Dialogue State Tracking On Simmc2 0

Act F1

Slot F1

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

			Paper Title
PaCE	97.1	87.0	PaCE: Unified Multi-modal Dialogue Pre-training with Progressive and Compositional Experts
BART-large	96.3	88.3	Learning to Embed Multi-Modal Contexts for Situated Conversational Agents
BART-base	95.2	82.0	Learning to Embed Multi-Modal Contexts for Situated Conversational Agents
GPT-2	94.5	81.7	Language Models are Unsupervised Multitask Learners
MTN	93.4	76.7	Multimodal Transformer Networks for End-to-End Video-Grounded Dialogue Systems

0 of 5 row(s) selected.