How to filter out NAN values from an array?

[kyrollosyanny]

Let's say I have an array that contains NAN values, I know that using '''mx.where''', we can replace the NAN values with other values. But here, I just want to remove them completely from the array. Is there a way to do this?

For example:

test=mx.array([1,2,3,mx.NAN,5,6])

test2= mx.remove_nan(test)

##output is test2=1,2,3,5,6

It seems that mlx does not support boolean indices so I am not sure how to filter out the NANs.

Thanks a lot for the help.

[awni]

MLX doesn't have operations where the output shape depends on the input data. You can read more about that here.

So to do what you want you could use mx.where with a sentinal or you'd need to go to numpy / python and then come back to MLX.