The official repository for ERNIE 4.5 and ERNIEKit – its industrial-grade development toolkit based on PaddlePaddle.
active 2025-04-02 → 2026-05-25 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 419 days from 2025-04-02 to 2026-05-25. Pushes: 164 total, peak 9 in a day. Pull requests: 454 total, peak 20 in a day. Issues: 95 total, peak 9 in a day. Comments: 354 total, peak 14 in a day. Stars: 766 total, peak 242 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 |
|---|---|---|---|---|
| nepeplwu | 259 | 61 | 47 | 66 |
| wtmlon | 97 | 23 | 33 | 13 |
| lugimzzz | 87 | 22 | 25 | 7 |
| paddle-bot[bot] | 62 | 0 | 13 | 49 |
| Jonathans575 | 59 | 0 | 37 | 16 |
| liym27 | 56 | 13 | 14 | 12 |
| zjjlivein | 46 | 11 | 30 | 1 |
| BossPi | 44 | 0 | 37 | 7 |
| waliwali777 | 42 | 0 | 26 | 9 |
| FeixLiu | 36 | 9 | 23 | 2 |
| Xing-lil | 32 | 4 | 15 | 4 |
| forBlank | 31 | 0 | 12 | 14 |
| stale[bot] | 26 | 0 | 0 | 14 |
| GITD245 | 19 | 0 | 12 | 5 |
| xinruiM | 19 | 0 | 15 | 2 |
| ZeyuChen | 16 | 4 | 4 | 2 |
| DesmonDay | 15 | 3 | 8 | 0 |
| CLAassistant | 14 | 0 | 0 | 14 |
| yongqiangma | 13 | 4 | 2 | 1 |
| jzhang533 | 11 | 2 | 1 | 2 |
Recent activity
Latest issues, pull requests and releases
- Issue#1451panxiaohuan92026-05-25 06:47sft后paddleocr-vl模型推理报错：ValueError: No Paddle model files were found in './PaddleOCR-VL-SFT-handw2'
- Pull request#14494ek02026-05-24 05:14
- Issue comment#1449paddle-bot[bot]2026-05-24 05:14[Docs] Add vLLM deployment guide for ERNIE 4.5 models (EN/ZH)
- Pull request#14494ek02026-05-24 05:14
- Issue#1439croviatrust2026-03-12 07:02📋 Documentation Enhancement Suggestion
- Issue comment#1382nepeplwu2026-02-26 03:01我之前ai studio花钱租卡创建的项目，现在不能运行了？
- Issue#1437megemini2026-02-25 05:02微调 PaddleOCR-VL 后模型名称建议重命名为 model.safetensors
- Issue comment#1372Blinkion2026-02-21 12:02baidu/ERNIE-4.5-VL-28B-A3B-Thinking本地图片示例代码
- Issue#1372nepeplwu2026-02-21 12:01baidu/ERNIE-4.5-VL-28B-A3B-Thinking本地图片示例代码
- Issue comment#1372nepeplwu2026-02-21 12:01baidu/ERNIE-4.5-VL-28B-A3B-Thinking本地图片示例代码
- Issue#1375nepeplwu2026-02-20 12:01ModuleNotFoundError: No module named 'paddleformers.transformers.ernie4_5_moe_vl'
- Issue#1370nepeplwu2026-02-19 12:02what's the format of inline formula?
- Issue comment#1363nepeplwu2026-02-18 12:01Token Count Calculation in SFT Data Distribution Curation
- Issue comment#1434Bobholamovic2026-02-10 07:28PaddleOCR-VL按照官方孟加拉语数据集finetuning后单样本推理报错
- Issue#1348nepeplwu2026-02-09 12:02paddleocr-vl finetuning dataset format
- Issue#1436nepeplwu2026-02-09 03:01Think mode for text generation model
- Issue comment#1432MYBao2172026-01-26 07:47监督微调PaddleOCR-VL，加载Checkpoint恢复训练的时候失败
- Issue comment#1427HWChatGPT42026-01-26 03:34paddleocr-vl经过全量sft训练后推理速度显著变慢的原因？
- Pull request#1433lshpku2026-01-23 06:28
- Issue#1431Hor1zon-thu2026-01-22 09:44ppocr-vl sft后效果变差
- Issue comment#1429forBlank2026-01-22 06:52paddleocr-vl进行sft训练时报错ValueError: PaddleRecall error(102): LossNan. Loss contains inf or nan values, its value is nan
- Issue comment#1396hustzxl2026-01-22 03:05PaddleOCR-VL 微调表格标注
- Issue comment#1405longzeyilang2026-01-22 02:06PaddleOCR-VL LoRA 微调后 怎么合并模型？
- Issue comment#1309nepeplwu2026-01-21 12:01使用FastDeploy和ERNIEkit部署训练后模型，均出现问题
- Issue comment#1429HWChatGPT42026-01-20 06:46paddleocr-vl进行sft训练时报错ValueError: PaddleRecall error(102): LossNan. Loss contains inf or nan values, its value is nan
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 766 stars here means stars gained during the window, not the repo's star count.