Views
No views yet
qwen_vl_native_loraQwen/Qwen3-VL-8B-InstructOSL-XFoul| Accuracy | Balanced Accuracy |
|---|---|
| 69.96% | 48.33% |
1from opensportslib.apis import VQAModel
2
3my_model = VQAModel(
4 config="opensportslib/configs/vqa/qwen3_vl_native.yaml",
5 weights="YOUR_HF_REPO_ID",
6)
7
8predictions = my_model.infer(
9 test_set="/path/to/test_annotations.json",
10)
11
12single_prediction = my_model.infer(
13 video_path="/path/to/video.mp4",
14 question="Was this a foul? What card should be given?",
15)
16
17print(predictions)
18print(single_prediction)opensportslib/configs/vqa/qwen3_vl_native.yaml.Qwen/Qwen3-VL-8B-Instruct.1@misc{opensportslib_qwen3_vl_vqa_xfoul_lora_2026,
2 title={OpenSportsLib Qwen3-VL LoRA for Soccer VQA},
3 author={OpenSportsLab},
4 year={2026},
5 howpublished={https://huggingface.co/OpenSportsLab}
6}