提示词反推A novel Multimodal Large Language Model (MLLM) architecture, designed to structurally align visual and textual embeddings.
active 2025-04-02 → 2026-05-16 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 410 days from 2025-04-02 to 2026-05-16. Pushes: 5 total, peak 1 in a day. Pull requests: 4 total, peak 2 in a day. Issues: 18 total, peak 2 in a day. Comments: 28 total, peak 4 in a day. Stars: 321 total, peak 31 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 |
|---|---|---|---|---|
| runninglsy | 7 | 3 | 0 | 2 |
| wsbagnsv1 | 5 | 0 | 0 | 5 |
| CLAassistant | 4 | 0 | 0 | 4 |
| JumpingRain | 4 | 1 | 3 | 0 |
| Dineshkumar-Anandan-ZS0367 | 2 | 0 | 0 | 1 |
| sshan-zhao | 2 | 1 | 1 | 0 |
| KeyaoZhao | 2 | 0 | 0 | 0 |
| hanshupe007 | 2 | 0 | 0 | 1 |
| elisaife | 2 | 0 | 0 | 1 |
| TragedyN | 2 | 0 | 0 | 2 |
| AlexisMDP | 1 | 0 | 0 | 1 |
| bbeiler-ridgeline | 1 | 0 | 0 | 1 |
| ecnuycxie | 1 | 0 | 0 | 0 |
| a-myth-biswas | 1 | 0 | 0 | 1 |
| XyWzzZ | 1 | 0 | 0 | 0 |
| cooleel | 1 | 0 | 0 | 0 |
| wonderisland | 1 | 0 | 0 | 0 |
| SkywalkerSZW | 1 | 0 | 0 | 0 |
| yogkul2000 | 1 | 0 | 0 | 0 |
| XueWei1997 | 1 | 0 | 0 | 1 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#125huazhenliu2025-12-25 06:25Ovis2.5 1B/4B model
- Issue comment#124hfassold2025-12-11 05:30Do NOT add Ovis2.5 model to transfromers package
- Issue comment#110a-myth-biswas2025-11-27 10:37cannot forbid "enable_thinking" through change to false
- Issue comment#123wsbagnsv12025-10-24 01:59Quantization support needed for Ovis2.5 9B
- Issue comment#105wsbagnsv12025-10-21 15:55Inference without Flash Attention is Needed!
- Issue comment#123Dineshkumar-Anandan-ZS03672025-09-30 02:43Quantization support needed for Ovis2.5 9B
- Issue#123Dineshkumar-Anandan-ZS03672025-09-26 05:25Quantization support needed for Ovis2.5 9B I have tried to quantize the model using ms-swift and this architecture is not supported yet. When can I expect or else share the quantization code for us.
- Issue comment#119XueWei19972025-09-22 06:28定位能力
- Issue#120ksooklall-arlo2025-09-19 02:28Finetuing Ovis2.5-2B for video caption
- Issue#118wonderisland2025-09-18 03:57reasoning parser for vllm
- Issue comment#112runninglsy2025-09-08 06:49Is the <think> process always in Chinese
- Issue comment#103wsbagnsv12025-09-05 09:52Quantization support
- Issue#113casillas11112025-09-04 07:35Multi-card inference of Ovis2.5-9B
- Issue comment#111anuramat2025-09-03 20:22Value error, Model architectures ['Ovis2_5'] are not supported for now.
- Issue#112yogkul20002025-09-03 19:19Is the <think> process always in Chinese
- Issue comment#109elisaife2025-09-03 06:11cannot use vllm to infer ovis2.5
- Issue comment#109TragedyN2025-09-03 06:02cannot use vllm to infer ovis2.5
- Issue comment#109TragedyN2025-09-03 03:32cannot use vllm to infer ovis2.5
- Issue#109elisaife2025-09-02 07:24cannot use vllm to infer ovis2.5
- Issue#107Graysonicc2025-09-01 07:33verl m目前支持ovis2.5吗
- Issue comment#98Magmanat2025-08-30 11:35Ovis2.5-2B Inference speed
- Issue comment#106hanshupe0072025-08-30 09:05Ovis 2.5 vs Qwen 2.5 VL
- Issue comment#103wsbagnsv12025-08-30 08:07Quantization support
- Issue#106hanshupe0072025-08-29 15:42Ovis 2.5 vs Qwen 2.5 VL
- Issue comment#98AlexisMDP2025-08-24 19:19Ovis2.5-2B Inference speed
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 321 stars here means stars gained during the window, not the repo's star count.