using Flux, Flux.Data.MNIST, Statistics
using Flux: onehotbatch, onecold, crossentropy, throttle
using Base.Iterators: repeated, partition
# load the images
imgs = MNIST.images()
raw_labels=MNIST.labels()
# average the values in the image
imgs0 = imgs[raw_labels .== 0]
avg_imgs0 = reduce(+, float.(imgs0))./length(imgs0)
I see that eltype(imgs[1]) is ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}} and imgs[1] is a 28x28 matrix. But I can’t seem to convert avg_imgs0 which has eltype of Float back to imgs[1]'s eltype
I tried
#import Base.ColorTypes
ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}.(avg_imgs0)
How do I convert avg_imgs0 back to a matrix containing ColorTypes.Gray{FixedPointNumbers.Normed{UInt8,8}}?
