Natural Language Generation on WebNLG
55.3BLEULoRA
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| LoRABackbone=GPT-2, # Trainable Parameters=0.39M, # Total Parameters=100%2021.10 | 55.3 | 0.41 | 0.39 | |
| AdaptersBackbone=GPT-2, # Trainable Parameters=11.48M, # Total Parameters=100%2021.10 | 55.2 | 0.41 | 0.39 | |
| DSEEBackbone=GPT-2, # Trainable Parameters=0.39M, # Total Parameters=80%2021.10 | 54.9 | 0.44 | 0.39 | |
| PrefixBackbone=GPT-2, # Trainable Parameters=0.35M, # Total Parameters=100%2021.10 | 54.4 | 0.41 | 0.41 | |
| Fine-tuneBackbone=GPT-2, # Trainable Parameters=354.92M, # Total Parameters=100%2021.10 | 47.6 | 0.39 | 0.5 | |
| CriSPO2024.10 | 44.3 | — | — | |
| DSEEBackbone=GPT-2, # Trainable Parameters=0.39M, # Total Parameters=50%2021.10 | 42 | 0.33 | 0.53 | |
| Manualshots=3-shot2024.10 | 34.8 | — | — | |
| FT-Top2Backbone=GPT-2, # Trainable Parameters=25.19M, # Total Parameters=100%2021.10 | 33.5 | 0.26 | 0.75 | |
| OPRO2024.10 | 33.3 | — | — | |
| Manualshots=0-shot2024.10 | 31.3 | — | — |