no code implementations • 31 Oct 2023 • Gabrielle Cohn, Rishika Agarwal, Deepanshu Gupta, Siddharth Patwardhan
We introduce EELBERT, an approach for compression of transformer-based models (e. g., BERT), with minimal impact on the accuracy of downstream tasks.
no code implementations • 4 Dec 2022 • Benjamin Muller, Deepanshu Gupta, Siddharth Patwardhan, Jean-Philippe Fauconnier, David Vandyke, Sachin Agarwal
For a given language, we are able to predict zero-shot performance, that increases on a logarithmic scale with the number of few-shot target language data points.