Multimodal Dialogue Generation on SynMSI (test)
4.838Context RelevanceSynMSI (GroundTruth)
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| SynMSI (GroundTruth)2025.12 | 4.838 | — | — | — | — | — | 4.893 | — | |
| ViBES2025.12 | 4.584 | — | — | — | — | — | 4.376 | — | |
| LLM+SpeechBackbone=Llama22025.12 | 3.859 | — | — | — | — | — | 3.157 | — | |
| AnyGPTtuning strategy=fine-tune2025.12 | 3.803 | — | — | — | — | — | — | — | |
| SOLAMIFine-tuning Strategy=full parameter2024.11 | 3.634 | 3.443 | 8.853 | 151.5 | 0.36 | 0.824 | 3.824 | 2.639 | |
| DLPAlternative name=MotionGPT2024.11 | 3.577 | 4.254 | 8.259 | 165.053 | 0.495 | 0.812 | 3.785 | 5.518 | |
| DLPBackbone=MotionGPT2025.12 | 3.577 | — | — | — | — | — | 3.785 | — | |
| SOLAMIPre-training Status=w/o pretrain2024.11 | 3.541 | 5.052 | 8.558 | 159.709 | 0.387 | 0.82 | 3.461 | 2.657 | |
| LLM+SpeechBackbone=Llama22024.11 | 3.527 | — | — | — | — | 0.818 | 3.859 | 3.157 | |
| AnyGPTFine-tuning Strategy=fine-tune2024.11 | 3.502 | — | — | — | — | 0.819 | 3.803 | 2.588 | |
| SOLAMIFine-tuning Strategy=LoRA2024.11 | 3.251 | 15.729 | 8.145 | 167.149 | 0.4 | 0.77 | 3.423 | 2.71 | |
| SOLAMItuning strategy=LoRA2025.12 | 0.824 | — | — | — | — | — | 3.634 | — | |
| SOLAMItuning strategy=full params2025.12 | 0.824 | — | — | — | — | — | 3.634 | — |