A Transformer-based Threshold-Free Framework for Multi-Intent NLU
Open paper
Benchmark
Model
Rank
Results
intent-detection-on-mixatis
TFMN
–
Accuracy: 79.8
intent-detection-on-mixsnips
TFMN
–
Accuracy: 97.7
slot-filling-on-mixatis
TFMN
–
Micro F1: 88.0
slot-filling-on-mixsnips
TFMN
–
Micro F1: 96.4
Rank counts only results with a code link.