Efficient GPT Model Pre-training using Tensor Train Matrix Representation.

Viktoria Chekalina, Georgiy Novikov, Julia Gusak, Alexander Panchenko, Ivan V. Oseledets

Browse the full PACLIC paper archive.