This is the official repository of our paper "What If We Recaption Billions of Web Images with LLaMA-3 ?"
active 2024-06-13 → 2026-01-29 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 596 days from 2024-06-13 to 2026-01-29. Pushes: 5 total, peak 5 in a day. Pull requests: 0 total, peak 0 in a day. Issues: 22 total, peak 3 in a day. Comments: 35 total, peak 5 in a day. Stars: 136 total, peak 37 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 |
|---|---|---|---|---|
| xhl-video | 9 | 1 | 0 | 8 |
| ImKeTT | 8 | 1 | 0 | 7 |
| LinB203 | 7 | 0 | 0 | 5 |
| Coobiw | 4 | 0 | 0 | 3 |
| cihangxie | 4 | 0 | 0 | 2 |
| MonolithFoundation | 3 | 0 | 0 | 2 |
| TheFllood | 3 | 3 | 0 | 0 |
| XavierHeart | 3 | 0 | 0 | 1 |
| HarborYuan | 3 | 0 | 0 | 1 |
| viyjy | 3 | 0 | 0 | 1 |
| YunjieYu | 3 | 0 | 0 | 3 |
| WANGSSSSSSS | 2 | 0 | 0 | 0 |
| YoadTew | 2 | 0 | 0 | 0 |
| taiyouZhang | 1 | 0 | 0 | 0 |
| kdwonn | 1 | 0 | 0 | 1 |
| vishaal27 | 1 | 0 | 0 | 0 |
| pbaylies | 1 | 0 | 0 | 1 |
| andimarafioti | 1 | 0 | 0 | 0 |
| YoPatapon | 1 | 0 | 0 | 0 |
| xin-ran-w | 1 | 0 | 0 | 0 |
Recent activity
Latest issues, pull requests and releases
- Issue#16vishaal272024-09-12 18:27Releasing the ViT-S/16 Recap-CLIP models
- Issue comment#6LinB2032024-09-03 05:19How much storage space is needed to save 1B images?
- Issue#15LinB2032024-09-03 04:27How to get distribution of resolution?
- Issue comment#11xhl-video2024-08-25 06:21A certain degree of mismatch between the `Key` and the original DataComp-1B
- Issue comment#11kdwonn2024-08-20 20:24A certain degree of mismatch between the `Key` and the original DataComp-1B
- Issue comment#2YunjieYu2024-08-14 10:01Code used to generate the dataset
- Issue comment#5YunjieYu2024-08-14 09:47The descriptions start with "The image depicts/features" instead of going directly to the object
- Issue comment#12YunjieYu2024-08-14 09:34Can you provide a demo for recaption?
- Issue#14XavierHeart2024-08-08 05:58Some questions about the Version of DataComp1B used by Recap-DataComp1B
- Issue comment#14XavierHeart2024-08-08 05:58Some questions about the Version of DataComp1B used by Recap-DataComp1B
- Issue comment#14xhl-video2024-08-07 16:56Some questions about the Version of DataComp1B used by Recap-DataComp1B
- Issue comment#11xhl-video2024-08-07 16:45A certain degree of mismatch between the `Key` and the original DataComp-1B
- Issue comment#11LinB2032024-08-07 04:23A certain degree of mismatch between the `Key` and the original DataComp-1B
- Issue comment#11xhl-video2024-08-07 04:20A certain degree of mismatch between the `Key` and the original DataComp-1B
- Issue comment#11LinB2032024-08-07 02:50A certain degree of mismatch between the `Key` and the original DataComp-1B
- Issue comment#11Coobiw2024-08-05 11:22A certain degree of mismatch between the `Key` and the original DataComp-1B
- Issue comment#11LinB2032024-08-05 07:38A certain degree of mismatch between the `Key` and the original DataComp-1B
- Issue#14XavierHeart2024-07-10 07:58Some questions about the Version of DataComp1B used by Recap-DataComp1B
- Issue comment#11Coobiw2024-07-04 07:25A certain degree of mismatch between the `Key` and the original DataComp-1B
- Issue#13xin-ran-w2024-07-03 15:10The GPU resource and time needed for training DiT.
- Issue comment#10xhl-video2024-07-01 23:05All zero value for both original and recap clip scores in parquets.
- Issue comment#11xhl-video2024-07-01 23:03A certain degree of mismatch between the `Key` and the original DataComp-1B
- Issue comment#11Coobiw2024-07-01 08:17A certain degree of mismatch between the `Key` and the original DataComp-1B
- Issue#11Coobiw2024-07-01 08:15A certain degree of mismatch between the `Key` and the original DataComp-1B
- Issue#10YoPatapon2024-07-01 02:26All zero value for both original and recap clip scores in parquets.
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 136 stars here means stars gained during the window, not the repo's star count.