Spaces:
Running on Zero
Running on Zero
| title: OpenCaption-4B-VL-SFT | |
| emoji: 📝 | |
| colorFrom: indigo | |
| colorTo: pink | |
| sdk: gradio | |
| sdk_version: 6.24.0 | |
| app_file: app.py | |
| short_description: Dense fine-grained image captioning with Qwen3-VL-4B | |
| python_version: "3.12" | |
| startup_duration_timeout: 30m | |
| license: apache-2.0 | |
| # OpenCaption-4B-VL-SFT-v1.0 | |
| Dense, fine-grained image captioning with **OpenCaption-4B-VL-SFT-v1.0** — a vision-language model fine-tuned from Qwen3-VL-4B-Instruct for rich, structured image descriptions. | |
| Upload an image and the model produces a detailed, structured caption with thematic sections covering subjects, backgrounds, lighting, atmosphere, and fine visual details. | |
| ## Model | |
| - **Base**: [Qwen/Qwen3-VL-4B-Instruct](https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct) | |
| - **Fine-tune**: [prithivMLmods/OpenCaption-4B-VL-SFT-v1.0](https://huggingface.co/prithivMLmods/OpenCaption-4B-VL-SFT-v1.0) | |
| - **License**: Apache-2.0 |