deepseek新训练方法
Python · active 2024-12-30 → 2026-06-19 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 537 days from 2024-12-30 to 2026-06-19. Pushes: 183 total, peak 9 in a day. Pull requests: 127 total, peak 6 in a day. Issues: 51 total, peak 5 in a day. Comments: 121 total, peak 10 in a day. Stars: 1,694 total, peak 521 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 |
|---|---|---|---|---|
| TongLi3701 | 147 | 44 | 35 | 24 |
| YeAnbang | 123 | 59 | 16 | 20 |
| ver217 | 56 | 15 | 18 | 11 |
| BurkeHulk | 46 | 17 | 11 | 2 |
| pre-commit-ci[bot] | 36 | 36 | 0 | 0 |
| flybird11111 | 29 | 4 | 11 | 11 |
| duanjunwen | 24 | 1 | 5 | 11 |
| Issues-translate-bot | 12 | 0 | 0 | 12 |
| binmakeswell | 8 | 4 | 2 | 0 |
| sdjasj | 7 | 0 | 0 | 2 |
| Sze-qq | 5 | 2 | 2 | 0 |
| botbw | 5 | 0 | 1 | 2 |
| ElcarimQAQ | 5 | 0 | 5 | 0 |
| Yanjia0 | 4 | 0 | 4 | 0 |
| Guodanding | 3 | 0 | 0 | 0 |
| wangbluo | 3 | 0 | 3 | 0 |
| littletomatodonkey | 3 | 0 | 0 | 1 |
| ericxsun | 3 | 0 | 0 | 1 |
| CelesteBeck | 3 | 0 | 0 | 2 |
| AmazDeng | 3 | 0 | 0 | 2 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#6387Truong57242026-04-14 04:49[BUG]: LowLevelZeroPlugin with use_fp8=true causes `shape mismatch error: shape [32, 512] is invalid for input of size 512`
- Pull request#6411Yanjia02026-04-09 08:33
- Pull request#6410Yanjia02026-04-09 01:49
- Pull request#6408Truong57242026-04-02 10:23
- Pull request#6405Mr-Neutr0n2026-02-11 18:17
- Pull request#6404thecaptain7892026-02-07 21:46
- Issue#6403williamyang49782026-01-15 02:39Do you have the latest version of Docker images?
- Issue#6402williamyang49782026-01-14 08:50Can you provide the latest Dockerfile?
- Issue#6401sdjasj2026-01-14 05:15[BUG]: FusedLAMB Fails with LowLevelZeroPlugin When Using Small initial_scale and CPU Offload
- Issue#6401sdjasj2026-01-14 05:15[BUG]: FusedLAMB Fails with LowLevelZeroPlugin When Using Small initial_scale and CPU Offload
- Issue#6400sdjasj2026-01-14 05:03[BUG]: Lamb Optimizer Fails with LowLevelZeroPlugin When Using Small initial_scale and CPU Offload
- Issue#6400sdjasj2026-01-14 05:03[BUG]: Lamb Optimizer Fails with LowLevelZeroPlugin When Using Small initial_scale and CPU Offload
- Issue comment#6389sdjasj2025-12-24 06:26[BUG]: TypeError in ColoAttention.attention() when enable_all_optimization=True in HybridParallelPlugin
- Issue comment#6396sdjasj2025-12-24 06:25[BUG]: `KeyError: 'sum'` When Using Adagrad Optimizer with LowLevelZeroPlugin (initial_scale=8.33)
- Issue comment#6389jkeo1802025-12-09 05:18[BUG]: TypeError in ColoAttention.attention() when enable_all_optimization=True in HybridParallelPlugin
- Issue comment#6393SamareshSingh2025-12-01 20:48Fix FP8 linear layer dimension check to prevent runtime error
- Pull request#6393ssam182025-11-12 17:19
- Issue comment#2938knightnemo2025-11-10 13:43[BUG]: AttributeError: type object 'Trainer' has no attribute 'add_argparse_args'
- Issue comment#6356YeAnbang2025-11-10 01:42[feat] Support Zero Bubble StreamRL-like RL Training
- Issue#6386sdjasj2025-11-01 10:24[BUG]: `AssertionError: div_scale should remain default` when using Gemini plugin with CPUAdam
- Pull request#6385Sze-qq2025-09-26 06:54
- Pull request#6385Yanjia02025-09-26 03:57
- Pull request#6384YeAnbang2025-09-16 08:33
- Pull request#6294TongLi37012025-08-26 13:51
- Issue comment#6381TongLi37012025-08-26 05:59[BUG]: colossal cannot split tensor evenly when using Sequential Parallelism in hybirdplugin
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 1,694 stars here means stars gained during the window, not the repo's star count.