Overview

Language model: Shobhank-iiitdwd/DistBERT-squad2-QA
Language: English
Training data: SQuAD 2.0 training set x 20 augmented + SQuAD 2.0 training set without augmentation
Eval data: SQuAD 2.0 dev set
Infrastructure: 1x V100 GPU
Published: Dec 8th, 2021

Details

• haystack's intermediate layer and prediction layer distillation features were used for training. bert-base-uncased-squad2 was used as the teacher model and DBERT_General_6L_768D was used as the student model.

Hyperparameters

Intermediate layer distillation

batch_size = 26
n_epochs = 5
max_seq_len = 384
learning_rate = 5e-5
lr_schedule = LinearWarmup
embeds_dropout_prob = 0.1
temperature = 1

Prediction layer distillation

batch_size = 26
n_epochs = 5
max_seq_len = 384
learning_rate = 3e-5
lr_schedule = LinearWarmup
embeds_dropout_prob = 0.1
temperature = 1
distillation_loss_weight = 1.0

Performance

"exact": 71.87736882001179
"f1": 76.36111895973675