Dynamic Self-Attention : Computing Attention over Words Dynamically for Sentence Embedding

Benchmark Model Rank Results
natural-language-inference-on-snli2400D Multiple-Dynamic Self-Attention Model#29% Test Accuracy: 87.4% Train Accuracy: 89.0Parameters: 7.0m
natural-language-inference-on-snli600D Dynamic Self-Attention Model#33% Test Accuracy: 86.8% Train Accuracy: 87.3Parameters: 2.1m