1 paper · 1 filter
Kevin Coakley, Christine R. Kirkpatrick, Odd Erik Gundersen
Reproducing published deep learning papers to validate their conclusions can be difficult due to sources of irreproducibility. We investigate the impact that implementation factors…