# Meaning of absolute tolerance and relative tolerance

Hello Data science community,

I am using a library nnet in R. The documentation, page no. 4-5, shows that it has default values for abstol and reltol as a stopping criteria.

nnet uses optim library for optimizing. I checked its documentation also. But I could not find the meaning abstol and reltol

abstol and reltol stands for absolute and relative convergence tolerance. I wish to change these values. But can someone please explain its meaning.

Thanks