ECG report generation on PTB-XL (test)
74.5PrecisionLLaMA-3-Instruct
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| LLaMA-3-InstructModel Size=8B, Fine-tuning=True2024.03 | 74.5 | 68.2 | 71.2 | |
| Mistral-InstructModel Size=7B, Fine-tuning=True2024.03 | 73 | 66.1 | 69.4 | |
| LLaMA-1Model Size=7B, Fine-tuning=True2024.03 | 72.5 | 65.7 | 68.9 | |
| LLaMA-2-InstructModel Size=7B, Fine-tuning=True2024.03 | 72.1 | 66.8 | 69.3 | |
| OPTModel Size=6.7B, Fine-tuning=True2024.03 | 71.2 | 64.8 | 67.8 | |
| MistralModel Size=7B, Fine-tuning=True2024.03 | 71.1 | 66.4 | 68.7 | |
| BLOOMModel Size=7B, Fine-tuning=True2024.03 | 70.1 | 64.5 | 67.2 | |
| GPT-JModel Size=6B, Fine-tuning=True2024.03 | 68.9 | 62.2 | 65.4 | |
| GPT-NeoModel Size=2.7B, Fine-tuning=True2024.03 | 67.5 | 58.8 | 62.8 | |
| GPT-NeoXModel Size=20B, Fine-tuning=True2024.03 | 65.4 | 57.9 | 61.4 | |
| GPT2-LargeModel Size=774M, Fine-tuning=True2024.03 | 62.5 | 55.3 | 58.6 | |
| GPT2-MediumModel Size=345M, Fine-tuning=True2024.03 | 53.4 | 46.5 | 49.7 |