Self-training from Self-memory in Data-to-text Generation
Open paper
Benchmark
Model
Rank
Results
data-to-text-generation-on-dart
self-mem + new data
#3
BLEU: 47.76
data-to-text-generation-on-e2e-nlg-challenge
Self-memory
#5
BLEU: 65.11
METEOR: 46.11
NIST: 8.35
ROUGE-L: 68.41
CIDEr: 2.16
Rank counts only results with a code link.