README.md

learning rate: 2e-5
training epochs: 3
batch size: 64
seed: 87
model: bert-base-uncased
trained on MNLI which is converted into two-way nli classification (predict entailment or not-entailment class)