Fix issue where ndims was incorrectly used to calculate shape of input

[umar456]

Fix the assertion to validate the input shape of the stride, padding and dilation array

Description

In C functions, the way the shapes of the input shapes are passed into functions is
using an array of dim_t values and an unsigned integer. The unsigned integer is used
to indicate how many values are in the input array. It does not indicate the number of
axis of the resulting dim4 object. This distinction caused an error with the new
convolve functions.

Changes to Users

Users will be able to pass any values to the stride_dims, padding_dims and dilation_dims
values.

Checklist

• Rebased on latest master
• Code compiles
• Tests pass
• [ ] Functions added to unified API
• [ ] Functions documented