NetX should allow user to utilize lava Sparse and DelaySparse synapses when converting SNNs from lava-dl to lava.

[Michaeljurado42]

User story

As a user, I want sparse lava-dl synapses to be interpreted as sparse or delay-sparse synapses in lava during netx conversion to reduce SNN memory usage and power consumption.

Conditions of satisfaction

• In my opinion the user should be able to specify a parameter in netx.hdf5.Network which specifies which layers should be converted to Sparse types

Acceptance tests

-- New unit tests will probably need to be written to test this feature.

[Michaeljurado42]

There is a draft pull request I have opened which enables this functionality. This feature was used by the SPANDEX team for the Intel Neuromorphic Deep Noise Suppression Challenge (Intel N-DNS Challenge). Before writing any test code, I was hoping that I could get some feedback about this new feature.

[bamsumit]

Thanks Michael, this sounds good. I just have one suggestion on the PR.