Official implementation of BLIP3o-Series
Python · active 2025-05-13 → 2026-05-23 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 376 days from 2025-05-13 to 2026-05-23. Pushes: 186 total, peak 30 in a day. Pull requests: 2 total, peak 1 in a day. Issues: 53 total, peak 3 in a day. Comments: 201 total, peak 11 in a day. Stars: 754 total, peak 73 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 |
|---|---|---|---|---|
| JiuhaiChen | 265 | 174 | 0 | 91 |
| zhiyangxu-VT | 13 | 10 | 0 | 3 |
| alfassy | 9 | 0 | 0 | 6 |
| petermg | 8 | 0 | 0 | 6 |
| mmderakhshani | 7 | 0 | 0 | 7 |
| dhevarghese | 7 | 0 | 0 | 6 |
| lky-violet | 7 | 0 | 0 | 6 |
| championbridge | 6 | 0 | 0 | 4 |
| Luciennnnnnn | 6 | 0 | 0 | 3 |
| Jaguar156 | 6 | 0 | 0 | 5 |
| Columbine21 | 6 | 0 | 0 | 2 |
| pruksmhc | 6 | 0 | 0 | 3 |
| henbucuoshanghai | 5 | 0 | 0 | 3 |
| zch1102 | 5 | 0 | 0 | 5 |
| fahadskhan | 4 | 0 | 0 | 3 |
| juexZZ | 4 | 0 | 0 | 2 |
| Amshaker | 3 | 0 | 0 | 1 |
| TtuHamg | 3 | 0 | 0 | 1 |
| guozhiyao | 3 | 0 | 0 | 2 |
| kuaikuaixym | 3 | 0 | 0 | 2 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#98kuaikuaixym2026-03-12 00:25Performance Discrepancy in Pretraining Replication on BLIP3o-NEXT Branch (28% vs 31% GenEval)
- Issue comment#98kuaikuaixym2026-03-11 05:39Performance Discrepancy in Pretraining Replication on BLIP3o-NEXT Branch (28% vs 31% GenEval)
- Issue#98kuaikuaixym2026-03-10 03:31Performance Discrepancy in Pretraining Replication on BLIP3o-NEXT Branch (28% vs 31% GenEval)
- Issue comment#97JiuhaiChen2026-02-23 18:18Bug of Understanding Inference.
- Issue comment#97KingH121382026-01-05 08:10Bug of Understanding Inference.
- Issue comment#90KingH121382025-12-25 07:42About the Understanding Performance.
- Issue comment#90KingH121382025-12-25 06:58About the Understanding Performance.
- Issue#92JuntongWang2025-11-29 13:29How to perform Image Editing tasks? Missing example code.
- Issue comment#91fahadskhan2025-11-25 21:40Release of BLIP3o-Edit
- Issue comment#75zzxslp2025-11-25 00:11Diverging training?
- Issue comment#75zzxslp2025-11-24 00:53Diverging training?
- Issue comment#91JiuhaiChen2025-11-20 18:00Release of BLIP3o-Edit
- Issue#90Columbine212025-11-16 09:00About the Understanding Performance.
- Issue comment#87lky-violet2025-10-31 01:17Questions about supervision labels and architecture consistency in BLIP-3o-NEXT
- Issue comment#57JiuhaiChen2025-10-27 23:54Autogressive model generates visiual features,as inputs to Dit?how?where are details?
- Issue comment#71taoyang11222025-10-25 23:02Wechat QRcode is invalid
- Issue comment#83JiuhaiChen2025-10-17 17:43Missing attention_mask in hidden_states calculation causes blurry images in batch inference
- Pull request#85qqzy2025-10-17 06:58
- Issue#84qqzy2025-10-15 07:57Performance degradation on DPG after RL
- Issue#83qqzy2025-10-15 07:00Missing attention_mask in hidden_states calculation causes blurry images in batch inference
- Issue comment#82JiuhaiChen2025-09-17 19:35The training loss curve during the fine-tuning stage
- Issue comment#81JiuhaiChen2025-09-15 18:54About the Pretraining and SFT stage differences
- Issue#81Columbine212025-09-15 03:04About the Pretraining and SFT stage differences
- Issue comment#76Jaguar1562025-09-12 19:43Inquiry about training compute setup
- Issue comment#76JiuhaiChen2025-09-12 03:14Inquiry about training compute setup
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 754 stars here means stars gained during the window, not the repo's star count.