1 paper · 1 filter
Igor V. Netay
We describe algorithms and data structures to extend a neural network library with automatic precision estimation for floating point computations. We also discuss conditions to mak…