[Fork] Training library for Megatron-based models with bi-directional Hugging Face conversion capability
Python · active 2025-08-08 → 2026-08-10 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 368 days from 2025-08-08 to 2026-08-10. Pushes: 7,532 total, peak 77 in a day. Pull requests: 1,812 total, peak 23 in a day. Issues: 1,055 total, peak 106 in a day. Comments: 4,098 total, peak 48 in a day. Stars: 227 total, peak 6 in a day.
- Pushes
- Pull requests
- Issues
- Comments
- Stars
Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.
Top contributors
Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity
| Contributor | Contributions | Pushes | PRs | Comments |
|---|---|---|---|---|
| copy-pr-bot[bot] | 3K | 2.4K | 0 | 613 |
| yaoyu-33 | 2.4K | 1.1K | 243 | 622 |
| ko3n1g | 1.4K | 926 | 248 | 77 |
| ananthsub | 1.2K | 175 | 232 | 520 |
| cuichenx | 845 | 465 | 87 | 167 |
| coderabbitai[bot] | 785 | 1 | 1 | 515 |
| malay-nagda | 643 | 369 | 95 | 83 |
| chtruong814 | 619 | 239 | 217 | 59 |
| svcnvidia-nemo-ci | 472 | 259 | 132 | 62 |
| maanug-nv | 387 | 91 | 41 | 113 |
| gautham-kollu | 299 | 136 | 36 | 74 |
| dimapihtar | 238 | 166 | 9 | 43 |
| github-actions[bot] | 224 | 17 | 80 | 55 |
| erhoo82 | 205 | 9 | 7 | 64 |
| snowmanwwg | 170 | 4 | 8 | 11 |
| liding-nv | 146 | 86 | 10 | 35 |
| suiyoubi | 140 | 85 | 13 | 24 |
| HollowMan6 | 125 | 3 | 19 | 59 |
| claude[bot] | 124 | 0 | 0 | 72 |
| dingqingy-nv | 119 | 13 | 27 | 21 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#5444yaoyu-332026-08-08 13:24fix(training): resolve MegatronMIMO evaluation batch defaults
- Pull request#5432yaoyu-332026-08-08 04:42
- Issue#5434yaoyu-332026-08-07 22:50[feature] Integrate Megatron-Core Generalized Tensor Parallelism (GTP)
- Issue#5434yaoyu-332026-08-07 22:50[feature] Integrate Megatron-Core Generalized Tensor Parallelism (GTP)
- Pull request#5327cuichenx2026-08-07 21:41
- Issue comment#5427copy-pr-bot[bot]2026-08-07 21:33cp: `[examples] refactor: direct GLM inference to model card (5346)` into `r0.6.0`
- Issue comment#5394nvegesna-netizen2026-08-07 20:42perf(recipe): llama3 8B GB200 NVFP4 — run DPA in FP8 current scaling
- Pull request#5368yaoyu-332026-08-07 19:47
- Pull request#5368yaoyu-332026-08-07 19:47
- Issue comment#5360copy-pr-bot[bot]2026-08-07 19:35dsv4 flash verification card and export fixes
- Issue comment#5369svcnvidia-nemo-ci2026-08-07 18:38test: AUT-1363 validate MCore dev compatibility
- Issue comment#5416svcnvidia-nemo-ci2026-08-07 18:13chore(deps): AUT-1409 update mlflow to 3.15.1
- Issue comment#5416copy-pr-bot[bot]2026-08-07 18:09chore(deps): AUT-1409 update mlflow to 3.15.1
- Issue comment#5415copy-pr-bot[bot]2026-08-07 17:44ci: AUT-1408 replace cicd PAT with GitHub App tokens
- Pull request#5410svcnemo-autobot2026-08-07 17:25
- Issue comment#5369svcnvidia-nemo-ci2026-08-07 17:25test: AUT-1363 validate MCore dev compatibility
- Issue#4596cuichenx2026-08-07 17:14[tracking] SFT dataset unification / upstreaming with Megatron Core
- Issue comment#5414copy-pr-bot[bot]2026-08-07 17:07cp: `fix(peft): preserve grouped expert SwiGLU LoRA order (5376)` into `r0.6.0`
- Pull request#5414svcnvidia-nemo-ci2026-08-07 17:07
- Issue comment#5402yaoyu-332026-08-07 17:06fix(recipes): use max-pool FSDP for Qwen3.5-VL
- Issue comment#5369svcnvidia-nemo-ci2026-08-07 14:37test: AUT-1363 validate MCore dev compatibility
- Pull request#5373yaoyu-332026-08-06 16:19
- Issue comment#4998claude[bot]2026-07-28 23:18fix(training): count expert parameter norms with TP and EP
- Pull request#5062yaoyu-332026-07-24 20:28
- Pull request#5029yaoyu-332026-07-22 18:17
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 227 stars here means stars gained during the window, not the repo's star count.