多模态推理模型及训练方法
active 2025-05-21 → 2026-06-14 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 390 days from 2025-05-21 to 2026-06-14. Pushes: 13 total, peak 3 in a day. Pull requests: 25 total, peak 5 in a day. Issues: 62 total, peak 3 in a day. Comments: 193 total, peak 9 in a day. Stars: 607 total, peak 36 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 |
|---|---|---|---|---|
| JaaackHongggg | 43 | 3 | 1 | 39 |
| dependabot[bot] | 28 | 0 | 17 | 11 |
| ChenShawn | 25 | 9 | 6 | 8 |
| lky-violet | 11 | 0 | 0 | 9 |
| w-zhih | 10 | 0 | 0 | 5 |
| mengzchen | 8 | 0 | 0 | 5 |
| Alisaleelee | 6 | 0 | 0 | 3 |
| dddraxxx | 6 | 0 | 0 | 5 |
| RenlyH | 5 | 0 | 0 | 5 |
| wcpcp | 5 | 0 | 0 | 5 |
| FolSpark | 4 | 0 | 0 | 4 |
| zapqqqwe | 4 | 0 | 0 | 2 |
| EugeneLiu01 | 4 | 0 | 0 | 4 |
| lbj00jay | 4 | 0 | 0 | 3 |
| Daisy-Zhang | 3 | 0 | 0 | 2 |
| Copilot | 3 | 0 | 0 | 3 |
| qinguangming1999 | 3 | 0 | 0 | 3 |
| FloSophorae | 3 | 0 | 0 | 2 |
| sfc-gh-bzhai | 3 | 0 | 0 | 1 |
| chunyu-li | 3 | 0 | 0 | 3 |
Recent activity
Latest issues, pull requests and releases
- Issue#137liuchaohu2026-02-09 11:16The evaluation differences of the Vstar dataset
- Issue comment#31Michael49332026-01-27 12:11我得到了一个很奇怪的答案，这个答案合理么？
- Issue#136Tristan-Liu-132026-01-21 13:12请问下huggingface的模型怎么输出思维链呢？
- Issue comment#122Liac-li2025-12-04 01:27Qwen2.5VL 3B training with lora OOM on H100
- Issue comment#117JaaackHongggg2025-11-25 14:00模型推理慢
- Issue comment#103JaaackHongggg2025-11-25 13:48Model Response Score
- Issue comment#99JaaackHongggg2025-11-25 13:44About the training data
- Issue comment#80JaaackHongggg2025-11-25 13:32Does the Model Memorize Answers on the V* Benchmark?
- Issue#133snow-like-kk2025-11-13 13:15How can I remove flash-attn?
- Issue comment#131ckx5067720992025-11-03 08:25Using DeepEyes checkpoint fine-tuning, many <im_start> are repeatedly output when rollout.
- Issue#130Cyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy2025-10-31 07:50how to deploy to vllm and the deepspeed version unknown
- Issue#125pspdada2025-10-07 10:17Cannot find VISUAL_DATASET_TRAIN_0_8 file
- Issue comment#16pspdada2025-09-30 12:11request for evalution/inference code
- Issue comment#103Jam1ezhang2025-09-27 08:52Model Response Score
- Issue#123yuyi03122025-09-19 07:23How to get the tool_call curve if wandb is offline?
- Issue comment#113chunyu-li2025-09-12 10:17ray cpu memory OOM/cpu memory leak
- Issue comment#113chunyu-li2025-09-11 12:45ray cpu memory OOM/cpu memory leak
- Issue comment#113woshitff2025-09-11 12:23ray cpu memory OOM/cpu memory leak
- Issue comment#113chunyu-li2025-09-11 12:14ray cpu memory OOM/cpu memory leak
- Issue#122Liac-li2025-09-11 05:25Qwen2.5VL 3B training with lora OOM on H100
- Issue comment#99dearjohn122025-09-09 06:50About the training data
- Issue comment#102lbj00jay2025-09-09 06:34cannot import name 'process_image' from 'verl.utils.dataset.rl_dataset'
- Issue comment#78wcpcp2025-09-08 07:51感觉论文偏假，根本复现不出来，工具调用是0
- Issue comment#70lbj00jay2025-09-08 02:06How do I use these saved models?
- Issue comment#113woshitff2025-09-05 13:11ray cpu memory OOM/cpu memory leak
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 607 stars here means stars gained during the window, not the repo's star count.