deepset/tinybert-6l-768d-squad2

Overview

Language model: deepset/tinybert-6L-768D-squad2
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 (based on TinyBERT). deepset/bert-base-uncased-squad2 was used as the teacher model and huawei-noah/TinyBERT_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

Authors

• Timo Möller: timo.moeller [at] deepset.ai
• Julian Risch: julian.risch [at] deepset.ai
• Malte Pietsch: malte.pietsch [at] deepset.ai
• Michel Bartels: michel.bartels [at] deepset.ai

About us

We bring NLP to the industry via open source!
Our focus: Industry specific language models & large scale QA systems.

Some of our work:

• German BERT (aka "bert-base-german-cased")
• GermanQuAD and GermanDPR datasets and models (aka "gelectra-base-germanquad", "gbert-base-germandpr")
• FARM
• Haystack

Get in touch: Twitter | LinkedIn | Discord | GitHub Discussions | Website

By the way: we're hiring!