Skip to content

Actions: AlignmentResearch/megatron-lm-contrib

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
42 workflow runs
42 workflow runs

Filter by Workflow

⌥ + click/return to exclude

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
fix(ssm): refuse packed sequences without seq_idx in the Mamba training path fork-base #13: Pull request #7 opened by SamAdamDay
15s sam/assert-packed-seq-idx
sam/assert-packed-seq-idx
15s
Merge pull request #5 from AlignmentResearch/tigist/fix-hybrid-optimi… fork-base #12: Commit 834e071 pushed by tigist-far
11s farai/main
farai/main
11s
[Optimizer] HybridDeviceOptimizer resumes from a checkpoint without a second set of pinned copies, and exactly fork-base #11: Pull request #5 synchronize by tigist-far
12s tigist/fix-hybrid-optimizer-load-reuses-cpu-copies
tigist/fix-hybrid-optimizer-load-reuses-cpu-copies
12s
[CI] fork-base follows a rename of this repository fork-base #10: Pull request #6 opened by tigist-far
15s tigist/fork-base-allows-a-repository-rename
tigist/fork-base-allows-a-repository-rename
15s
[Optimizer] HybridDeviceOptimizer resumes from a checkpoint without a second set of pinned copies, and exactly fork-base #9: Pull request #5 synchronize by tigist-far
13s tigist/fix-hybrid-optimizer-load-reuses-cpu-copies
tigist/fix-hybrid-optimizer-load-reuses-cpu-copies
13s
[Optimizer] HybridDeviceOptimizer resumes from a checkpoint without a second set of pinned copies, and exactly fork-base #8: Pull request #5 synchronize by tigist-far
21s tigist/fix-hybrid-optimizer-load-reuses-cpu-copies
tigist/fix-hybrid-optimizer-load-reuses-cpu-copies
21s
[Optimizer] HybridDeviceOptimizer resumes from a checkpoint without a second set of pinned copies, and exactly fork-base #7: Pull request #5 synchronize by tigist-far
13s tigist/fix-hybrid-optimizer-load-reuses-cpu-copies
tigist/fix-hybrid-optimizer-load-reuses-cpu-copies
13s
[Optimizer] HybridDeviceOptimizer resumes from a checkpoint without a second set of pinned copies, and exactly fork-base #6: Pull request #5 opened by tigist-far
16s tigist/fix-hybrid-optimizer-load-reuses-cpu-copies
tigist/fix-hybrid-optimizer-load-reuses-cpu-copies
16s
Merge pull request #3 from AlignmentResearch/tf-at/torch-gmm-full-rec… fork-base #5: Commit 7f47384 pushed by taufeeque9
15s farai/main
farai/main
15s
perf(moe): add frozen grouped GEMM for adapter training fork-base #4: Pull request #3 synchronize by taufeeque9
15s tf-at/torch-gmm-full-recompute
tf-at/torch-gmm-full-recompute
15s
Graph Update: pip in /docker/lts, /examples/post_training/modelopt, /megatron/core, /megatron/core/distributed/fsdp/src #1544878844 Dependency Graph #2: by dependabot Bot
1m 2s farai/main
farai/main
1m 2s
Graph Update: uv in /. #1544878843 Dependency Graph #1: by dependabot Bot
50s farai/main
farai/main
50s
chore(fork): prune NVIDIA-only CI and record the upstream base Community Bot #3: Issue comment #2 (comment) created by taufeeque9
8s
8s
chore(fork): prune NVIDIA-only CI and record the upstream base Request NVSkills CI #8: Issue comment #2 (comment) created by taufeeque9
1s
1s
chore(fork): prune NVIDIA-only CI and record the upstream base Claude Fix PR #3: Issue comment #2 (comment) created by taufeeque9
1s
1s
chore(fork): prune NVIDIA-only CI and record the upstream base Community Request Assignee #3: Issue comment #2 (comment) created by taufeeque9
1s
1s
chore(fork): prune NVIDIA-only CI and record the upstream base Claude Code Review #3: Issue comment #2 (comment) created by taufeeque9
1s
1s
chore(fork): prune NVIDIA-only CI and record the upstream base Claude Copy PR to Main #3: Issue comment #2 (comment) created by taufeeque9
10s
10s
Merge pull request #4 from AlignmentResearch/tf-c/fork-base-scaffoldi… fork-base #3: Commit e92d99e pushed by taufeeque9
12s farai/main
farai/main
12s
chore(fork): prune NVIDIA-only CI and record upstream base Claude Complexity Label #2: Pull request #4 ready_for_review by taufeeque9
21s
21s
chore(fork): prune NVIDIA-only CI and record upstream base fork-base #2: Pull request #4 opened by taufeeque9
16s tf-c/fork-base-scaffolding-reviewed
tf-c/fork-base-scaffolding-reviewed
16s
Auto Swap Labels Auto Swap Labels #1: completed by SamAdamDay
2s
2s
perf(moe): add frozen grouped GEMM for adapter training Review Trigger #1: Pull request #3 submitted by SamAdamDay
1s
1s
perf(moe): add frozen grouped GEMM for adapter training Claude Complexity Label #1: Pull request #3 ready_for_review by taufeeque9
35s
35s
fix(moe): reject TE op fuser with torch backend Request NVSkills CI #7: Commit 8447c0d pushed by taufeeque9
1s tf-at/torch-gmm-full-recompute
tf-at/torch-gmm-full-recompute
1s