Training models using DeepSpeed in multi-cluster distributed way
active 2025-02-04 → 2026-07-28 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 540 days from 2025-02-04 to 2026-07-28. Pushes: 33 total, peak 5 in a day. Pull requests: 32 total, peak 4 in a day. Issues: 7 total, peak 2 in a day. Comments: 31 total, peak 3 in a day. Stars: 310 total, peak 8 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 |
|---|---|---|---|---|
| tohtana | 23 | 13 | 10 | 0 |
| hwchen2017 | 19 | 8 | 7 | 2 |
| tjruwase | 18 | 5 | 6 | 2 |
| sfc-gh-truwase | 11 | 4 | 2 | 1 |
| Antlera | 8 | 0 | 2 | 4 |
| delock | 5 | 2 | 0 | 2 |
| neeldani | 3 | 0 | 0 | 1 |
| Buddingpopp | 2 | 0 | 0 | 0 |
| inkcherry | 2 | 0 | 0 | 2 |
| jaydeepborkar | 2 | 0 | 0 | 0 |
| 2460707925 | 2 | 0 | 0 | 2 |
| GuanhuaWang | 2 | 0 | 0 | 2 |
| jouw | 2 | 0 | 1 | 1 |
| StoKou | 2 | 0 | 0 | 0 |
| loadams | 2 | 1 | 0 | 0 |
| TamerSoliman | 1 | 0 | 1 | 0 |
| therealnaveenkamal | 1 | 0 | 0 | 1 |
| xylian86 | 1 | 0 | 1 | 0 |
| raviguptaamd | 1 | 0 | 1 | 0 |
| Nas-Azzam | 1 | 0 | 0 | 1 |
Recent activity
Latest issues, pull requests and releases
- Pull request#998tohtana2026-02-07 21:34
- Pull request#998tohtana2026-01-22 00:13
- Pull request#994tohtana2025-12-19 20:29
- Pull request#994tohtana2025-12-11 23:23
- Pull request#993KairongChen-l2025-12-09 14:21
- Pull request#992TamerSoliman2025-11-22 20:01
- Pull request#990sfc-gh-truwase2025-09-29 17:03
- Pull request#990xylian862025-09-14 01:58
- Issue comment#524luanshang2025-09-09 01:36RuntimeError: Error building extension 'transformer_inference'
- Issue#989delock2025-09-08 08:40One example, multiple config files
- Issue#988jaydeepborkar2025-08-28 05:40Gradient overflow issue while using deepspeed
- Issue#988jaydeepborkar2025-08-28 00:37Gradient overflow issue while using deepspeed
- Pull request#987Antlera2025-08-20 17:22
- Pull request#982sfc-gh-truwase2025-08-16 21:18
- Issue#986Buddingpopp2025-08-11 06:04Is there a micro-benchmark that can test the last two optimizations mentioned in the FastPersist paper, "Parallelizing checkpoint writes across DP ranks and Pipelining checkpoint writes"?
- Issue comment#982delock2025-08-08 14:54Add Benchmarking and Fine-Tuning Support for ZenFlow
- Issue comment#982Antlera2025-08-07 03:42Add Benchmarking and Fine-Tuning Support for ZenFlow
- Issue comment#982delock2025-08-06 05:41Add Benchmarking and Fine-Tuning Support for ZenFlow
- Issue comment#982Antlera2025-08-04 18:47Add Benchmarking and Fine-Tuning Support for ZenFlow
- Issue comment#186smartparrot2025-07-23 06:15OOM despite ZeRO stage 3
- Issue comment#983tohtana2025-07-18 17:29fix init weights issue for critic/reward model
- Issue comment#983jouw2025-07-08 04:29fix init weights issue for critic/reward model
- Pull request#981hwchen20172025-07-04 05:24
- Issue comment#983hwchen20172025-07-03 18:15fix init weights issue for critic/reward model
- Issue comment#981hwchen20172025-07-03 16:42Update submodule link to reflect https style
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 310 stars here means stars gained during the window, not the repo's star count.