- Main function are
FSE_signal
andMRF_Signal
- Parameters should be broadcastable in leading batch dimensions
- All tensor inputs should work with autograd
- Calculation will run in parallel along batch dimensions. This might cause OOM issues for large batches.