Hey there,
In week 4 we see a function so-called sp.ndimage.zoom for CAM purpose.
In week 3 we know that for blowing up encoded features we use UpSampling2D sometimes.
What’s the difference between them?
Thank you.
https://docs.scipy.org/doc/scipy/reference/generated/scipy.ndimage.zoom.html
One difference I noted is that
“The array is zoomed using spline interpolation of the requested order.”
https://www.tensorflow.org/api_docs/python/tf/keras/layers/UpSampling2D
Here, the interpolation is specified using a string, either nearest or bilinear .
So, the interpolation technique seems to be different in the two methods.
Am I to understand that these two functions are doing the same thing, or at least similar, but with different implementations and different results?
PS: There are many functions with the same or different names between np and tf, or keras, and it turns out that sometimes they actually do the same thing.
Hi @Chris.X ,
I think you’re right… Thank you…