[ICLR 2026] pi-Flow: Policy-Based Few-Step Generation via Imitation Distillation
active 2025-10-17 → 2026-04-24 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 190 days from 2025-10-17 to 2026-04-24. Pushes: 27 total, peak 3 in a day. Pull requests: 1 total, peak 1 in a day. Issues: 21 total, peak 2 in a day. Comments: 35 total, peak 4 in a day. Stars: 95 total, peak 12 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 |
|---|---|---|---|---|
| Lakonik | 49 | 27 | 0 | 14 |
| oljike | 5 | 0 | 0 | 3 |
| yic666 | 3 | 0 | 0 | 2 |
| zwukong | 3 | 0 | 0 | 2 |
| BryantGary | 3 | 0 | 0 | 1 |
| coder4nlp | 3 | 0 | 0 | 1 |
| YunkangTao | 3 | 0 | 0 | 3 |
| sitatec | 2 | 0 | 0 | 2 |
| alexpuliatti | 2 | 0 | 0 | 1 |
| cbyan1003 | 2 | 0 | 0 | 0 |
| vibe-viscot | 2 | 0 | 0 | 2 |
| itsuzef | 1 | 0 | 1 | 0 |
| pnotp | 1 | 0 | 0 | 1 |
| changhy666 | 1 | 0 | 0 | 0 |
| lisandropuzzolo | 1 | 0 | 0 | 1 |
| ItsThanhTung | 1 | 0 | 0 | 1 |
| yuanzhi-zhu | 1 | 0 | 0 | 0 |
| kaptainkory | 1 | 0 | 0 | 0 |
| RandomGuyWithIssues | 1 | 0 | 0 | 1 |
Recent activity
Latest issues, pull requests and releases
- Pull request#27itsuzef2026-04-24 02:26
- Issue#9Lakonik2026-02-12 01:08Please add tensor parallelism support
- Issue comment#9Lakonik2026-02-12 01:07Please add tensor parallelism support
- Issue#26yic6662026-01-09 06:55关于少卡训练
- Issue comment#25yic6662026-01-08 02:49Does it work well with Qwen-Image-2512
- Issue comment#25sitatec2026-01-08 02:37Does it work well with Qwen-Image-2512
- Issue comment#25sitatec2026-01-07 20:07Does it work well with Qwen-Image-2512
- Issue comment#25Lakonik2026-01-07 18:13Does it work well with Qwen-Image-2512
- Issue comment#25yic6662026-01-07 09:53Does it work well with Qwen-Image-2512
- Issue comment#6BryantGary2026-01-05 04:13Looking forward to the qwen image edit version
- Issue#24BryantGary2026-01-04 08:11Error when cancel the line of load_adapter_lora in the inference script.
- Issue#23Lakonik2026-01-02 03:46where is the "data/t2i_prompts_3m/preproc_qwen.jsonl.gz" and the "data/qwen_empty_prompt_embeds.pth" ?
- Issue#23BryantGary2025-12-30 15:11where is the "data/t2i_prompts_3m/preproc_qwen.jsonl.gz" and the "data/qwen_empty_prompt_embeds.pth" ?
- Issue comment#8YunkangTao2025-12-26 15:52视频生成有noise
- Issue#22cbyan10032025-12-21 09:40How to distill other model on piFlow?
- Issue comment#8YunkangTao2025-12-16 08:55视频生成有noise
- Issue comment#8YunkangTao2025-12-13 06:05视频生成有noise
- Issue#14Lakonik2025-12-13 03:55Why adding lora on text stream
- Issue#18Lakonik2025-12-13 03:55Does the algorithm rely on previous work GMFlow?
- ReleaseLakonik2025-12-13 02:54LakonLab v0.1.0
- Issue#20kaptainkory2025-12-08 03:41Scaled FP8 errors with Qwen-Image default workflow
- Issue comment#2alexpuliatti2025-12-05 23:36edit pipeline
- Issue#8Lakonik2025-12-05 18:52视频生成有noise
- Issue comment#8Lakonik2025-12-05 18:52视频生成有noise
- Issue comment#19Lakonik2025-12-05 18:51蒸馏后图片质量比较差
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 95 stars here means stars gained during the window, not the repo's star count.