Create module for interfacing with the clifford python package

[RobinKa]

Create a module for interfacing with Clifford.

• Convert from clifford's MVArray to TFGA tf.Tensor
• Convert from TFGA tf.Tensor to MVArray (or maybe that should be part of Clifford instead)

[hugohadfield]

I think it should be possible to convert from clifford MVArray by simply calling .value on the array. This should give an array of the dense arrays under the hood. Currently in clifford I think this will only work for 1D arrays of multivectors, but it is definitely something we would like to work on

[hugohadfield]

See:
https://github.com/pygae/clifford/blob/cc1c6eef29d511c4cab8c622e57107d960cde9ca/clifford/_mvarray.py#L28
and in fact we can go the other way as well from a numpy array:
https://github.com/pygae/clifford/blob/cc1c6eef29d511c4cab8c622e57107d960cde9ca/clifford/_mvarray.py#L35
I'll have a go today making a notebook to explore working with both libraries..

One thing that @eric-wieser and I were also discussing was building/converting the current clifford tools to be backend agnostic and work (hopefully) with any syntax compatible library

[pompeii222]

Could you please provide a version of pyTorch of TFGA? Because the pyTorch framework is used as the open source for many academic papers, we can hardly embde the TFGA into the open source code.

[pompeii222]

Hello, at present there is only a example of conv.ipynb, can you provide a complete example of convolutional neural network by using tfga