English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Large-scale transformer models have shown remarkable performance in languagemodelling tasks. However, such models feature billions of parameters, leadingto difficulties in their deployment and prohibitive training costs fromscratch. To reduce the number of the parameters in the GPT-2 architecture, wereplace the matrices of fully-connected layers with the corresponding TensorTrain Matrix~(TTM) structure. Finally, we customize forward and backwardoperations through the TTM-based layer for simplicity and the stableness offurther training. % The resulting GPT-2-based model stores up to 40% fewerparameters, showing the perplexity comparable to the original model. On thedownstream tasks, including language understanding and text summarization, themodel performs similarly to the original GPT-2 model. The proposed tensorizedlayers could be used to efficiently pre-training other Transformer models.

Efficient GPT Model Pre-training using Tensor Train Matrix Representation