| base_model: Qwen/Qwen3-VL-8B-Instruct | |
| license: apache-2.0 | |
| tags: | |
| - camera-movement | |
| - video-understanding | |
| - qwen3-vl | |
| - sft | |
| - vggt-injection | |
| library_name: transformers | |
| pipeline_tag: video-text-to-text | |
| # CamInject-8B | |
| Camera-movement VGGT-Direct 注入 SFT 微调模型，基于 `Qwen/Qwen3-VL-8B-Instruct`。 | |
| - Checkpoint: `checkpoint-1326` | |
| - 训练框架: [ms-swift](https://github.com/modelscope/ms-swift) | |
| This model is described in the paper [Temporally Grounded Compositional Camera Motion Understanding via Geometric Knowledge Distillation](https://huggingface.co/papers/2608.10932). | |
| Project page: https://ddz16.github.io/cammotion.github.io | |
| Code: https://github.com/ddz16/CamDistill | |
| ## 使用 | |
| ```python | |
| from transformers import AutoModelForCausalLM, AutoProcessor | |
| model = AutoModelForCausalLM.from_pretrained("ddz16/CamInject-8B", torch_dtype="bfloat16", device_map="auto") | |
| processor = AutoProcessor.from_pretrained("ddz16/CamInject-8B") | |
| ``` |