actions: NVIDIA/Megatron-LM

Actions

Community Bot

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
community-bot.yml
2,500+ workflow runs
2,500+ workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

Enable AG/RS overlap with explicit process group passing Community Bot #9827: Issue comment #3249 (comment) edited by cspades
16s
16s
fix: wait for async P2P send before deallocating output tensor Community Bot #9826: Issue comment #4047 (comment) created by ZhiyuLi-Nvidia
11s
11s
fix: wait for async P2P send before deallocating output tensor Community Bot #9825: Issue comment #4047 (comment) created by ZhiyuLi-Nvidia
8s
8s
Revert "Multiturn rollout support prep (#2966)" Community Bot #9824: Issue comment #3153 (comment) created by yaoyu-33
11s
11s
Claude: add respond-to-issue skill Community Bot #9823: Issue comment #4141 (comment) created by svcnvidia-nemo-ci
14s
14s
fix: Replace polynomial rolling hash with SHA-256 for prefix caching Community Bot #9822: Issue comment #4158 (comment) created by copy-pr-bot bot
8s
8s
fix(dist_ckpt): Remove step tensor from sharded param state to fix MoE/EP checkpoint validation Community Bot #9821: Issue comment #2895 (comment) created by Phlip79
14s
14s
Fix build_sequences_per_dataset output path arg usage Community Bot #9820: Issue comment #4144 (comment) created by Phlip79
14s
14s
Fix build_sequences_per_dataset output path arg usage Community Bot #9819: Issue comment #4144 (comment) created by Phlip79
16s
16s
add the logic to enable chunked mlp during training Community Bot #9818: Issue comment #3656 (comment) created by Phlip79
15s
15s
Enable AG/RS overlap with explicit process group passing Community Bot #9817: Issue comment #3249 (comment) edited by cspades
12s
12s
Enable AG/RS overlap with explicit process group passing Community Bot #9816: Issue comment #3249 (comment) created by cspades
10s
10s
Enable AG/RS overlap with explicit process group passing Community Bot #9815: Issue comment #3249 (comment) edited by cspades
13s
13s
Enable AG/RS overlap with explicit process group passing Community Bot #9814: Issue comment #3249 (comment) created by cspades
15s
15s
Fix layerwise expt dp size 1 and contention with element wise distributed optimizer Community Bot #9813: Issue comment #4138 (comment) created by skyw
13s
13s
Fix muon getter backward compatability Community Bot #9812: Issue comment #4157 (comment) created by skyw
12s
12s
Fix muon getter backward compatability Community Bot #9811: Issue comment #4157 (comment) created by skyw
13s
13s
Fix muon getter backward compatability Community Bot #9810: Issue comment #4157 (comment) created by copy-pr-bot bot
12s
12s
Enable AG/RS overlap with explicit process group passing Community Bot #9809: Issue comment #3249 (comment) created by jeffnvidia
9s
9s
Applying NVRx backend to both Megatron-FSDP and non-FSDP paths Community Bot #9808: Issue comment #3839 (comment) created by dimapihtar
12s
12s
Applying NVRx backend to both Megatron-FSDP and non-FSDP paths Community Bot #9807: Issue #3839 closed by dimapihtar
9s
9s
Fix RerunStateMachine crash (TypeError: 'NoneType' object is not subscriptable) by not saving a checkpoint after a transient NaN / Inf Community Bot #9806: Issue comment #3981 (comment) created by svcnvidia-nemo-ci
15s
15s
revert: Revert emerging optimizer refactor (#4113, #4119) Community Bot #9805: Issue comment #4156 (comment) created by copy-pr-bot bot
10s
10s
[DEV] fix(megatron-fsdp): preserve non-meta tensors during meta device materialization Community Bot #9804: Issue comment #4155 (comment) created by copy-pr-bot bot
10s
10s
fix(megatron-fsdp): preserve non-meta tensors during meta device materialization Community Bot #9803: Issue comment #4154 (comment) created by copy-pr-bot bot
10s
10s