Entity Recognition on Public Liver Ultrasound Datasets OOD (test)
91.06Hamming AccuracyEchoVLM
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| EchoVLMParameter Count=11B2025.09 | 91.06 | 58.57 | 60 | 59.25 | |
| LingshuParameter Count=7B2025.09 | 89.61 | 68.54 | 58.75 | 58.91 | |
| Gemma-3Parameter Count=12B2025.09 | 85.97 | 64.16 | 55.72 | 57.55 | |
| Qwen2.5-VLParameter Count=7B2025.09 | 85.28 | 60.85 | 54 | 56.19 | |
| HuatuoGPT-VisionParameter Count=7B2025.09 | 76.06 | 58.43 | 45.31 | 50.64 | |
| EchoVLMParameter Count=3B2025.09 | 63.63 | 64.31 | 32.97 | 40.78 | |
| MedgemmaParameter Count=4B2025.09 | 58.04 | 64.44 | 26.48 | 34.67 | |
| LLaVA-OneVisionParameter Count=7B2025.09 | 57.04 | 61 | 27.54 | 34.84 | |
| Qwen2-VLParameter Count=7B2025.09 | 56.32 | 63.83 | 24.77 | 32.5 | |
| LLaVA-NeXTParameter Count=13B2025.09 | 52.76 | 59.23 | 21.53 | 27.37 | |
| LLaVA1.5Parameter Count=7B2025.09 | 51.18 | 69.26 | 20.89 | 28.25 | |
| LLaVA-MedParameter Count=7B2025.09 | 50.73 | 65.78 | 20.35 | 27.36 |