CORTEXA
← Browse

Vladimer Khasia

1 paper indexed

openalexZenodo (CERN European Organization for Nuclear Research)2026-07-24

Memory-Efficient Exact Backpropagation

Vladimer Khasia

Training deep neural networks is constrained by the memory footprint of the computational graphduring automatic differentiation. This memory requirement scales linearly with model depth andsequence length, while the final projections contribute significantly to the spatial lower…

View free PDFSource page