traceopt-ai

Issues somebody may already be working on are listed last. List everything by date

• Show a TraceML Training Report in W&B and MLflow

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 5/5 Over a week Newbie friendliness 25/100

  traceopt-ai/traceml#620 · 1 comment · Opened Nov 12, 2026

  Maintainers usually reply within 1 day

• Score rank stragglers before averaging across steps

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  traceopt-ai/traceml#617 · Opened Nov 12, 2026

  Maintainers usually reply within 1 day

• Add a Weights & Biases telemetry exporter

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 5/5 Over a week Newbie friendliness 35/100

  traceopt-ai/traceml#614 · 1 comment · Opened Nov 11, 2026

  Maintainers usually reply within 1 day

• [Bug]: code_manifest.json reports a PyTorch Lightning script as using the Hugging Face Trainer

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 2/5 1-3 hours Newbie friendliness 78/100

  traceopt-ai/traceml#612 · Opened Nov 11, 2026

  Maintainers usually reply within 1 day

• A loop with no input pipeline never gets a Step Time verdict

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 5/5 Over a week Newbie friendliness 22/100

  traceopt-ai/traceml#581 · Opened Nov 2, 2026

  Maintainers usually reply within 1 day

• `--enable-logging` raw logs land outside the run's folder and lose their tail

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 3/5 1-2 days Newbie friendliness 55/100

  traceopt-ai/traceml#580 · Opened Nov 2, 2026

  Maintainers usually reply within 1 day

• Aggregator keeps running after its `traceml run` launcher is killed

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 58/100

  traceopt-ai/traceml#573 · Opened Oct 27, 2026

  Maintainers usually reply within 1 day

• Step-time summary: first (compile) step skews the averages, and runs without a DataLoader fall back to CPU-clock phase times

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 5/5 Over a week Newbie friendliness 15/100

  traceopt-ai/traceml#552 · 2 comments · Opened Oct 9, 2026

  Maintainers usually reply within 1 day

• System cluster view averages an unreported GPU into the node mean

  May be free again @giovannicozzolongo claimed this 66 days ago, and no pull request is open. Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  traceopt-ai/traceml#440 · 1 comment · 1 assignee · Opened Sep 2, 2026

  Maintainers usually reply within 1 day

• Export aggregator telemetry as structured OTLP Logs

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  enhancement Integration P1 runtime

  Difficulty 5/5 Over a week Newbie friendliness 35/100

  traceopt-ai/traceml#396 · Opened Aug 25, 2026

  Maintainers usually reply within 1 day

• Generate diagnosis notes from canonical evidence

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 42/100

  traceopt-ai/traceml#383 · Opened Aug 14, 2026

  Maintainers usually reply within 1 day

• Unify live trend analysis and dashboard inputs

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  traceopt-ai/traceml#382 · Opened Aug 14, 2026

  Maintainers usually reply within 1 day

• Add a metric-honest global-rank × aligned-step heatmap

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  dashboard distributed enhancement P2

  Difficulty 4/5 3-5 days Newbie friendliness 42/100

  traceopt-ai/traceml#375 · Opened Aug 10, 2026

  Maintainers usually reply within 1 day

• Add temporal recurrence and culprit stability to rank-straggler evidence

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  distributed enhancement P1

  Difficulty 5/5 Over a week Newbie friendliness 38/100

  traceopt-ai/traceml#374 · Opened Aug 10, 2026

  Maintainers usually reply within 1 day

• Umbrella: frontend hygiene sweep (waves)

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  bug P2

  Difficulty 5/5 Over a week Newbie friendliness 35/100

  traceopt-ai/traceml#363 · 2 comments · Opened Aug 7, 2026

  Maintainers usually reply within 1 day

• Design: watch-specific verdict promotion

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  enhancement RFC

  Difficulty 4/5 3-5 days Newbie friendliness 45/100

  traceopt-ai/traceml#362 · Opened Aug 7, 2026

  Maintainers usually reply within 1 day

• RFC: schema governance for summary consumers

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  documentation enhancement RFC

  Difficulty 5/5 Over a week Newbie friendliness 35/100

  traceopt-ai/traceml#361 · Opened Aug 7, 2026

  Maintainers usually reply within 1 day

• RFC: declared coverage registry + presentation tiers

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  enhancement RFC

  Difficulty 5/5 Over a week Newbie friendliness 22/100

  traceopt-ai/traceml#360 · Opened Aug 7, 2026

  Maintainers usually reply within 1 day

• Examples validation harness: scripts/validate_examples.py + manifest + CI tiers

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 5/5 Over a week Newbie friendliness 28/100

  traceopt-ai/traceml#323 · 2 comments · Opened Aug 4, 2026

  Maintainers usually reply within 1 day

• Distribute TraceML via conda-forge

  May be free again @abhinavsriva claimed this 113 days ago, and no pull request is open. Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  enhancement

  traceopt-ai/traceml#248 · 1 comment · 1 assignee · Opened Jul 23, 2026

  Maintainers usually reply within 1 day

• Add a GPU-capable demo container variant

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  enhancement help wanted

  Difficulty 4/5 3-5 days Newbie friendliness 56/100

  traceopt-ai/traceml#244 · Opened Jul 22, 2026

  Maintainers usually reply within 1 day

• [RFC] Container strategy: from one demo image to a complete Docker story

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  enhancement help wanted

  Difficulty 5/5 Over a week Newbie friendliness 25/100

  traceopt-ai/traceml#241 · Opened Jul 22, 2026

  Maintainers usually reply within 1 day

• [RFC] Scale distributed step time telemetry with rank-local windows

  May be free again @ppraneth claimed this 116 days ago, and no pull request is open. Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  distributed runtime

  traceopt-ai/traceml#229 · 1 reaction · 3 assignees · Opened Jul 20, 2026

  Maintainers usually reply within 1 day

• [Feature]: Add low-overhead per-step collective communication timing

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  distributed enhancement help wanted

  Difficulty 5/5 Over a week Newbie friendliness 35/100

  traceopt-ai/traceml#224 · 1 comment · Opened Jul 17, 2026

  Maintainers usually reply within 1 day

• init() aggregator preflight accepts any TCP listener (no session handshake)

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  traceopt-ai/traceml#211 · 3 comments · Opened Jul 15, 2026

  Maintainers usually reply within 1 day

• [Feature]: Add Kubernetes and Kubeflow examples for TraceML distributed training

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  enhancement help wanted Integration

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  traceopt-ai/traceml#191 · Opened Jul 10, 2026

  Maintainers usually reply within 1 day

• [Feature]: Communication Profiling Beyond DDP: FSDP and Process-Group-Level Visibility

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  enhancement

  Difficulty 5/5 Over a week Newbie friendliness 30/100

  traceopt-ai/traceml#145 · Opened Jun 5, 2026

  Maintainers usually reply within 1 day

• Add Stable Baselines 3 integration for reinforcement learning workloads

  Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  enhancement help wanted

  Difficulty 4/5 3-5 days Newbie friendliness 42/100

  traceopt-ai/traceml#134 · Opened May 28, 2026

  Maintainers usually reply within 1 day

• Store batch size samples in DB for all ranks

  May be free again A pull request for this issue was closed without being merged. Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  enhancement help wanted

  Difficulty 4/5 3-5 days Newbie friendliness 20/100

  traceopt-ai/traceml#81 · 1 comment · Opened Apr 23, 2026

  Maintainers usually reply within 1 day

• Support telemetry-only execution without local SQLite or diagnostics

  Possibly taken @abhinavsriva claimed this 4 days ago. Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel

  Difficulty 4/5 3-5 days Newbie friendliness 48/100

  traceopt-ai/traceml#605 · 1 assignee · Opened Nov 5, 2026

  Maintainers usually reply within 1 day

• Demo container: add a documented dashboard-mode run

  Possibly taken A pull request linked to this issue is open or already merged. Open

  ⋯Report a problem with this issue

  • Is something wrong?

  • Already taken
  • Not beginner-friendly
  • Stale or invalid
  • Something else…
  • Send Cancel
  enhancement good first issue help wanted

  Difficulty 4/5 3-5 days Newbie friendliness 57/100

  traceopt-ai/traceml#243 · 3 comments · Opened Jul 22, 2026

  Maintainers usually reply within 1 day