Acoustic Signal Fidelity on Sarcastic Speech Synthesis Dataset
9.6MCDw/ BERT
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| w/ BERTSemantic Cue Type=✓, Prosodic Cue Type=-2025.10 | 9.6 | 265.6 | 4.5 | |
| BaselineSemantic Cue Type=-, Prosodic Cue Type=-2025.10 | 9.8 | 261.9 | 4.4 | |
| w/ LoRA + RAGSemantic Cue Type=✓, Prosodic Cue Type=✓2025.10 | 9.8 | 259.6 | 4.4 | |
| w/ RAGSemantic Cue Type=-, Prosodic Cue Type=✓2025.10 | 10 | 261 | 4.4 | |
| w/ LLaMA 3-LoRASemantic Cue Type=✓, Prosodic Cue Type=-2025.10 | 10.1 | 282.6 | 4.4 | |
| w/ LLaMA 3Semantic Cue Type=✓, Prosodic Cue Type=-2025.10 | 10.4 | 261.6 | 4.4 |