InternVL · Issues· 315 open
Open on GitHubLocally synced open issues (discussions stay on GitHub)
- #1101
How to output attention visualization of the InternVL3 model on images
Updated Sep 11, 2026 - #1254
RuntimeError: Tensor.item() cannot be called on meta tensors with Transformers 5.x InternVL 3.5 fails to load with Transformers v5 due to meta tensor initialization
Updated Sep 11, 2026 - #1223
[Bug] There is no `preprocessor_config.json` and `processor_config.json` in hf format model after custom2hf.py
Updated Sep 11, 2026 - #1249
evaluation_strategy
Updated Aug 20, 2026 - #1241
Generation repetition after MPO post-training
Updated Jul 13, 2026 - #1273
[Feature] 没有新模型了，是不是团队解散了？
Updated Jun 14, 2026 - #1270
[Docs] LM on which InternVL3.5 is built
Updated Jun 9, 2026 - #1272
纯视觉检测模型效果如何
Updated Jun 7, 2026 - #1268
[Feature] 3.5还没有微调脚本
Updated May 14, 2026 - #1262
坐标提取不准确
Updated Apr 24, 2026 - #1204
Internvl3.5-8b 和 internvl3.5-38b vllm推理乱码
Updated Apr 21, 2026 - #1150
关于InternVL3.5的Router training
Updated Apr 1, 2026 - #1267
Clarification on InternVL3-9B Architecture: Language Model Codebase (InternLM2 vs. InternLM3) and Wqkv Layout
Updated Mar 24, 2026 - #1265
[Bug] Invalid JSON format in `debug_mpo.json` causes JSONDecodeError
Updated Mar 17, 2026 - #1250
做video评测的时候16帧及以上会超上下文
Updated Mar 11, 2026