Model Description
BioMobileBERT is the result of training the MobileBERT-uncased model in a continual learning scenario for 200k training steps using a total batch size of 192 on the PubMed dataset.
Initialisation
We initialise our model with the pre-trained checkpoints of the MobileBERT-uncased model available on Huggingface.
Architecture
MobileBERT uses a 128-dimensional embedding layer followed by 1D convolutions to up-project its output to the desired hidden dimension expected by the transformer blocks. For each of these blocks, MobileBERT uses linear down-projection at the beginning of the transformer block and up-projection at its end, followed by a residual connection originating from the input of the block before down-projection. Because of these linear projections, MobileBERT can reduce the hidden size and hence the computational cost of multi-head attention and feed-forward blocks. This model additionally incorporates up to four feed-forward blocks in order to enhance its representation learning capabilities. Thanks to the strategically placed linear projections, a 24-layer MobileBERT (which is used in this work) has around 25M parameters.
Citation
If you use this model, please consider citing the following paper:
@article{rohanian2023effectiveness,
title={On the effectiveness of compact biomedical transformers},
author={Rohanian, Omid and Nouriborji, Mohammadmahdi and Kouchaki, Samaneh and Clifton, David A},
journal={Bioinformatics},
volume={39},
number={3},
pages={btad103},
year={2023},
publisher={Oxford University Press}
}
Support
If this model helps your work, you can keep the project running with a one-off or monthly contribution:
https://github.com/sponsors/nlpie-research