[Fork] Qwen3-VL is the multimodal large language model series developed by Qwen team, Alibaba Cloud.
Jupyter Notebook · active 2025-09-23 → 2026-08-10 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 322 days from 2025-09-23 to 2026-08-10. Pushes: 25 total, peak 5 in a day. Pull requests: 40 total, peak 6 in a day. Issues: 257 total, peak 20 in a day. Comments: 639 total, peak 30 in a day. Stars: 2,471 total, peak 245 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 |
|---|---|---|---|---|
| JJJYmmm | 109 | 0 | 2 | 107 |
| ShuaiBai623 | 68 | 25 | 9 | 9 |
| wulipc | 39 | 0 | 3 | 36 |
| ywang96 | 18 | 0 | 1 | 17 |
| zkailinzhang | 16 | 0 | 0 | 6 |
| John-Ge | 15 | 0 | 2 | 13 |
| HumanZhong | 13 | 0 | 0 | 13 |
| ycsun1972 | 10 | 0 | 0 | 10 |
| Xqle | 8 | 0 | 0 | 6 |
| Arslan-Mehmood1 | 6 | 0 | 0 | 4 |
| Pig255 | 6 | 0 | 0 | 4 |
| sibosutd | 6 | 0 | 0 | 6 |
| Xiaohui9607 | 6 | 0 | 0 | 5 |
| litengfeigithub | 5 | 0 | 0 | 3 |
| ludunjie1219 | 5 | 0 | 0 | 5 |
| JerryZ01 | 5 | 0 | 0 | 2 |
| Asunatan | 5 | 0 | 0 | 1 |
| dyhuachi | 5 | 0 | 0 | 4 |
| Caesar107 | 5 | 0 | 0 | 3 |
| maxgameone | 4 | 0 | 0 | 2 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#2095JJJYmmm2026-06-26 12:21Output text can't exceed 375 words
- Issue comment#716bandomatteo2026-05-31 20:55Issue with add_vision_ids=True argument
- Issue comment#1904aperezr202026-05-15 21:19about odinw evaluation
- Issue#2089FoundaryCore2026-04-26 20:34Qwen2.5-VL-7B vs 32B
- Issue comment#65ainiyo-wj2026-04-24 06:59KeyError: 'video_fps'
- Issue#2087ymp50782026-04-23 23:43How was patch_embed.proj (Conv3d) initialized from SigLIP2 weights (Conv2d)?
- Issue comment#2058soundarmoorthy2026-04-14 22:13AttributeError: Qwen2Tokenizer has no attribute all_special_tokens_extended. Did you mean: 'num_special_tokens_to_add'?
- Issue comment#2068JJJYmmm2026-04-09 04:22qwen3vl preprocess video: smart_resize in transformers is different with smart_resize in qwen-vl-utils/vision_process.py
- Issue comment#2064JJJYmmm2026-04-09 04:18Qwen3-VL，支持华为升腾300系列的npu部署吗？
- Issue comment#2072JJJYmmm2026-04-09 04:17多图推理时输出被截断
- Issue#2082cwgxy2026-04-07 03:015090推理时，报CUDA error: an illegal memory access was encountered
- Issue comment#2059Zhh-Fan2026-03-28 07:27请问是否有Qwen3VL-8B 或同系列模型在AndroidControl上的分数？
- Issue comment#1748FiveONE51512026-03-27 00:50Dose qwen3vl has a different evaluation strategy from qwen2.5vl in MATH-Vision? As the improvements in MATH-Vision is quite large.
- Issue#2072Timeviewerzero2026-03-24 07:35多图推理时输出被截断
- Issue comment#2061linhan122026-03-24 06:29Qwen3-VL-4B-Instruct在zero bench_sub上的评测复现
- Issue comment#1832chenxl1245782026-03-22 08:34NCCL timeout when training Qwen3 with mixed video and image datasets
- Issue comment#1844edx-h2026-03-21 03:37使用官方给的Huggingface推理代码，推理8b模型显存溢出，而且推理过程只要加了图片就推理的特别慢
- Issue comment#2026zkw522026-03-20 03:03Qwen3-VL LoRA 微调 OOM
- Issue#2066CharlesChen242026-03-17 10:05Qwen3-VL-4B-Instruct 如何设置参数，输入一致，保证输出一致
- Issue#2061linhan122026-03-12 03:31Qwen3-VL-4B-Instruct在zero bench_sub上的评测复现
- Issue comment#2057ChencongZJU2026-03-09 06:19Qwen3VL-8B-Instruct 在LVBench上指标的复现
- Issue comment#2041JJJYmmm2026-03-09 05:59Pre-processing error
- Issue comment#2058JJJYmmm2026-03-09 05:51AttributeError: Qwen2Tokenizer has no attribute all_special_tokens_extended. Did you mean: 'num_special_tokens_to_add'?
- Issue comment#2058JJJYmmm2026-03-09 05:51AttributeError: Qwen2Tokenizer has no attribute all_special_tokens_extended. Did you mean: 'num_special_tokens_to_add'?
- Issue comment#2052JJJYmmm2026-03-09 04:43For object detection tasks, how to set up the training format.
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 2,471 stars here means stars gained during the window, not the repo's star count.