[ ] Refactoring (no functional changes, no api changes)
[ ] Build related changes
[ ] Documentation changes
[ ] Other (please describe):
What is the current behavior?
Cannot save/load recurrent network with NetX.
What is the new behavior?
A new hdf5 layer category for dense recurrent layers allows saving recurrent networks to an hdf5 file.
NetX can load into Lava a recurrent network from an hdf5 file.
A unit test shows example usage. An example 2-layer 5-neuron dense recurrent network is created in Lava-dl, saved to a hdf5 file, and loaded into Lava from the hdf5 file. The network is run in both Lava-dl and Lava on an example input to confirm matching activity level within an error margin of 2%.
Issue Number: https://github.com/lava-nc/lava-dl/issues/327
Objective of pull request: As a user, I want to be able to save and load dense recurrent networks from lava-dl into lava.
Pull request checklist
Your PR fulfills the following requirements:
flakeheaven lint src/lava tests/
) and (bandit -r src/lava/.
) pass locallypytest
) passes locallyPull request type
Please check your PR type:What is the current behavior?
Cannot save/load recurrent network with NetX.
What is the new behavior?
A new hdf5 layer category for dense recurrent layers allows saving recurrent networks to an hdf5 file.
NetX can load into Lava a recurrent network from an hdf5 file.
A unit test shows example usage. An example 2-layer 5-neuron dense recurrent network is created in Lava-dl, saved to a hdf5 file, and loaded into Lava from the hdf5 file. The network is run in both Lava-dl and Lava on an example input to confirm matching activity level within an error margin of 2%.
Does this introduce a breaking change?