powered by
threshold sets values of a vector/matrix below min to min; values above max are set to max.
threshold
min
max
threshold is mainly used to project sparsified weight vectors (sparsify_weights) back onto the probability simplex (thus min = 0 and then normalize).
sparsify_weights
min = 0
normalize
threshold(x, min = -Inf, max = Inf)
a numeric matrix(like object)
minimum value
maximum value
# NOT RUN { print(threshold(c(1, 4, 2, -1, 10), min = 0)) # }
Run the code above in your browser using DataLab