Feature - parallel training

[Ondrysak]

Are there any plans to implement training in parallel manner as shown in

https://arxiv.org/pdf/2102.11417.pdf

[arvoelke]

This has been implemented as keras_lmu.LMUFFT, which will be automatically selected if you use keras_lmu.LMU and satisfy these conditions:

keras-lmu/keras_lmu/layers.py

Lines 398 to 403 in ab07757

	if (
	not self.hidden_to_memory
	and not self.memory_to_memory
	and self.memory_d == 1
	and input_shapes[1] is not None
	):

There is still a bit of support that can be added for the RNN flags in #35 but let us know if this works for your use case.

[NarsimhaChilkuri]

For now, you might want to look at the implementation here. This is essentially the same as keras_lmu.LMUFFT , with two exceptions: 1) it supports multi dimensional input; and 2) when return_sequences=False, it implements equation (25) from the paper, which is more efficient.

[drasmuss]

Just a note that the multi dimensional input for LMUFFT is now supported in master.